Sie sind auf Seite 1von 208

T0 c

o
e
x
i
s
t
e
n
c
eo
fpeaceandharmonybetweent
h
ed
i
f
f
e
r
e
n
t
melodieso
fpeopleeverywhere

MA
n.ABi
sar
e
g
i
s
l
e
r
e
dl
r
a
d
e
m
a
r
ko
f l
eMalhWorksI
n
c
.
l
e
do
nn
c
i
d

T
h
i
sb
o
o
ki
sp

paper.@

yJ
o
h
nW
i
l
e
y& S
o
n
sNewY
o
r
k
.A
I
Ir
i
g
h
l
s e
d
.
C
o
p
y
r
i
g
h
l
@
2 2b
P
u
b
l
i
s
h
e
ds
i
m
u
l
t
a
n
e
o
u
s
l
yi
nC
n
n
a
d
n
.
Nop
n
r
to
fI
h
i
sp
u
b
l
i
c
l
l
l
i
o
nmayber
e
p
r
o
d
u
d

s
t
o
r
e
di
na t
r
i
e
v
s
y
s
t
e
mo
r1
n
s
m
i
t
t
e
di
nn
n
yf
o
r
mo
r
n
s

e
l
e
c
l
r
o
n
i
cm
e
c
h
a
n
i
c
a
J
p
h
o
l
o
c
o
p
y
i
n
g
c
o
r
d
i
n
g

s
c
a
n
n
i
n
go
ro
t
h
e
r
w
i
s
e
e
x
c
e
p
ta
s
byanym
p
e
r
m
i
t
t
c
du
n
d
e
rS
e
c
t
i
o
n
s1
0
7o
r1
0
8o
f
t
h
e1
9
7
6U
n
i
t
e
d
S
t
a
l
c
sC
o
p
y
r
i
g
h
lA
c
r
w
i
r
h
o
u
le
i
t
h
e
rt
h
ep
r
i
o
r
w
r
i
t
t
c
n
p
r
m
i
s
s
i
o
no
fI
h
eP
u
b
l
i
s
h
e
r

o
ro
u
t
h
o
r
i
7
.
3l
i
o
nt
h
r
o
u
g
hp
a
y
m
e
n
lo
ft
h
ea
p
p
r
o
p
r
i
u
l
ep
e
r
c
o
p
yf
c
c1
0
I
h
eC
o
p
y
r
i
g
h
lC
l
e
a
r
a
n
c
eC
e
n
l
e
r

2
2
2Ro
W
dD
r
i
v
eD
n
n
v
e
r

MA0
1
9
2
3(
9
7
8
)7
5
0
8
4
0
0

f
n
x(
9
7
8
)
4
7
4
4
.R
e
q
u
I
S1
0I
h
eP
u
b
l
i
s
h
e
rf
o
rp
e
r
m
i
s
s
i
o
ns
h
o
u
l
dben
d
d
r
e
s
s
e
d1
0t
h
eP
e
m
l
i
s
s
i
o
n
sD
e
p
a
r
t
m
e
n
l
7
5
0
J
o
h
nW
i
l
c
y& S
o
n
sI
n
c
.
6
0
5 i
r
dA
v
e
n
u
e
NewY
o
r
kNY1
0
1
5
8
- 1
2
(
2
1
2
)8
5
0

1
1 (
2
1
2
)
M
a
i
l
:PERMREQ@WILEY.COM.
8
50
-6
0
0
8E

u
b
l
i
c
a
l
i
o
ni
sd
e
s
i
g
n
e
d1
0p
r
o
v
i
d
ea
c
c
u
l
en
n
da
u
t
h
o
r
i
l
a
l
i
v
ei
n
f
o
m
t
a
l
i
o
ni
n E
1
0es
u
b
j
e
c
l
sp
d
.I
ti
s8
0
l
dw
i
t
ht
h
eu
n
d
e
r
s
l
u
n
d
i
n
gI
h
a
lt
h
ep
u
b
l
i
s
h
e
ri
sn
o
le
n
g
a
g
e
di
nr
e
n
d

n
g
m
a
t
t
e
r
c
o
v
e
u
i
r
e
des
e
r
v
i
c
e
so
fa
p
r
o
f
e
s
s
i
o
n
a
ls
e
r
v
i
c
e
s
.I
fp
r
o
f
e
s
s
i
o
n
n
ln
d
v
i
c
eo
ro
t
h
c
re
x
p
e
r
ta
s
s
i
s
l
a
n
c
ei
s q
as
h
o
u
l
db
es
o
u
g
hl
.
c
o
m
p
e
l
e
n
tp
r
o
f
e
s
s
i
o
n
a
lp
e n

U
b
r
a
r
yo
f
C
o
n
g
r
e
s
sC
a
t
a
l
o
g
;
l
I
g
;
nP
"
b
l
;
c
a
t
i
o
t
lDat
a
:
V
e
n
k
o
l
u
r
o
m
n

P
.

A
p
p
l
i
e
do
p
l
i
m
iZlll
i
o
nw
i
t
hMATLAB~ P
r
o
g
r
a
m
m
i
n
gIP
.V
e
n
k
a
l
a
r
m
l
l
n
.
p
.c
o
m
.

AW
i
l
e
y
I
n
t
e
r
s
c
i
e
n
c
cp
u
b
l
i
c
n
l
i
o
n
.
"
5
8

.
5(
c
1o
t
h:
a
l
k
.p
n
p
e
r
)
ISBN0
4
7
1 9
i
n
g2
.MATLAB.J
.T
i
t
l
e
.
1
.M
n
l
h
e
m
a
l
i
c
a
Jo
p
t
i
m
i
z
a
t
i
o
n
D
o
l
l
lp
r
o
c
Q
A
4
0
2
.
5
.
V
4
22
0
0
1
5
1
9
.
3
d
c
2
1

2
0
0
1
0
2
6
9
3
8

P
r
in
l
e
di
nI
h
eU
n
i
l
e
dS
l
a
l
e
so
fA
m
e
r
i
c
n
.
10987654321

~.4r;r :T
--

CONTENTS

PREFACE
1 I
n
t
r
o
d
u
c
t
l
o
n

x
i
i
i
1

1
.
1 O
p
t
i
m
i
z
a
t
i
o
nF
u
n
d
a
m
e
n
t
a
l
sI2
1
.
1
.
1 E
l
e
m
e
n
t
so
fP
r
o
b
l
e
mF
o
r
m
u
l
a
t
i
o
nI4
1
.
1
.2 M
a
t
h
e
m
a
t
i
c
a
lM
o
d
e
l
i
n
gI1
0
1
.
1
.3 N
a
t
u
r
eo
fS
o
l
u
t
i
o
nI1
6
1
.
1.
4 C
h
a
r
a
c
t
e
r
i
s
t
i
c
so
ft
h
eS
e
a
r
c
hP
r
o
c
e
d
u
r
eI20
1
.
.
As I2
5
1
.
2 I
n
t
r
o
d
u
c
t
i
o
nt
oMA1
1
.2
.
1 WhyMATLAB?I25
1
.2
.
2 MATLABI
n
s
t
a
l
l
a
t
i
o
nI
s
s
u
e
sI26
1
.2.
3 UsingMATLABt
h
eF
i
r
s
tTimeI2
7
1
.
2.
4 U
s
i
n
gt
h
eE
d
i
t
o
r/3
3
1
.
2.
5 CreatingaCodeS
n
i
p
p
e
t/37
1
.2
.
6C
r
e
a
t
i
n
gaProgramI40

P
r
o
b
l
e
m
sI44
2 GraphlcalOptlmizatlon
2
.
1

45

b
lemD
e
f
i
n
i
t
i
o
nI45
2
.1
.
1 E
xample2.
1 I46
2
.1
.2 F
o
r
m
a
tf
o
rl
h
eG
r
a
p
h
i
c
a
lD
i
s
p
l
a
yI47

2
.
2 G
r
a
p
h
i
c
a
lS
o
l
u
t
i
o
nI48
2
.
2
.
1 MATLABH
i
g
h
-Lev
e
lG
r
a
p
h
i
c
sF
u
n
c
t
i
o
n
sI48
2
.
2
.
2 Example2
.
I
G
r
a
p
h
i
c
a
lS
o
l
u
t
i
o
nI50
2
.
2.
3D
i
s
p
l
a
y
i
n
gt
h
eG
r
a
p
h
i
c
sI5
3
2
.
2.
4 C
u
s
t
o
m
i
z
i
n
gt
h
eF
i
g
u
r
eI54
2
.
3 A
d
d
i
t
i
o
n
a
lExamplesI56
v
l
l

v
l
l
l

CONTENTS

CONTENTS

2.
3.
1 Example2
.
2I56
2.
3.
2 Example2.
3I64
2.
3.
3 Example2.
4I7
3

4
.
1
.
1 P
r
o
b
l
e
mFormulation-Example4
.
1 /1
5
5
a
i
n
t
s/1
5
7
4
.
1
.
2D
i
s
c
u
s
s
i
o
no
fCons
4
.
2 M
a
t
h
e
m
a
t
i
c
a
lC
o
n
c
e
p
t
s/1
5
9
4
.
2
.
1 S
y
m
b
o
l
i
cC
o
m
p
u
t
a
t
i
o
nU
s
i
n
gMATLAB/1
5
9
a
t
h
e
m
a
t
i
c
a
lC
o
n
c
e
p
t
sI1
6
2
4
.
2
.
2 Ba~ic M
o
r
e
m
/
S
e
r
i
e
s/1
6
9
4
.
2
.
3 T
a
y
l
o
r
'
sThe

2.
4A
d
d
i
t
i
o
n
a
lMA
1
1
.
.
A
BG
r
a
p
h
i
c
sI79
4.
1 H
a
n
d
l
eG
r
a
p
h
i
c
sI80
2.
4.
2 G
r
a
p
h
i
c
a
lU
s
e
rI
n
t
e
r
f
a
c
eI8
1
2.
2
.
4
.
3 GUICodeI84
R
e
f
e
r
e
n
c
e
sI9
1
P
r
o
b
l
e
m
sI9
2
3 LlnearProgrammlng

93

3
.
2 G
r
a
p
h
i
c
a
lS
o
l
u
t
i
o
n/1
0
7
3
.
2
.
1 Example3
.
1 /1
1
0
3
.
2
.
2 C
h
a
r
a
c
t
e
r
i
s
t
i
c
so
ft
h
eS
o
l
u
t
i
o
n/I
I
1
3 D
i
f
f
e
r
e
n
tS
o
l
u
t
i
o
nTypesI1
1
4
3
.
2.

n
p
l
e
s/1
9
4
4
.
5 Ex
4
.
5
.
1 Example4.2/ 1
9
4
3/1
9
6
4
.
5
.
2 Example4.

3
.
3 N
u
m
e
r
i
c
a
lS
o
l
u
t
i
o
n
t
h
eS
i
m
p
l
e
xMethodI1
1
5
3
.
3
.
1 F
e
a
t
u
r
e
so
ft
h
eS
i
m
p
l
e
xMethod/1
1
5
3.
2 A
p
p
l
i
c
a
t
i
o
no
fS
i
m
p
l
e
xMethod/1
1
7
3.
3.
3 S
o
l
u
t
i
o
nU
s
i
n
gMA
1
1
.
.
A
B/1
2
0
3.
4 S
o
l
u
t
i
o
nU
s
i
n
gM
A
T
L
A
B
'
SO
p
t
i
m
i
z
a
t
i
o
nT
o
o
l
b
o
xI123
3
.
3.

R
e
f
e
r
e
n
c
e
s/200
P
r
o
b
l
e
m
s/2
0
1

e
-DlmenslonalProblem
5 NumericalTechnlquesTheOn

3
.
4 A
d
d
i
t
i
o
n
a
lE
x
a
m
p
l
e
s/1
2
4
4.
1 Example3
.
2
T
r
a
n
s
p
o
r
t
a
t
i
o
nP
r
o
b
l
e
m/1
2
4
3.
4.
2 Example3.3-Equ
a
l
i
t
yC
o
n
s
t
r
a
i
n
t
sa
n
dU
n
r
e
s
c
t
e
d
3.
V
a
r
i
a
b
l
e
s/1
3
0
4.
3 Example3.
4- AF
o
u
r
V
a
r
i
a
b
l
eP
r
o
b
l
e
m/1
3
4
3.

203

r
oblemD
e
f
i
n
i
t
i
o
nI204
5
.
1 P
5
.1
.
1 C
o
n
s
t
r
a
i
n
e
dOne-D
i
m
e
n
s
i
o
n
a
lP
r
oblem/204
5
.
2 S
o
l
u
t
i
o
nt
ot
h
eP
r
o
b
l
e
m/2
0
5
5
.
2
.
1 G
r
a
p
h
i
c
a
lS
o
l
u
t
i
o
n/2
0
5
5
.
2
.
2 Newton-RaphsonT
e
c
h
n
i
q
u
e/206
3B
i
s
e
c
t
i
o
nT
e
c
h
n
i
q
u
eI2
0
9
5
.
2.
5
.
2.
4P
o
l
y
n
o
m
i
a
lA
p
p
r
o
x
i
m
a
t
i
o
n/2
1
1
l
d
e
nS
e
c
t
i
o
nMethodI214
5
.
2
.
5 Go

3.
5 A
d
d
i
t
i
o
n
a
lT
o
p
i
c
si
nL
i
n
e
a
rProgramming/1
3
8
3
.
5
.
1 P
r
i
m
a
la
n
dD
u
a
lP
r
o
b
l
e
m/1
3
8
3
.
5
.
2 S
e
n
s
i
t
i
v
i
t
yA
n
a
l
y
s
i
s/1
4
8
n
c
e
s/1
5
1
R
e
f
e
r

5
.
3 I
m
p
o
r
t
a
n
c
eo
ft
h
eO
n
e
D
i
m
e
n
s
i
o
n
a
lP
r
oblem/217

P
r
o
b
l
e
m
s/1
5
2

r
oblemD
e
f
i
n
i
t
i
o
nI1
5
5
4
.
1 P

3 G
r
a
p
h
i
ca
1S
o
l
u
t
i
o
n
s/1
7
1
4.
4
.
3
.
1 U
n
c
o
n
s
t
r
a
i
n
e
dP
r
o
b
l
e
m/1
7
1
4
.
3
.
2E
q
u
a
l
i
t
yC
o
n
s
t
r
a
i
n
e
dP
r
o
b
l
e
m/1
7
2
3 I
n
e
q
u
a
l
i
t
yC
o
n
s
t
r
a
i
n
e
dP
r
o
b
l
e
m/1
7
3
4
.
3.
3
.
4 Eq
u
a
l
i
t
ya
n
dI
n
e
q
u
a
l
i
t
yC
o
n
s
t
r
a
i
n
t
s/1
7
4
4.

4 A
n
a
l
y
t
i
ca
1C
o
n
d
i
t
i
o
n
s/1
7
5
4.
4
.
1 U
n
c
o
n
s
t
r
a
i
n
e
dP
r
ob
lem/1
7
6
4.
4.
2 Eq
u
a
l
i
t
yCons
i
n
e
dP
r
o
b
l
e
mI1
7
9
4.
4.
4.
3 I
n
e
q
u
a
l
i
t
yC
o
n
s
t
r
a
i
n
e
dO
p
t
i
m
i
z
a
t
i
o
nI1
8
6
4
.
4G
e
n
e
r
a
lO
p
t
i
m
i
z
a
t
i
o
nP
r
o
b
l
e
m/1
9
1
4.

3
.
1 P
r
oblemD
e
f
i
n
i
t
i
o
nI94
3
.
1
.
1 S
t
a
n
d
a
r
dF
o
r
m
a
t/94
3
.
1
.
2 M
o
d
e
l
i
n
gI
s
s
u
e
s/98

4 NonlinearProgramming

I
x

154

4 A
d
d
i
t
i
o
n
a
lE
x
a
m
p
l
e
sI2
1
9
5.
m
p
l
e5.2I
l
1
u
s
t
i
o
no
fG
e
n
e
m
lGo
l
d
e
nS
e
c
t
i
o
nM
e
t
h
o
d/2
1
9
5
.
4
.
1 a
4.
2 Example5
.
3 Tw
o
P
o
i
n
tBoundaryV
a
l
u
eP
r
o
b
l
e
mI220
5.
5.
4.
3 Example5.4-RootF
i
n
d
i
n
gw
i
t
hG
o
l
d
e
nS
e
c
t
i
o
nI223

7
.
3.
3 Gener
a
1i
z
e
dR
e
d
u
c
e
dG
r
a
d
i
e
n
t(GRG)MethodI297
7.
3.
4 S u
e
n
t
i
a
lG
r
a
d
i
e
n
tR
e
s
t
o
r
a
t
i
o
nA
J
g
o
r
i
t
h
m(SGRA)I302

R
e
f
e
r
e
n
c
e
sI225

P
r
o
b
l
e
m
sI225
6 NumerlcalTechnlquesf
o
rUnconstralnedOptlmlzatlon

x
l

CONTENTS

CONTENTS

227

6
.
1 P
r
o
b
l
e
mD
e
f
i
n
i
t
i
o
nI2
2
7
6
.1
.
1 Example6
.
1 I2
2
8
6
.
1
.
2 N
e
c
e
s
s
a
r
ya
n
dS
u
f
f
i
c
i
e
n
tC
o
n
d
i
t
i
o
n
sI2
2
8
6
.
1
.
3 E
l
e
m
e
n
t
so
faNumeric
a
1T
e
c
h
n
i
q
u
eI229

7.
4A
d
d
i
t
i
o
na
1ExampJesI307
7.
4.
1 ExampJe7
.
2
F
J
a
g
p
o
J
eP
r
o
b
l
e
mI307
7.
4.
2 ExampJe7.3-BeamD
e
s
i
g
nI310
7.
4.
3 ExampJe7.
4
Optim
a
1C
o
n
t
r
o
lI313
R
e
f
e
r
e
n
c
e
sI316
P
r
o
b
l
e
m
sI316

6
.
2 N
u
m
e
r
i
c
a
lT
e
c
h
n
i
q
u
e
s
N
o
n
g
r
a
d
i
e
n
tM
e
t
h
o
d
sI230
6
.
2
.
1 RandomWa
1kI230
6
.
2
.
2 P
a
t
t
e
mS
e
a
r
c
h/234
6
.
2.
3 Powe
J
1
'sM
ethodI238

8D
l
s
c
r
e
t
eOptlmlzation

318

8
.
1 C
o
n
c
e
p
t
si
nD
i
s
c
r
e
t
eP
r
o
g
r
a
m
m
i
n
g/320
8
.1
.
1 ProblemR
e
l
a
x
a
t
i
o
nI3
2
1
.2 D
i
s
c
r
e
t
eO
p
t
i
m
a
lS
o
l
u
t
i
o
nI322
8
.1

6
.
3 N
u
m
e
r
i
c
a
lTh
n
i
q
u
e
s
G
r
a
d
i
e
n
t
B
a
s
e
dMethodsI2
4
1
'
6.
3.
1 S
t
e
e
p
e
s
tD
e
s
c
e
n
tMethodI2
4
1
6.
3.
2 C
o
n
j
u
g
a
t
eG
r
a
d
i
e
n
t(
F
1
e
t
c
h
e
r
R
e
e
v
e
s
)MethodI244
6
.
3.
3 D
a
v
i
d
o
n
F
J
e
t
c
h
e
r
P
o
w
el
IMethodI246
6
.
3.
4B
r
o
y
d
o
n
-F
l
e
t
c
h
e
r
G
o
l
d
f
a
r
t
S
h
a
n
n
oMethodI249

8
.
2 D
i
s
c
r
e
t
eO
p
t
i
m
i
z
a
t
i
o
nT
e
c
h
n
i
q
u
e
sI324
8
.
2
.
1 E
x
h
a
u
s
t
i
v
eE
n
u
m
e
r
a
t
i
o
nI326
8
.
2
.
2 B
r
a
n
c
ha
n
dBoundI329
8
.
2.
3 DynamicProgramming/336

6
.
4N
u
m
e
r
i
c
a
lTechniques-SecondO
r
d
e
rI2
5
1
6
.
5 A
d
d
i
t
i
o
n
a
lE
x
a
m
p
l
e
sI253
6
.
5
.
1 Example6.2-RosenbrockP
r
o
b
l
e
m/2
5
3
6
.
5
.
2 E
x
a
m
p
l
e6
.3Three
D
i
m
e
n
s
i
o
n
a
lAown
r
aRo
n
g
D
i
s
kI2
5
5
6
.
5
.
3 ExampJe6.
4
F
i
t
t
i
n
gB
e
z
i
e
rP
a
r
a
m
e
t
r
i
cC
u
r
v
c
sI258

8
.
3 A
d
d
i
t
i
o
n
a
lE
x
a
m
p
l
e
sI3
4
1
8
.
3
.
1 Exnmple8
.
41BeumD
e
s
i
g
nI3
4
1
8
.
3
.
2 Zero-OneI
n
t
e
g
e
rP
r
o
g
r
a
m
m
i
n
gI343
R
e
f
e
r
e
n
c
e
sI348
P
r
o
b
l
e
m
sI348

R
e
f
e
r
e
n
c
e
sI262
9 GlobalOptlmlzation

P
r
o
b
l
e
m
sI263
7 NumerlcalTechniquesf
o
rConstralnedO
p
t
l
m
l
z
a
t
l
o
n
7
.
1 P
r
o
b
l
e
mD
e
f
i
n
i
t
i
o
nI266
7
.1
.
1 ProblemFormulation-Example7
.
1 I266
7
.1
.2 N
e
c
e
s
s
a
r
yC
o
n
d
i
t
i
o
n
s/2
6
7
7.
1
.
3 E
l
e
m
e
n
t
so
faN
u
m
e
r
i
c
a
lT
e
c
h
n
i
q
u
eI269
7
.
2 I
n
d
i
r
e
c
tM
e
t
h
o
d
sf
o
rC
o
n
s
t
r
a
i
n
e
dO
p
t
i
m
i
z
a
t
i
o
nI270
7
.
2
.
1 E
x
t
e
r
i
o
rP
e
n
a
l
t
yF
u
n
c
t
i
o
n(EPF
)Me
o
dI2
7
1
7
.
2
.
2 AugmentedL
a
g
r
a
n
g
eM
u
l
t
i
p
l
i
e
r(ALM)MethodI276
7
.
3 D
i
r
e
c
tM
e
t
h
o
d
sf
o
rCons
a
i
n
e
dO
p
t
i
m
i
z
a
t
i
o
nI2
8
1
7
.
3
.
1 S
e
q
u
e
n
t
ia
1L
in
e
a
rP
r
o
g
r
a
m
m
i
n
g(
S
L
P
)I284
7.
3.
2 S
e
q
u
e
n
t
i
a
lQ
u
a
d
r
a
t
i
cP
r
ogramming(
S
Q
P
)I289

265

350

9
.
1 P
r
o
b
l
e
mD
e
f
i
n
i
t
i
o
nI3
5
1
9
.1
.
1 G
l
o
b
a
lMinimum/3
5
1
9
.
1
.
2 N
a
t
u
r
eo
ft
h
eS
o
l
u
t
i
o
nI354
9
.1
.3 E
l
e
m
e
n
t
so
faN
u
m
e
r
i
ca
1T
e
c
h
n
i
q
u
eI356
9
.
2N
u
m
e
r
i
c
a
lTh
n
i
q
u
s
a
n
dA
d
d
i
t
i
o
n
a
lEx
a
m
p
l
e
sI357
9
.
2
.
1 S
i
m
u
l
a
t
e
dA
n
n
e
a
l
i
n
g(SA)I358
9
.
2
.
2G
e
n
e
t
i
cA
l
g
o
r
i
t
h
m(GA)I366
R
e
f
e
r
e
n
c
e
sI3
7
7
7
8
P
r
o
b
l
e
m
sI3
10 O
p
t
l
m
i
z
a
t
i
o
nToolboxfromMA
TlA
B
1
0
.
1 TheO
p
t
i
m
i
z
a
t
i
o
nT
o
o
l
b
o
xI380

379

x
l
l CONTENTS
1
0
.1
.1 ProgramsI380
1
0
.1
.2 UsingProgramsI382
1
0
.
1
.
3 S
e
t
t
i
n
gO
p
t
i
m
i
z
a
t
i
o
nP
a
r
a
m
e
t
e
r
sI384
1
0
.
2 E
x
a
m
p
l
e
sI3
8
5
1
0
.
2.
1
1
0
.
2
.
2
1
0
.
2
.
3
1
0
.
2.
4

PREFACE

L
in
e
a
rP
r
ogrammingI3
8
5
Q
u
a
d
r
a
t
i
cP
r
ogrammingI386
U
n
c
o
n
s
t
r
a
i
n
e
dO
p
t
i
m
i
z
a
t
i
o
nI388
C
o
n
s
t
r
a
i
n
e
dO
p
t
i
m
i
z
a
t
i
o
nI3
8
9

R
e
f
e
r
e
n
c
eI3
9
1
Index

393

l
esu~ject o
fo
p
t
i
m
i
z
a
t
i
o
ni
sr
e
c
e
i
v
i
n
gs
e
r
i
o
u
sa
t
t
e
n
t
i
o
nome
n
g
i
n
e
e
r
s
s
c
i
e
n
t
i
s
t
s

m
a
n
a
g
e
r
s
a
n
dm
o
s
te
v
e
r
y
b
o
d
ye
l
s
e
.T
h
i
si
sd
r
i
v
e
nbycompe t
i
o
n
q
u
a
l
i
t
ya
s
s
u
r
a
n
c
e
c
o
s
to
fp
r
o
d
u
c
t
i
o
na
n
df
i
n
a
l
l
yt
h
es
u
c
c
e
s
so
ft
h
eb
u
s
i
n
e
s
se
n
t
e
e
.
l
g
n
o
r
i
n
ge
p
r
a
c
t
i
c
eo
fo
p
t
i
m
i
z
a
t
i
o
ni
sn
o
tano
p
t
i
o
nd
u
r
i
n
gc
u
r
r
e
n
tt
i
m
e
s
.
O
p
t
i
m
i
z
a
t
i
o
ni
sp
r
a
c
t
i
c
e
dt
h
r
o
u
g
hs
o
f
t
w
a
r
p
r
o
g
r
a
m
sa
n
dr
e
q
u
i
r
e
ss
i
g
n
i
f
i
c
a
n
t
et
e
c
h
n
i
q
u
e
so
fo
p
t
i
m
i
z
a
t
i
o
nh
a
v
en
o
tc
h
a
n
g
e
ds
i
g
n
i
f
i
c
a
n
t
l
yi
n
c
o
m
p
u
t
e
rr
e
s
o
u
r
c
e
s
.Th
c
e
n
ty
e
a
r
sb
u
tt
l
l
ea
r
e
a
so
fa
p
p
l
i
c
a
t
i
o
n
sh
a
v
emushroomeda
tas
i
g
n
i
f
i
c
a
n
tr
a
t
e
.
S
u
c
c
e
s
s
f
u
l
l
yembeddingt
h
eu
s
eo
fo
p
t
i
m
i
z
a
t
i
o
ni
np
r
o
f
e
s
s
i
o
n
a
lp
r
a
c
t
i
c
er
e
q
u
i
r
e
sa
t
q
u
i
s
i
t
e
s
.

b
e
yi
n
c
l
u
d
em
a
t
h
e
m
a
t
i
c
a
lm
o
d
e
l
i
n
go
ft
h
ed
e
s
i
g
np
r
o
b
l
e
m

l
e
a
s
tt
h
r
e
eP
knowledgeo
fc
o
m
p
u
t
e
rprogramminga
n
dknowledgeo
fo
p
t
i
m
i
z
a
t
i
o
nt
e
c
h
n
i
q
u
e
s
.
e
l
a
x r
e
q
u
i
r
e
d
Many s
p
e
c
i
a
l
p
u
r
p
o
s
eo
p
t
i
m
i
z
a
t
i
o
ns
o
f
t
w
a
r
ep
a
c
k
a
g
e
s r
knowledge o
fprogramming a
r
ea
v
a
i
l
a
b
l
et
o
d
a
y
. To u
s
e them e
f
f
i
c
i
e
n
t
l
yt
h
e
r
e
m
a
i
n
i
n
gtwoa
r
e
a
ss
t
i
l
lh
a
v
et
obea
d
d
r
e
s
s
e
d
.
l
e
r
ea
r
es
e
v
e
r
a
le
x
c
e
l
l
e
n
tb
o
o
k
sont
h
es
u
b
j
e
c
to
fo
p
t
i
m
i
z
a
t
i
o
nafewo
fthem
r
t
h
es
u
b
j
e
c
ti
nd
e
p
t
h
.whichi
sn
e
c
e
s
s
a
r
yb
e
c
a
u
s
e
r
e
l
e
a
s
e
dr
e
c
e
n
t
l
y
.Mosto
fthemc
o
v
t
h
m
a
t
h
e
m
a
t
i
c
a
lm
o
d
e
l
sa
r
en
o
n
l
i
n
e
a
ra
n
dr
e
q
u
i
r
es
p
e
c
i
a
lt
e
c
h
n
i
q
u
s
t
h
a
ta
r
eu
s
u
a
l
l
y
notpa
o
fa
n
yc
o
r
ec
u
r
r
i
c
u
l
u
m
.A
l
lo
feb
o
o
k
sassume er
e
a
d
e
ri
sf
a
m
i
l
i
a
rw
i
l
h
aprogrammingl
a
n
g
u
a
g
e
a
d
i
t
i
o
n
a
l
l
yFORTRAN
andr
e
c
e
n
t
l
y
C
.A
n
o
t
h
e
ra
s
s
u
m
p
l
i
o
n
hq
u
e
n
t
l
ymadei
s t
h
eo
p
t
i
m
i
z
a
t
i
o
nc
h
n
i
q
u
e
swi
1
lbeimplemented i
na
mainframec
o
m
p
u
t
i
n
ge
n
v
i
r
o
n
m
e
nl
.S
uchac
o
m
b
i
n
a
t
i
o
ni
sex
e
melyd
i
f
f
i
c
u1
tf
o
rt
h
e
s
e
l
f
I
e
a
r
n
e
re
v
e
nw
i
t
henormousm
o
t
i
v
a
t
i
o
n
.
Ane
x
p
l
o
s
i
o
no
fi
n
e
x
p
e
n
s
i
v
ed
e
s
k
t
o
pc
o
m
p
u
t
i
n
gr
e
s
o
u
r
c
e
sa
i
d
se
n
g
i
n
e
e
r
i
n
gd
e
s
i
g
n
o
d
a
y
.P
a
r
a
l
l
e
l
i
n
gt
h
i
sd
e
v
e
l
o
p
m
e
n
ti
sea
v
a
i
l
a
b
i
l
i
t
yo
fe
x
t
e
n
s
i
v
e s
o
u
p
r
a
c
t
i t
ont
h
eI
n
t
e
r
n
e

tb
o
t
hf
o
rl
e
a
r
n
i
n
ga
n
ddeployment
.Thi
si
sj
o
i
n
e
dbyt
h
ea
v
a
i
l
a
b
i
l
i
t
yo
f
s
y
s
t
e
m
st
h
a
tp
r
o
v
i
d
e o
p
p
o
r
t
u
n
i
t
yf
o
rh
a
n
d
l
i
n
gm
a
t
h
e
m
a
t
i
c
sg
r
a
p
h
i
c
s
.a
n
d
s
o
f
t
1
ye
a
s
yt
o
programmingi
nac
o
n
s
i
s
t
e
n
tm
a
n
n
e
r
.Theses
o
f
t
w
a
r
es
y
s
t
e
m
sa
r
es
i
g
n
i
f
i
c
a
n
m
a
s
t
e
rcompar
t
ot
h
eh
i
g
h
e
r
l
e
v
e
lprogrammingl
a
n
g
u
a
g
e
so
ft
h
ep v
i
o
u
sye
S
.
l
i
sbookp
r
i
m
a
r
i
l
ys
e
e
k
st
oh
a
r
n
e
s
st
h
i
si
a
n
g
u
l
a
t
i
o
no
fs
r
v
i
c
e
st
op
r
o
v
i
d
ea
ebooku
s
e
sMA
T
I
.
A
B@t
o
p
r
a
c
t
i
c
a
la
p
p
r
o
a
c
ht
ot
h
es
t
u
d
yo
fd
e
s
i
g
no
p
t
i
m
i
z
a
t
i
o
n
. l
i
1
1u
s
t
r
a
t
ea
n
di
m
p
l
e
m
e
n
tt
h
ev
a
r
i
o
u
st
e
c
h
n
i
q
u
e
so
fo
p
t
i
m
i
z
a
t
i
o
n
.MATLABi
sap
r
o
d
u
c
t
fromMathWorks.I
nc
.
c
c
o
m
p
l
i
s
htwoi
m
p
o
r
t
a
n
to
b
j
t
i
v
e
s
.Thef
i
r
T
h
i
sbooka
t
t
e
m
p
t
s10a

x
l
v P
R
E
F
A
C
E
t
u
r
e
sofMATLABl
ti
n
t
e
g
r
a
t
e
si
sp
o
w
e
r
f
u
l
o
fs
y
m
b
o
l
i
c
.n
u
m
e
r
i
c
a
l
.a
n
dg
r
a
p
h
i
c
a
ll
f
s
l
a
t
i
o
no
fmany a
l
g
o
r
i
t
h
m
si
n
t
oa
p
p
1
ied numerical
c
o
m
b
i
n
a
t
i
o
nd
u
r
i
n
g n
e
r
ei
s ac
o
n
s
t
a
n
te
n
h
a
n
c
e
m
e
n
to
ft
h
e
t
e
c
h
n
i
q
u
e
sf
o
rd
e
s
i
g
no
p
t
i
m
i
z
a
t
i
o
n
. Th
programmings
k
i
l
l
st
h
r
o
u
g
h
o
u
tt
h
b
o
o
k
.Thes
e
c
o
n
d0
e
c
t
i
v
ei
st
h
p
r
i
m
a
r
yt
a
s
ko
f
b
ook-toc
o
m
m
u
n
i
c
a
t
eandd
e
m
o
n
s
t
r
a
t
ev
a
r
i
o
u
sl
Iu
m
e
r
i
c
a
lt
e
c
h
n
i
q
u
e
st
h
a
ta
r
e
usedi
nt
l
t
ea
r
e
ao
l
o
p
t
i
m
a
ld
e
s
.
c
u
r
r
e
n
l
AHo
ft
h
cn
u
m
e
r
i
c
a
ltechniquesMEsupportedbyMMLABc
o
d
ea
v
a
i
l
a
b
l
e
c
o
m
p
u
t
e
r f
i
l
e
s
. These f
i
l
e
s a
r
e a
v
a
i
l
a
b
l
e on a companion web s
i
t
e
t
.l
t
i
sn
e
c
e
s
s
a
r
yt
ov
i
s
i
tt
h
ewebs
i
t
et
odownloadt
h
cf
i
l
e
st
o
www.wilev.comlvenka
f
o
l
l
owa
l
l
o
ft
heE
Xt
l
I
npi
e
sTh
eboo
kwi
l
l
b
ee
n
t
i
a
l
l
y
i

c
omple
t

.
e
w
i

t
h

u
t

t
h
c
s
cf
i
l
e
s
.
awebcourseon
Att
h
ewebs
i
t
eyouw
i
l
la
l
s
of
i
n
du
s
c
f
u
ll
i
n
k
s1
0o
t
h
e
rr
e
s
o
u
r
c
e
s
MATLABprogramming
a
ndu
p
d
a
t
e
di
n
f
o
r
m
a
t
i
o
na
b
o
u
tA
p
p
l
i
e
dO
p
t
i
n
t
i
m
u
o
nw
i
t
h
MAT
l
.
A
B @P
r
o
g
r
a
m
m
i
n
g
.
Thed
e
c
i
s
i
o
nt
ol
o
c
a
t
et
h
ef
i
l
e
sonawebs
i
t
epmvidEsana
v
e
n
u
et
oshadowt
h
e
dynamicc
h
a
n
g
e
si
nMATLAss
o
f
1
w
a
r
ei
t
s
c
l
f
whichi
sb
c
i
n
gt
r
a
n
s
f
o
r
m
e
di
nam
a
j
o
r
waya
l
m
o
s
te
v
e
r
yy
c
a
r
.
l
tp
r
o
v
i
d
e
samechanism1
0c
o
r
r
e
c
tb
u
g
si
ec
o
d
ei
nat
i
m
e
l
y
m
a
n
n
e
r
.I
te
s
t
a
b

is
h
e
sa
no
p
p
o
r
t
u
n
i
t
yt
oi
m
p
r
o
v
e c
o
d
et
h
r
o
u
g
hr
e
a
d
e
rs
u
g
g
c
s
t
i
o
n
s
u
t
h
o
r1
0k
e
e
pt
h
cr
c
a
d
e
re
n
g
a
g
e
dw
i
t
hnewd
c
v
c
l
o
p
m
e
n
t
s
.
andp
r
o
v
i
d
e
sawayf
o
rea
I
tdocutdownt
h
es
i
z
co
ft
h
ebooke
n
o
r
m
o
u
s
l
ybyl
o
c
a
t
i
n
ga
l
lo
ft
h
ep
a
g
o
fcode

o
u
t
s
i
d
et
h
eb
o
o
k
.
Anu
n
u
s
u
a
lf
e
a
t
u
r
eo
ft
h
ebooki
sci
n
c
l
u
s
i
o
no
fd
i
s
c
r
e
t
eo
p
t
i
m
i
z
a
t
i
o
nandg
l
o
b
a
l
r
a
d
i
l
i
o
n
a
lc
o
n
t
i
n
u
o
u
sd
e
s
i
g
no
p
t
i
m
i
z
a
t
i
o
nt
e
c
h
n
i
q
u
e
sdot
a
k
eupa
o
p
t
i
m
i
z
a
t
i
o
n
.T
s
i
g
n
i
f
i
c
a
n
tp
o
r
t
i
o
no
ft
h
eb
o
o
k
.C
o
n
t
i
n
u
o
u
sp
r
o
b
l
e
m
sg
e
n
e
r
a
l
l
ya
l
l
o
wt
h
ea
l
g
o
r
i
t
h
m
s
t
ob
em
a
t
h
e
m
a
t
i
c
a
l
l
yd
e
v
e
l
o
p
e
di
nac
o
n
v
i
n
c
i
n
gmanncr

a
n
dei
d
e
a
sc
o
u
l
dbe
c
o
r
eo
fo
r
i
g
i
n
a
la
1
g
o
r
i
t
h
m
sd
c
v
c
1
0
p
c
dbyt
h
er
e
a
d
e
r
st
oa
d
d
r
e
s
st
h
e
i
rowns
p
e
c
i
a
l
to
f
t
c
np
r
e
s
c
n
ta
n
e
c
d
s
.E
v
e
r
yc
h
a
p
t
c
ri
n
c
l
u
d
e
sa
d
d
i
t
i
o
n
a
ln
o
n
t
r
i
v
i
a
lc
x
a
m
p
1
e
sa
em
a
t
h
c
m
a
t
i
c
a
l
n
o
v
e
lu
s
eo
re
x
t
c
n
s
i
o
no
ft
h
co
p
t
i
m
i
z
a
t
i
o
nt
e
c
h
n
i
q
u
e
si
nt
h
ec
h
a
p
t
e
r
.Th
a
l
g
o
r
i
t
h
m
s and c
x
a
m
p
l
e
sa
r
a
c
c
o
m
p
a
n
i
e
d by MA'
A
Bc
o
d
ea
v
a
i
l
a
b
l
ea
tt
h
e
companionwebsue-Thenumericalt
e
c
h
n
i
q
u
e
su
s
u
a
l
l
yi
n
c
l
u
d
es
t
r
o
n
gg
r
a
p
h
i
c
a
l
l
1u
s
t
r
a
t
eco
p
c
r
a
t
i
o
no
ft
h
ct
e
c
h
n
i
q
u
e
s
.
s
u
p
p
o
r
tt
oi
D
i
s
c
r
e
t
ea
n
d
g
l
o
b
a
lo
p
t
i
m
i
z
a
t
i
o
nt
e
c
h
n
i
q
u
e
sa
r
ebasedonv
e
r
y
d
i
f
f
e
r
e
n
t
a
l
g
o
r
i
t
h
m
s
and .
c
ho
fthcm c
a
ne
a
s
i
l
yj
u
s
t
i
f
yf
u
l
l
f
l
e
d
g
e
di
n
d
e
p
e
n
d
e
n
tc
o
u
r
s
c
sd
e
t
c
d
e
x
c
l
u
s
i
v
e
l
y1
0t
h
c
i
rS
1
u
d
y
.I
nac
o
m
p
r
c
h
e
n
s
i
v
cbooks
u
c
h

P
R
E
F
A
C
E

XV

c
l
a
s
s
r
o
o
ms
e
t
t
i
n
gw
i
t
h
o
u
1c
o
m
p
u
t
a
t
i
o
n
a
lc
x
p
e
r
i
e
n
c
ei
sa
l
s
of
e
a
s
i
b
1
e
.I
ni
sc
a
s
et
h
e
ep
r
e
s
e
n
t
e
dandt
h
en
u
r
n
e
r
i
c
a
lr
e
s
u
l
t
si
l
l
u
s
t
r
a
t
e
d
.Thes
e
n
i
o
r
/
g
r
a
d
u
a
1
e
a
1
g
o
r
i
t
h
m
sc
a
nb
i
n
e
s
.e
s
p
e
c
i
a
l
l
ye
n
g
i
n
c
e
r
i
n
ga
r
et
h
et
a
r
g
c
1a
u
d
i
e
n
c
e
.
s
1
u
d
e
n
t
si
nv
a
r
i
o
u
sd
i
s
c
i
p
O
p
t
i
m
i
z
a
t
i
o
nt
e
c
h
n
i
q
u
c
sa
r
ea
ni
m
p
o
r
t
a
n
l1
0
0
1t
os
o
l
v
ed
c
s
i
g
np
r
o
b
l
e
m
si
na
l
l
h
ea
r
e
ao
fm
e
c
h
a
n
i
c
a
l
p
r
o
f
e
s
s
i
o
n
a
la
r
e
a
so
fs
t
u
d
y
.Manyi
l
l
u
s
t
r
a
t
i
o
n
sa
r
eomt
e
n
g
i
n
e
e
r
i
n
gr
e
f
l
e
c
t
i
n
g e
x
p
e
r
i
e
n
c
eo
feau
o
r
.
I
n
d
e
p
e
n
d
e
n
tl
e
a
r
n
e
l
'
s

p
a
r
t
i
c
u
Ia
r
l
yp
r
o
f
e
s
s
i
o
n
a
l
swhon
e
e
dt
ou
n
d
e
r
s
t
a
n
dt
h
es
u
b
j
e
c
t
s
h
o
u
l
da
l
s
of
i
n
dt
h
ebookv
e
r
yu
s
e
f
u
.
lA r
e
a
d
e
rwhoh
a
sa
c
c
e
s
st
o
o
fo
p
t
i
m
i
z
a
t
i
o
n
MATLABs
o
f
t
w
a
r
ec

u
s
t
h
ebookt
oi
t
sb
e
s
ta
d
v
a
n
t
a
g
e
.The book i
sl
a
r
g
e
l
y
s
e
l
f
c
o
n
t
a
i
n
e
dandd
e
v
e
l
o
p
sa
l
ln
c
c
c
s
s
a
r
ym
a
t
h
e
m
a
t
i
c
a
lc
o
n
c
e
p
t
s when n
c
e
d
c
d
.
A
b
s
t
r
a
c
tm
a
t
h
e
m
a
t
i
c
a
li
d
c
a
so
fo
p
t
i
m
i
t
i
o
n i
n
t
r
o
d
u
c
c
dg
r
a
p
h
i
c
a
l
l
ya
n
dt
h
r
o
u
g
h
i
I
l
u
s
t
r
a
t
i
v
ee
x
a
m
p
l
e
si
nt
h
cb
o
o
k
.I
nmanyi
n
s
t
a
n
c
e
s
t
h
eo
p
e
r
a
t
i
o
no
ft
h
cn
u
m
c
r
i
c
a
l
t
c
dt
oshowhowt
h
ed
e
s
i
g
ni
schangingw
i
t
hi
t
e
r
a
t
i
o
n
s
.T
h
ish
a
s
t
e
c
h
n
i
q
u
ei
s a
o
r
a
t
c
.
b
e
e
np
o
s
s
i
b
l
eb
e
c
a
u
s
eMATLABg
r
a
p
h
i
c
sa
r
er
i
c
hi
nf
c
a
t
u
r
e
sa
n
ds
i
m
p
l
et
oi
n
c
o
d
c
a
s
.I
ti
sf
o
rt
h
i
sr
e
a
s
o
ner
e
a
d
e
ri
s
makingUlcmc
f
f
c
c
t
i
v
ef
o
rc
o
n
v
c
y
i
n
gei
nc
Iu
d
i
n
g ac
h
a
p
t
e
rong
r
a
p
h
i
c
a
l
c
x
p
o
s
e
dt
og
r
a
p
h
i
c
a
l programming e
a
r
l
y oni
a
l
a
n
c
et
h
cd
e
v
e
l
o
p
m
e
n
to
fprogrammings
k
i
l
l
sa
n
dt
h
p
r
e
s
e
n
t
a
t
i
o
n
o
p
t
i
m
i
z
a
t
i
o
n
.Tob
o
fo
p
t
i
m
i
z
a
t
i
o
nc
o
n
c
e
p
t
sprogramming i
s
s
u
e
sa
r
ei
n
c
r
e
m
e
n
t
a
l
l
ye
x
p
l
o
r
e
da
n
d
e
l
o
p
i
n
gprogrammings
k
i
l
l
si
sem
o
s
t
i
m
p
l
e
m
e
n
t
e
d
.I
n
v
e
s
t
m
e
n
to
ft
i
m
ei
ndv
e
f
f
e
c
t
i
v
ewayt
oi
m
b
i
b
et
h
e
m
.Thca
u
t
h
o
ri
sn
O
la
w
a
r
eo
fs
h
o
r
t
c
u
t
s
.S
t
u
d
c
n
t
sa
r
c
debuge
r
r
o
r
s
.a
n
di
n
c
o
r
p
o
r
a
t
e
e
x
p
e
c
t
e
dt
op
r
o
g
r
a
mi
n
d
e
p
e
n
d
e
n
t
l
y
.makem
i
s
t
a
k
e
s
i
m
p
r
o
v
e
m
e
n
t
sa
sp to
ft
h
e
i
rl
e
a
r
n
i
n
ge
x
p
e
r
i
e
n
c
ed
e
l
i
v
e
r
e
dt
h
r
o
u
g
hi
sb
o
o
k
.
Thebookp
r
o
c
e
e
d
sa
tab
r
i
s
kp
a
c
et
op
r
o
v
i
d
ee
x
p
e
r
i
e
n
c
ci
nMATLA8programming
andcommunicatei
d
e
a
si
no
p
t
i
m
i
z
a
t
i
o
ni
nar
e
a
s
o
n
a
b
l
enumbero
fp
a
g
e
s
.Th
i
sc
o
u
l
d
n
o
tb
ep
o
s
s
i
b
l
ew
i
t
h
o
u
tt
a
k
i
n
ga
d
v
a
n
t
a
g
eo
fI
h
eaccompanyingwebs
i
t
e1
0l
o
c
a
t
emuch
o
ft
h
ec
o
d
ef
o
ra
l
g
o
r
i
t
h
m
s and e
x
a
m
p
l
e
s
. Merely r
u
n
n
i
n
gt
h
ec
o
d
ew
i
l
ln
0
1
c
et
h
er
e
a
d
e
r
'
sprogrammings
k
i
l
l
s
.I
ti
si
m
p
o
r
t
a
n
tf
o
rt
h
es
t
u
d
e
n
t
s
i
g
n
i
f
i
c
a
n
t
l
ye
i
s
ti
ni
sp
r
o
c
e
s
s
t
h
ec
o
d
ec
a
r
r
i
e
sI
ib
e
r
a
lcomments
t
ou
n
d
e
r
s
t
a
n
dt
h
ec
o
d
ea
l
s
o
.To s
o
f
t
e
nc
o
r
r
e
s
p
o
n
d
i
n
gt
ot
h
ea
l
g
o
r
i
t
h
mt
h
a
ti
sb
e
i
n
gi
m
p
l
e
m
c
n
t
e
d
.Thea
uo
re
x
p
e
c
t
s
er
e
a
d
e
rt
oh
a
v
er
e
a
dec
o
m
m
e
n
t
s
.Th
eaccompanyingc
o
d
ea
l
l
o
w
st
h
ei
n
s
t
r
u
c
t
o
r
t
oa
s
s
i
g
nb
o
t
hprogramminga
s
s
i
g
n
m
e
n
t
sa
sw
e
l
la
ss
o

x
v
l P
R
E
F
A
C
E

dllpdLY

a
p
p
r
o
a
c
h
e
s
t
h
ei
n
s
t
r
u
c
t
o
rc
a
nc
h
o
o
s
et
oi
n
c
l
u
d
eo
n
l
y as
u
b
s
e
to
fc
o
n
t
i
n
u
o
u
s
o
p
t
i
m
i
z
a
t
i
o
na
l
g
o
r
i
t
h
m
st
oa
l
l
o
wt
h
ei
n
c
l
u
s
i
o
no
fe
s
ep
o
p
u
l
a
rt
e
c
h
n
i
q
u
e
s
.Ife
p
t
i
m
i
z
a
t
i
o
nT
o
o
l
b
o
xf
r
o
mMATlABe
ni
tc
a
nb
ei
n
t
e
g
r
a
t
e
d
s
t
u
d
e
n
th
a
sa
c
c
e
s
st
ot
h
eO
o
u
r
s
ef
o
rh
a
n
d
l
i
n
gt
h
ef
i
n
a
ld
e
s
i
g
np
r
o
j
e
c
.
tThel
a
s
tc
h
a
p
t
e
rp
r
o
v
i
d
e
sa
n
i
n
t
oec
o
v
e
r
v
i
e
wont
h
eu
s
eo
ft
h
et
o
o
l
b
o
x
.
T
h
isbooks
t
a
r
t
e
do
u
tu
s
i
n
gMA
T
I
.
A
BV
e
r
s
i
o
n5
.Abouth
a
l
f
w
a
yt
h
r
o
u
g
ht
h
ebook
et
r
a
n
s
i
t
i
o
nt
oMATlAB V
e
r
s
i
o
n5.
3wasc
o
m
p
l
e
t
e
.Att
h
ec
o
m
p
l
e
t
i
o
no
ft
h
e
m u
s
c
r
i
p
tMATLABV
e
r
s
i
o
n6wshippingf
o
rsomet
i
m
e
.Aso
ft
h
i
st
i
r
n
e
Version
6
.
1i
sa
v
a
i
l
a
b
l
e
.Mosti
n
s
t
i
t
u
t
i
o
n
a
lMATlABl
i
c
e
n
s
e
sa
r
eu
s
u
a
l
l
yb
a
s
e
do
ns
u
b
s
c
r
i
p
t
i
o
n
h
e
yw
i
l
lh
a
v
e l
a
t
e
s
tv
e
r
s
i
o
no
fes
o
f
t
w
a
r
e
.I
nas
e
n
s
e

a
n
di
si
st
r
u
eo
f
s
o t
a
l
lb
o
o
k
st
h
a
td
e
p
e
n
donp
a
r
t
i
c
u
l
a
rs
o
f
t
w
a
r
es
y
s
t
e
m
s
t
h
ebookr
n
a
ya
p
p
e
a
ro
u
to
fd
a
t
e
o
f
t
w
a
r
e
.T
h
i
si
sn
o
tr
e
a
l
l
ys
o
.Whilet
h
eu
s
a
g
eofMATLABf
u
n
c
t
i
o
n
s
w
i
t
hr
e
s
p
e
c
tt
oes
r
n
a
yc
h
a
n
g
eb
e
t
w
e
e
nv
e
r
s
i
o
n
s
t
h
e
r
es
h
o
u
l
dbes
u
i
t
a
b
l
ewamingsandi
n
s
t
r
u
c
t
i
o
na
b
o
u
t
ed
i
f
f
e
r
e
n
tu
s
a
g
e
Ani
m
p
o
r
t
a
n
ti
n
s
u
r
a
n
c
ei
sb
u
i
l
ti
n
t
ot
h
ebook
p
a
r
t
i
c
u
l
a
r
l
yt
o
h
a
n
d
l
es
u
c
hc
h
a
n
g
e
si
ft
h
e
ya
r
i
s
e
.S
p
r
i
n
k
l
e
dt
h
r
o
u
g
h
o
u
tt
h
ebooki
st
h
eh
a
b
i
to
fe
u
s
i
n
geo
n
l
i
n
eh
e
l
p
.I
nmosti
n
s
t
a
n
c
e
si
swasp
r
i
m
a
r
i
l
ydonet
oe
x
p
o
s
et
h
er
e
a
d
e
r
t
oa
l
t
e
r
n
a
t
i
v
eu
s
eo
ft
h
esamecommand.A
n
o
t
h
e
rs
i
g
n
i
f
i
c
a
n
tc
h
a
n
g
et
h
em
a
n
u
s
c
r
i
p
t
h
a
dt
oe
n
d
u
r
ewas el
o
s
so
fa
c
c
e
s
st
oaUNIXv
e
r
s
i
o
nb
e
c
a
u
s
eo
ft
h
u
n
i
v
e
r
s
i
t
y
'
s
e
c
h
a
n
g
et
od
i
s
t
r
i
b
u
t
e
dc
o
m
p
u
t
i
n
gonPCc
l
u
s
t
e
r
s
.Thea
u
t
h
o
r
'
se
x
p
e
r
i
e
n
c
ew
i
e
a
r
l
yc
h
a
p
t
e
r
ss
u
g
g
s
1
8t
h
a
tt
h
i
ss
h
o
u
l
dn
o
tma
t
1
e
r
.E
x
c
e
p
tf
o
re
x
t
e
r
n
a
lf
i
l
e
sa
n
d
u
s
eo
fMA
TlA
Ba
si
l
l
u
s
t
r
a
t
e
di
nt
h
ebooki
si
n
d
i
f
f
e
r
e
n
tt
ot
h
ev
a
r
i
o
u
s
r
e
s
o
u
r
c
e
s
o
p
e
r
a
t
i
n
gs
y
(
WindowsMacUNIXo
rL
i
n
u
x
)
.
Thet
o
p
i
c
sc
h
o
s
e
nf
o
ri
n
c
l
u
s
i
o
nandt
h
es
i
m
p
l
i
c
i
t
yo
fp
r
e
s
e
n
t
a
t
i
o
no
fet
o
p
i
c
sa
r
e
d
i
r
e
c
t
l
yr
e
l
a
t
e
dt
ot
h
ee
x
p
e
r
i
e
n
c
eo
ft
e
a
c
h
i
n
gt
h
ec
o
u
r
s
eono
p
t
i
m
i
z
a
t
i
o
na
tt
h
e
p
a
r
t
m
e
n
to
fM
e
c
h
a
n
i
c
a
lE
n
g
i
n
e
e
r
i
n
gf
o
ro
v
e
rt
e
ny
e
a
r
s
.
s
e
n
i
o
r
/
g
r
a
d
u
a
t
el
e
v
e
li
neD
tc
o
m
p
r
e
h
e
n
s
i
o
ni
r
n
p
r
o
v
e
dw
i
t
hs
i
m
p
l
ei
I
Iu
s
t
r
a
t
i
o
n
s

however
E
x
p
e
r
i
e
n
c
ep
r
o
v
e
d a

s
t
r
a
i
g
h
o
r
w
a
r
dm
a
t
h
e
r
n
a
t
i
c
a
l
c
o
m
p
l
i
c
a
t
e
dt
h
em
a
t
h
e
m
a
t
i
c
a
li
d
e
a
s
.Ont
h
eo
t
h
e
rh
a
n
d
e
x
p
r
e
s
s
i
o
n
se
l
i
c
i
t
e
dnoi
n
t
e
r
e
s
to
ru
n
d
e
r
s
t
a
n
d
i
n
g
.Thee
m
p
h
a
s
i
sona
p
p
l
i
c
a
t
i
o
nwas
i
m
p
o
r
t
a
n
tt
ok
e
e
pt
h
es
t
u
d
e
n
t
s
'a
t
t
e
n
t
i
o
n
.I
nt
h
i
sc
o
n
n
e
c
t
i
o
nt
h
ea
u
t
h
o
rwouldl
i
k
et
o
1ka
l
lh
i
ss
t
u
d
e
n
1
8f
o
rp
u
s
h
i
n
gh
i
r
nf
o
rs
i
m
p
l
ee
x
p
l
a
n
a
t
i
o
n
s
f
o
rp
r
o
v
i
d
i
n
gp
o
s
i
t
i
v
e
i
n
t
e
r
a
c
t
i
o
n

andf
o
rt
h
e
i
rc
o
n
t
r
i
b
u
t
i
o
no
fo
r
i
g
i
n
a
la
n
du
s
e
f
u
li
d
e
a
si
nt
h
ed
i
s
c
u
s
s

P
R
E
F
A
C
E x
v
i
i
Thebookwasmadep
o
s
s
i
b
l
et
h
r
o
u
g
hs
u
p
p
o
r
tfromJ
o
h
nW
i
l
e
y dSonsI
n
c
. d
n
c
.S
i
n
c
e
r
et
h
a
n
k
sa
r
eowedt
oBobA
r
g
e
n
t
i
e
r
i
s
e
n
i
o
re
d
i
t
o
ra
tJ
o
b
n
MathWorksI
Wileyf
o
ra
c
c
e
p
t
i
n
gt
h
ep
r
o
p
o
s
a
l
a
n
dwhoa
l
la
l
o
n
gd
i
s
p
l
a
y
e
dal
o
to
fp
a
t
i
e
n
c
e
i
n
g
e
t
t
i
n
gt
h
ebookmovingforward.Samei
sduet
oBobH
i
l
b
e
d
a
s
s
o
c
i
a
t
em
a
n
a
g
i
n
g
f
o
rh
i
simp
s
s
i
v
eworka
tc
1
e
a
n
i
n
gu
pt
h
em
a
n
u
s
c
r
i
p
.
tB
r
i
a
n
e
d
i
t
o
ra
tJohnWiley
SnappNew Media e
d
i
t
o
ra
tJ
o
h
n Wileyc
r
e
a
t
e
d companion web s
i
t
e
l
s
ob
er
n
a
i
n
t
a
i
n
i
n
gi
t
.NaomiFemandesom
(
w
w
w
.
w
i
l
e
y
.
c
o
m
l
v
e
n
k
a
t
)a
n
dw a
I
n
c
.
sawt
oi
ta
t1h
a
dt
h
el
a
t
e
s
tv
e
r
s
i
o
no
fMATLABa
ss
o
o
na
si
twas
MathWorks

P
r
o
f
e
s
s
o
rE
m
e
r
i
t
u
s
a
tR
ic
eU
n
i
v
e
r
s
i
t
yi
s
a
v
a
i
l
a
b
l
e
.Myr
e
g
a
r
df
o
rD
r
.A
n
g
e
l
oM
i
e
l
e
moret
h
a
nc
a
nb
ee
x
p
r
e
s
s
e
di
nt
h
e
s
eI
i
n
e
s
.I
twash
ewhoi
n
t
r
o
d
u
c
e
dmet
o1
h
s
u
b
j
e
c
l
o
fo
p
t
i
m
i
z
a
t
i
o
nandd
e
m
o
n
s
t
r
a
t
e
dt
h
ee
f
f
e
c
t
i
v
e
n
e
s
so
fs
i
m
p
l
ep
r
e
s
e
n
u
t
i
o
n
Iw
i
l
l

myf
a
m
i
l
yd
e
s
e
r
v
e
ss
p
e
c
i
a
lr
n
e
n
t
i
o
n
a
l
w
a
y
sr
e
g
a
r
dhima
sag
r
e
a
tt
e
a
c
h
e
r
.Ofc
o
u
r
s
e
i
n
2
f
o
rp
u
t
t
i
n
gupw
i
t
ha
l
lt
h
e
Notnow
"
L
a
t
e
r

"
Howa
b
o
u
tt
o
r
n
o
r
r
o
w
?
"d
d
e
b
u
g
g
i
n
gec
o
d
e
.Sp
i
a
l k
sa
r
et
oA a
n
a dV
i
n
a
y
a
k

r
n
yo
f
f
s
p
r
i
n
g

i
t

t
h
e
i
rp
a
t
i
e
n
c
e
u
n
d
e
r
s
t
a
n
d
i
n
g
andencouragement
.Thea
u
o
ra
p
o
l
o
g
i
z
e
sf
o
ra
n
y
c
r
i
t
i
c
i
s
m
s
and
s
h
o
r
t
c
o
r
n
i
n
g
so
nt
h
ep
r
e
s
e
n
t
a
t
i
o
n and welcomes comments
s
u
g
g
e
s
t
i
o
n
sf
o
ri
m
p
r
o
v
e
m
e
n
ta
ta
l
lt
i
m
e
s
.

R
o
c
h
e
s
t
e
NewY
o
r
k

P
.VENKATARAMhN

a. .

tJFJJ-

p
d
r
h
wAbrr=353thtY4y

INTRODUCTION

O
p
t
i
m
i
z
a
t
i
o
nh
a
s become a n
e
c
e
s
s
a
r
yp
a
r
to
fd
e
s
i
g
na
c
t
i
v
i
t
yi
na
l
lm
a
j
o
r
o
t
i
v
a
t
i
o
nt
o
d
i
s
c
i
p
l
i
n
e
s
.Thesed
i
s
c
i
p
l
i
n
e
sa
r
en
o
tr
e
s
t
r
i
c
t
e
dt
oe
n
g
i
n
e
e
r
i
n
g
.Them
p
r
o
d
u
c
ee
c
o
n
o
m
i
c
a
l
l
yr
e
l
e
v
a
n
tp
r
o
d
u
c
t
so
rs
e
r
v
i
c
e
sw
i
t
hembeddedq
u
a
l
i
t
yi
st
h
e
p
r
i
n
c
i
p
a
lr
e
a
s
o
nf
o
ri
si
n
c
l
u
s
i
o
n
.Improvedp
r
o
d
u
c
t
i
o
na
n
dd
e
s
i
g
nt
o
o
l
sw
i a
s
y
n
e
r
g
i
s
t
i
ct
h
r
u
s
tt
h
r
o
u
g
hi
n
e
x
p
e
n
s
i
v
ec
o
m
p
u
t
a
t
i
o
n
a
lr
e
s
o
u
r
c
e
sh
a
v
ea
i
d
e
dt
h
e
c
o
n
s
i
d
e
r
a
t
i
o
no
fo
p
t
i
m
i
z
a
t
i
o
n methods i
n new d
e
v
e
l
o
p
m
e
n
t
sp
a
r
t
i
c
u
l
a
r
l
y
e
n
g
i
n
e
e
r
i
n
gp
r
o
d
u
c
t
s
.Eveni
nt
h
ea
b
s
e
n
c
co
fat
a
n
g
i
b
l
ep
r
o
d
u
c
to
p
t
i
m
i
z
a
t
i
o
n
i
d
e
a
sp
r
o
v
i
d
et
h
ea
b
i
l
i
t
yt
od
e
f
i
n
ea
n
de
x
p
l
o
r
ep
r
o
b
l
e
m
sw
h
i
l
ef
o
c
u
s
i
n
gon
s
o
l
u
t
i
o
n
s a
ts
u
b
s
c
r
i
b
et
osomem
e
a
s
u
r
eo
fu
s
e
f
u
l
n
e
s
s
.G
e
n
e
r
a
l
l
y
t
h
eu
s
eo
ft
h
e
wordo
p
t
i
m
i
z
a
t
i
o
ni
m
p
l
ist
h
cb
e
s
tr
e
s
u
l
tu
n
d
e
rt
h
c
i
r
c
u
m
s
t
a
n
c
e
s
.T
h
i
si
n
c
l
u
d
e
s
t
h
ep
a
r
t
i
c
u
l
a
rs
e
to
fc
o
n
s
t
r
a
i
n
t
sont
h
ed
e
v
e
l
o
p
m
e
n
tr
c
s
o
u
r
c
e
s
c
u
r
r
e
n
tknowledge
m
a
r
k
e
tc
o
n
d
i
t
i
o
n
s

a
n
ds
oo
n
.Everyoneo
fu
sh
a
sp
r
o
b
a
b
l
yu
s
e
dt
h
ct
e
r
ma
tsome
t
i
m
et
od
e
s
c
r
i
b
e ep
r
i
m
a
r
yq
u
a
l
i
t
yo
fo
u
rworko
re
n
d
e
a
v
o
r
.I
ti
sp
r
o
b
a
b
l
ye
m
o
s
lu
s
e
do
ra
b
u
s
e
dl
e
r
mi
na
d
v
e
r
t
i
s
i
n
gandp
r
e
s
e
n
t
a
t
i
o
n
s
.N
e
v
e
r
t
h
c
l
e
s
sc
a
b
i
l
i
t
yt
omaket
h
eb
e
s
tc
h
o
i
c
ei
sap
e
r
p
e
l
u
a
ld
e
s
i
r
eamongu
sa
ll
.
O
p
t
i
m
i
z
a
t
i
o
ni
sf
r
e
q
u
e
n
t
l
ya
s
s
o
c
i
a
t
e
dw
i
t
hd
e
s
i
g
n
b
ei
tap
r
o
d
u
c

ts
e
r
v
i
c
e
o
r
s
t
r
a
t
e
g
y
.A
e
r
o
s
p
a
c
ed
e
s
i
g
nwasamong e
a
r
l
i
e
s
td
i
s
c
i
p
l
i
n
e
st
oe
m
b
r
a
c
eo
p
t
i
m
i
z
a
t
i
o
n
i
nas
i
g
n
i
f
i
c
a
n
twayd
r
i
v
e
nb
yan
a
t
u
r
a
1n
e
e
dt
ol
o
w
e
rt
h
et
r
e
m
e
n
d
o
u
sc
o
s
ta
s
s
o
c
i
a
t
e
d
w
i
t
hc
a
r
r
y
i
n
gu
n
n
e
c
e
s
s
yw
e
i
g
h
ti
na
e
r
o
s
p
a
c
ev
e
h
i
c
l
e
s
.Minimummasss
t
r
u
c
t
u
r
e
sa
r
l

n
o
r
m
.O
p
t
i
m
i
z
a
t
i
o
nf
o
r
m
sp
a
r
to
ft
h
ep
s
y
c
h
eo
fe
v
e
r
ya
e
r
o
s
p
a
c
ed
e
s
i
g
n
e
r
.S
a
v
i
n
g
onf
u
e
lt
h
r
o
u
g
ht
r
j
e
c
t
o
r
yd
e
s
i
g
nwasa
n
o
t
h
e
rp
r
o
b
l
e
mt
h
a
ts
u
g
g
e
s
t
e
di
t
s
e
l
f
.V
e
r
ys
o
o
n
e
n
t
i e
n
g
i
n
e
e
r
i
n
gcommunityc
o
u
l
dr
e
c
o
g
n
i
z
ecn
e
e
dt
od
e
f
i
n
es
o
l
u
t
i
o
n
sb
a
s
e
d
onm
e
r
it
.R
e
c
o
g
n
i
z
i
n
gt
h
ed
e
s
i
r
ef
o
ro
p
t
i
m
i
z
a
t
i
o
na
n
da
c
t
u
a
1l
yi
m
p
l
e
m
e
n
t
i
n
gwe
t
wo
d
i
f
f
e
r
e
n
ti
s
s
u
e
s
.

1
.
1O
P
T
I
M
I
Z
A
T
I
O
NF
U
N
D
A
M
E
N
T
A
L
S 3

I
N
T
R
O
D
U
C
T
I
O
N

U
n
t
i
lr
e
c
e
n
t
l
y
.f
o
rmuchoft
h
et
i
m
e
.o
p
t
i
m
i
z
a
t
i
o
nwasu
s
u
a
l
l
ya
t
t
e
m
p
t
e
do
n
l
yi
n
t
h
o
s
es
i
t
u
a
t
i
o
n
swheret
h
e
r
eweres
i
g
n
i
f
i
c
a
n
tp
e
n
a
l
t
i
e
sf
o
rg
e
n
e
r
i
cd
e
s
i
g
n
s
.Th
e
a
p
p
l
i
c
a
t
i
o
no
fo
p
t
i
m
i t
i
o
ndemandedl
a
r
g
ec
o
m
p
u
t
a
t
i
o
n
a
lr
e
s
o
u
r
c
e
s
.I
nt
h
en
a
s
c
e
n
t
y
e
a
r
so
fd
i
g
i
t
a
lc
o
m
p
u
t
a
t
i
o
nt
h
e
s
ewerea
v
a
i
l
a
b
l
eo
n
l
yt
ol
a
r
g
en
a
t
i
o
n
a
ll
a
b
o
r
a
t
o
r
i
e
s
andr
e
s
e
a
r
c
hp
r
o
g
r
a
m
s
.Theser
e
s
o
u
r
c
e
sweren
e
c
e
s
s
a
r
yt
oh
a
n
d
l
et
h
n
o
n
l
i
n
e
a
r
p
r
o
b
l
e
m
sa
t a
s
s
o
c
i
a
t
e
dw
i
t
he
n
g
i
n
e
e
r
i
n
go
p
t
i
m
i
.
t
i
o
n
.Asar
e
s
u
l
toft
h
e
s
e
c
o
n
s
t
r
a
i
n
t
s most of t
h
ee
v
e
r
y
d
a
yp
r
o
d
u
c
t
s were d
e
s
i
g
n
e
dw
i
t
h
o
u
tr
e
g
a
r
dt
o
o
p
t
i
m
i
z
a
t
i
o
n
.Th
i
si
n
c
l
u
d
e
se
v
e
r
y
t
h
i
n
gyous
o
undyouo
ru
s
ei
nyourd
a
i
l
yl
i
f
e
.I
t
i
si
n
c
o
n
c
e
i
v
a
b
l
ea
tt
h
enewg
e
n
e
r
a
t
i
o
no
fr
e
p
l
a
c
e
m
e
n
tp
r
o
d
u
c
t
s
.l
i
k
ct
h
ec
a
r
.t
h
e
e
n
c
i
l
.r
e
d
e
s
i
g
n
e
da
n
dm
a
n
u
f
a
c
t
u
r
e
dt
o
d
a
ya
r
cn
o
td
e
s
i
g
n
e
d
h
o
u
s
e
.t
h
ed
e
s
k
.o
rep
o
p
t
i
m
a
l
l
yi
nones
e
n
s
eo
ra
n
o
t
h
c
r
.
Today.youwouldd
e
f
m
i
t
I
ye
x
p
l
o
r
ep
r
o
c
e
d
u
r
e
st
oo
p
t
i
m
i
z
ey
o
u
ri
n
v
e
s
l
e
n
t
sby
t
a
i
l
o
r
i
n
g your p
o
r
t
f
o
l
i
o
. You would o
p
t
i
m
i
z
ey
o
u
rb
u
s
i
n
e
s
st
r
a
v
e
lt
i
m
e by
a
p
p
r
o
p
r
i
a
I
yc
h
o
o
s
i
n
gyourd
c
s
t
i
n
a
t
i
o
n
s
.Youc
a
no
p
t
i
m
i
z
eyourcommutingt
i
m
b
y
c
h
o
o
s
i
n
gy
o
u
rt
i
m
eandr
o
u
t
e
.Youc
a
no
p
t
i
m
i
z
ey
o
u
rn
e
c
e
s
s
a
r
ye
x
p
e
n
d
i
t
u
r
ef
o
rI
iving
byc
h
o
o
s
i
n
gy
o
u
rdaya
n
ds
t
o
r
ef
o
rs
h
o
p
p
i
n
g
.Youc
a
no
p
t
i
m
i
z
ct
h
eu
s
e
f
u
lt
i
m
eyou
c
o
n
n
e
c
t
t
o I
n
t
e
m
e
tbyd
e
t
c
r
m
i
n
i
n
gy
o
u
rt
i
m
eo
fc
o
n
n
e
c
t
i
o
n
.Youc
a
nbuys
o
f
t
w
a
r
e
w
i
l
lo
p
t
i
m
i
z
ey
o
u
rc
o
n
n
c
c
t
i
o
nt
ot
h
cl
n
t
e
m
c
t
.Youc
a
nbuybookso
rr
e
a
da
r
t
i
c
l
e
s
t
c
l
lyou howt
op
e
r
f
o
r
mt
h
e
s
ev
a
r
i
o
u
so
p
t
i
m
i
z
a
t
i
o
n
s
.Theabovea
c
t
i
v
i
t
i
c
s
te
v
e
r
ya
c
t
i
v
i
t
y
e
x
c
e
p
t
p
r
i
m
a
r
i
l
yr
e
l
a
t
ct
o:
;
c
r
v
I
c
e
so
rs
t
r
a
t
e
g
y
.I
ti
snowa
p
p
a
r
c
n
t a
S
e
t
i
c
p
r
o
v
i
d
e
st
h
es
c
o
p
ef
o
ro
p
t
i
m
i
z
a
t
i
o
n
.l
i
sj
u
s
t
i
f
i
e
sl
o
o
k
i
n
ga
t es
t
u
d
yo
f
a
d
o
o
t
i
m
i
z
o
na
sat
o
o
l c
a
nbea
p
p
l
i
e
d1
0av
a
r
i
e
t
yo
fd
i
s
c
i
p
l
i
n
e
s
.I
fs
o
t
h
emyri

m
a
t
h
e
m
a
t
i
c
a
lmode.
lbedi .i

c
h
a
r
a
c
t
c
r
i
z
a
t
i
o
n
. di
t
sc
i
r
c
u
m
s
t
a
n
c
e
smustbe
e
x
p
r
e
s
s
e
dm
a
t
h
e
m
a
t
i
c
a
l
l
y
.C
o
n
s
i
d
e
rt
h
ed
e
s
i
g
na
c
t
i
v
i
t
yi
nt
h
ef
o
l
l
o
w
i
n
gc
a
s
e
s
:
Newconsumcrr
e
s
e
a
r
c
hi
n
d
i
c
a
t

t
h
a
tp
e
o
p
l
el
i
k
e1
0d
r
i
n
ka
b
o
u
t0
.
5l
i
t
c
ro
fs
o
d
a
ef
a
b
r
i
c
a
t
i
o
nC
O
S
Io
ft
h
er
e
d
e
s
i
g
n
e
d
popa
lat
i
m
ed
u
r
i
n
gt
h
esummerm
o
n
t
h
s
.Th
u
r
f
a
c
ea
r
e
a
.a
n
dc
a
nbee
s
t
i
m
a
t
c
da
t$
1
.
0
0p
c
r
s
o
d
ac
a
ni
sp
r
o
p
o
r
t
i
o
n
a
lt
o s
s
q
u
a
r
em
e
t
e
ro
ft
h
em
a
t
e
r
i
a
Iu
s
e
d
.Ac
i
r
c
u
l
a
rc
r
o
s
ss
e
c
t
i
o
ni
st
h
emostp
l
a
u
s
i
b
l
e
g
i
v
e
nc
u
r
r
e
n
tt
o
o
l
i
n
ga
v
a
i
l
a
b
l
ef
o
rm
a
n
u
f
a
c
t
u
r
e
.Fora
e
s
t
h
e
t
i
cr
e
a
s
o
n
se
th
o
l
d
i
n
gc
o
m
f
o
r
t
h
e
i
g
h
tmustbea
tl
e
a
s
tt
w
i
c
et
h
ed
i
a
m
e
t
c
r
.S
t
u
d
i
e
si
n
d
i
c
a
t
ea
r
e
q
u
i
r
e
sad
i
a
m
e
t
e
rb
e
t
w
e
e
n6and9cm.

;:.~

"
.

"

dt
oc
a
r
r
yap
o
i
n
tl
o
a
dFa
tt
h
ecndo
fa
Ac
a
n
t
i
l
e
v
e
rbeamn
c
e
d
st
obed
c
s
i
g
n
beamofl
e
n
g
t
hL
.Thec
r
o
s
ss
e
c
t
i
o
no
ft
h
ebeamwi
I
Ibei
nt
h
es
h
u
p
eo
ft
h
l
t
t
e
r
edas I
b
m
)
.Thebeams
h
o
u
l
dmeetp
r
e
s
c
r
i
b
e
df
a
i
l
u
r
ec
r
i
t
e
r
i
a
.
I(
e
r
ei
sa
l
s
oal
i
m
i
tonide
f
I
t
i
o
n
.A beamo
fminimummassi
sr
e
q
u
i
r
t
o
bed
e
s
i
g
n
e
d
.

MyPCCompanyh
a
sd
e
c
i
d
c
dt
oi
n
v
c
s
t$12m
i
l
l
i
o
ni
na
c
q
u
i
r
i
n
gs
e
v
e
r
a
ln
W
Componenl Placem
n
l Machines t
o m
a
n
u
f
a
c
t
u
r
e d
i
f
f
e
r
c
n
t k
i
n
d
s o
f
m
o
t
h
c
r
b
o
a
r
d
sf
o
rancwg
c
n
c
r
a
t
i
o
no
fp
e
r
s
o
n
a
lc
o
m
p
u
l
e
r
s
.Threemodelso
f
t
h
c
s
cm
a
c
h
i
n
e
sa
r
eu
n
d
e
rc
o
n
s
i
d
e
r
a
t
i
o
n
.T
o
t
a
lnumbcro
fo
p
e
r
a
t
o
r
sa
v
a
i
l
a
b
l
ei
s
tA f
Ioorspaccc
o
n
s
t
r
a
i
n
tn
e
e
d
s1
0b
c
1
0
0b
e
c
a
u
s
co
ft
h
el
o
c
a
ll
a
b
o
rmarke.
s
a
t
i
s
f
i
e
db
e
c
a
u
s
eo
ft
h
ed
i
f
f
c
r
e
n
td
i
m
c
n
s
i
o
n
so
ft
h
c
s
em
a
c
h
i
n
e
s
.A
d
d
i
t
i
o
n
a
l
i
n
f
o
r
m
a
t
i
o
nr
c
l
a
t
i
n
gt
oe
a
c
hoft
h
emachinesi
sg
i
v
e
ni
nT
a
b
l
e1
.
1
.Th
ecompany
w
i
s
h
e
st
od
e
t
e
r
m
i
n
ehowmanyofe
a
c
hk
i
n
di
sa
p
p
r
o
p
r
i
a
t
e1
0maximizet
h
e
numbero
fb
o
a
r
d
sm
a
n
u
f
a
c
t
u
r
e
dp
e
rd
a
y
.

of
o

t
i

m
i

z
a

t
i

on
p

r
o

bl

m
n

1
1
I
Th
i
s
i

seemph
a
s
i

so
f
t

heb
o
o
k
.
The p
a
r
t
n
e
r
s
h
i
pb
e
t
w
e
e
n ds
i
g
nand o
p
t
i
m
i
z
a
t
i
o
na
c
t
i
v
i
t
yi
so
f
t
e
nf
o
u
n
di
n
e
n
g
i
n
e
e
r
i
n
g
.T
h
i
sbookr
e
c
o
g
n
i
z
e
st
h
a
tc
o
n
n
e
c
t
i
o
nandmanyo
ft
h
cp
r
o
b
l
e
m
su
s
e
df
o
r
i
l
l
u
s
t
r
a
t
i
o
n
sandp
r
a
c
t
i
c
ea
r
efrome
n
g
i
n
r
i
n
gprima
r
i
I
ym
e
c
h
a
n
i
c
a
lc
i
v
i
land
a
e
r
o
s
p
a
c
ed
e
s
i
g
n
.N
c
v
e
r
t
h
e
l
e
s
st
h
es
t
u
d
y ofo
p
t
i
m
i
z
a
t
i
o
n
.p
a
r
t
i
c
u
l
a
r
l
ya
p
p
l
i
e
d
o
p
t
i
m
i
z
a
t
i
o
n
.
i
sn
o
tm e
x
c
l
u
s
i
v
epmpertyofmys
p
t
c
i
n
cd
i
s
c
i
p
u
r
l
e
.
I
ti
n
v
o
l
v
e
st
h
e
dECOVEryanddesignof u
t
i
o
n
st
h
r
o
u
g
ha
p
p
r
o
p
r
i
a
t
et
e
c
h
n
i
q
u
e
s o
c
i
a
t
e
dw
i
t
ht
h
e
f
o
r
m
u
l
a
t
i
o
no
ft
h
eproblemi
nas
p
i
f
i
cm
a
n
n
e
r
.T
h
i
sc
a
nbedonef
o
rexamplei
n
c
h
e
m
i
s
t
r
y
andb
u
s
i
n
e
s
sman
economics

Theabovel
i
s
tr
c
p
r
e
n
t
st
h
r
e
ep
r
o
b
l
e
m
sa
twi
I
Ibeused1
0d
e
f
i
n
cf
o
r
m
a
le
l
e
m
e
n
t
s
o
fano
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
e
m
.Eac
hproblem q
u
i
r
e
si
n
f
o
r
m
a
t
i
o
nfromt
h
es
p
e
c
i
f
i
ca
r
c
a
o
rd
i
s
c
i
p
l
i
n
et
owhichi
tr
e
f
e
r
s
.Tor
c
c
o
g
n
i
z
eo
rd
e
s
i
g
nt
h
e
s
ep
r
o
b
l
e
m
sassumesa
t
d
e
s
i
g
n
c
ri
sc
o
n
v
e
r
s
a
n
twi
ep
a
r
t
i
c
u
l
a
rs
u
b
j tm
a
t
t
e
r
. l
ep
r
o
b
l
e
m
sa
r
ek
e
p
t
s
i
m
p
l
et
of
o
c
u
sono
p
t
i
m
i
z
a
t
i
o
ni
s
s
u
c
s
.Problemso
fs
u
c
hv
a
r
i
e
l
ya
r
ei
m
p
o
r
t
a
n
tt
o
d
a
y
.
1
I
u
s
t
r
a
t
e a
tAlcoa(
a
n
R
e
c
e
n
ta
d
v
e
r
t
i
s
e
m
e
n
t
si
ng
e
n
e
r
a
lconsumerm
a
g
a
z
i
n
e
si
f
a
c
l
u
r
e
r
)i
sq
u
i
t
h
appyt
oh
a
v
er
c
d
u
c
e
d
Lh
cw
e
i
g
h
to
ft
h
es
t
a
n
d
a
r
d
aluminumm u
1aimbyt
h
ep
l
a
s
t
i
c
si
n
d
u
s
t
r
yw
i
t
h
s
o
d
apopc
a
nbyo
v
e
r30%i
nr
e
c
e
n
ly
e
a
r
s
.As
i
m
i
l
a
rc
I
l
o
n
)i
sa
l
5
0e
v
i
d
e
n
ti
nt
h
e
s
em
a
g
a
z
i
n
e
s
a
l
t
h
o
u
g
h
r
e
s
p
e
c
tt
ot
h
es
t
a
n
d
a
r
dm
i
l
kj
u
g(
g
a
sc
a
s
et
h
enumberi
s409

.Roofc
o
U
a
p
s
e
si
n cN
o
r
t
h
e
a
s
td
u
e1
0e
x
c
e
s
s
i
v
esnow
i
ni

1
.
1 OPTIMIZATIONFUNDAMENTALS
O
p
t
i
m
i
z
a
t
i
o
nc
a
nbea
p
p
l
i
e
dt
oa
l
ld
i
s
c
i
p
l
i
n
c
s
.Q
u
a
l
i
t
a
t
i
v
e
l
yt
h
i
sa
s
s
e
r
t
i
o
ni
m
p
l
i
e
s
m
u
l
t
i
p
l
ed
e
c
i
s
i
o
nc
h
o
i
c
e
s
;i
m
p
l
i
c
i
t
l
yr
e
c
o
g
n
i
z
i
n
gt
h
en
c
e
s
s
i
t
yo
fc
h
o
o
s
i
n
gamong
a
1t
e
m
a
t
i
v
e
s
.Th
i
sb
o
o
k
sd
e
a
l
sw
i
t
ho
p
t
i
r
n
i
z
a
t
i
o
ni
naq
u
a
n
t
i
t
a
t
i
v
ew
a
y
.T
h
i
smeanst
h
a
t
anoutcomeo
fa
p
p
l
y
i
n
go
p
t
i
m
i
z
a
t
i
o
nt
oeproblemd
e
s
i
g
n
.o
rs
e
r
v
i
c
emusty
i
e
l
d
number
a
tw
i
l
ld
e
f
i
n
et
h
es
o
l
u
t
i
o
n
.o
ri
no
t
h
e
rwords

numberso
rv
a
l
u
e
s w
i
l
l
c
h
a
r
a
c
t
e
r
i
z
et
h
ep
a
r
t
i
c
u
l
a
rd
e
s
i
g
no
rs
e
r
v
i
c
e
.Q
u
a
n
t
i
t
a
t
i
v
ed
c
s
c
r
i
p
t
i
o
no
f l
u
t
i
o
n
r
e
q
u
i
r
e
saq
u
a
n
t
i
t
a
t
i
v
ed
e
s
c
r
i
p
t
i
o
no
ft
h
eproblemi
t
s
e
l
f
.Th
i
sd
e
s
c
r
i
p
t
i
o
ni
sc
a
l
l
e
da

T
a
b
l
e1
.
1 ComponentP
l
a
c
e
m
e
n
tM
a
c
h
i
r

M
a
c
h
i
n
e
M
o
d
e
l
A
B
C

O
p
e
r
a
b
l
c
I
Ma
c
h
i
n
e
o
st
r
s
I
Da
y C
a
r
d
s
l
H
o
u
r Opemto
r
s
/
S
h
i
f
t Hou
B
o
a
r
d
T
y
p
e
s Bo
1
0
20
1
8

5
5
5
0
5
0

2
2

1
8
1
8
2
1

0
0
0
4
0
0
0
0
0
6
0
0
7
0
0
0
0
0

r
:
:
.
:

'
"

INTRODUCTION

d
u
r
i
n
gt
h
e 1998-1999 w
i
n
t
e
rw
i
l
lh
a
v
es
t
r
u
c
t
u
r
a
1d
e
s
i
g
n
e
r
sa
n
dh
o
m
e
b
u
i
l
d
e
r
s
e
x
p
l
o
r
i
n
g s
e
c
o
n
dp
r
o
b
l
e
m
.Th
v
i
b
r
a
n
ts
t
o
c
km
a
r
k
e
ti
nt
h
e
s
et
i
m
e
sh
a
smade
ba
1a
n
c
i
n
gt
h
ei
n
v
e
s
t
m
e
n
tp
o
r
t
f
o
l
i
omorec
h
a
l
l
e
n
g
i
n
g
.Th
ct
h
i
r
dc
a
s
emays
u
g
g
e
s
ta
1modela
p
p
r
o
p
r
i
a
t
ef
o
rs
u
c
hd
e
c
i
s
i
o
nm
a
k
i
n
g
.
m
a
t
h
e
m
a
t
i
ca
1
.
1
.
1 ElementsofProblemFormulatlon
I
nt
h
i
ss
e
c
t
i
o
n
wew
i
l
li
n
d
u
c
eef
o
r
ma
1e
l
e
m
e
n
t
so
ft
h
co
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
e
m
.I
n
t
h
i
sbookt
h
et
c
r
mp
r
o
d
u
c
ta
1
s
o f
e
r
st
oas
e
r
v
i
c
o
ras
t
r
a
t
e
g
y
.I
ts
h
o
u
l
dbe
u
n
d
e
r
s
t
o
o
da
top
z
a
t
i
o
np
r
e
s
u
p
p
o
s
e
st
h
eknowlcdgeo
fed
e
s
i
g
nr
u
l
e
sf
o
rt
h
e
s
p
i
f
i
cp
r
o
b
l
e
mp
r
i
m
a
r
i
l
yt
h
ea
b
i
l
i
t
yt
od
e
s
c
r
i
b
et
h
ed
e
s
i
g
ni
nm
a
t
h
e
m
a
t
i
c
a
1t
e
r
m
s
.
Th
e
s
et
e
r
m
scJuded
e
s
i
g
nv
a
r
i
a
b
l
e
sd
e
s
i
g
np
a
r
a
m
e
t
e
r
sandd
e
s
i
g
nf
u
n
c
t
i
o
n
s
.
T
r
a
d
i
t
i
o
na
1d
e
s
i
g
np
r
a
c
t
i
a
ti
sd
e
s
i
g
nw
i
t
h
o
u
tr
e
g
a
r
dt
oo
p
t
i
m
i
z
a
t
i
o
ni
n
c
Ju
d
e
sa
11
yweren
o
tf
o
r
m
a
l
l
yr
e
c
o
g
n
i
z
e
da
ss
u
c
h
.T
h
i
sa
l
s
o
o
ft
h
e
s
ee
l
e
m
e
n
t
s ughe
j
u
s
t
i
f
i
e
sep
r
e
r
e
q
u
i
s
i t
h
a
tyoumustb
ec
a
p
a
b
l
eo
fd
c
s
i
g
n
i
n
gt
h
e0
ti
fyoua
r
e
p
l
a
n
n
i
n
gt
oa
p
p
l
yet
e
c
h
n
i
q
u
c
so
fo
p
t
i
m
i t
i
o
n
.I
ti
sa
l
8
0agoodi
d
e
a1
0r
e
c
o
g
n
i
z
c
t
h
a
to
p
t
i
m
i
z
a
t
i
o
ni
sap
r
o
c
e
d
u
r
cf
o
rs
c
a
r
c
h
i
n
g1
h
eb
e
s
td
e
s
i
g
namongc
a
n
d
i
d
a
t
s

c
a
c
h
o
fwhichc
a
np
r
o
d
u
c
ea
na
c
c
e
p
t
a
b
l
ep
r
o
d
u
c
.
tTheo
c
e
df
o
rt
h
co
b
j
e
c
1o
rp
r
o
d
u
c
ti
sn
o
t
q
u
e
s
t
i
o
n
e
dh
e
r
e
.b
u
tt
h
i
smayb
eduet
oad
e
c
i
s
i
o
nb
a
s
e
d00o
p
t
i
m
i
z
a
t
i
o
na
p
p
l
i
e
di
n
a
n
o
t
h
e
rd
i
s
c
i
p
l
i
n
c
.

DeslgnV
a
r
l
a
b
l
e
s
: Designv
a
r
i
a
b
l
e
sa
r
ee
n
t
i
t
i
e
st
h
a
ti
d
e
n
t
i
f
yap
a
r
t
i
c
u
l
a
rd
e
s
i
g
n
.I
n
es
e
a
r
c
hf
o
r o
p
t
i
m
a
ld
e
s
i
g
ne
s
ee
n
t
i
t
i
e
sw
i
l
lc
h
a
n
g
eo
v
e
rap
r
c
s
c
r
i
b
e
dr
a
n
g
e
.
Thev
a
l
u
e
so
fac
o
m
p
l
e
t
es
e
to
ft
h
e
s
ev
a
r
i
a
b
l
e
sc
h
a
r
a
c
t
c
r
i
z
eas
p
e
c
i
f
i
cd
e
s
i
g
n
.T
h

numberandt
y
p
eo
fc
n
t
i
t
i
e
sb
c
l
o
n
g
i
n
gt
ot
h
i
ss
e
ta
r
ev
c
r
yi
m
p
o
r
t
a
n
ti
ni
d
e
n
t
i
f
y
i
n
ga
n
d
t
i
t
a
t
i
v
ed
e
s
i
g
np
r
o
b
l
e
m
.I
ti
se
s
s
e
n
t
i
a
li
sc
h
o
i
c
ec
a
p
t
u
e
s
e
t
t
i
n
gupequ
e
s
s
e
n
c
eo
f o
b
j
e
c
tb
e
i
n
gd
e
s
i
e
d da
tesamct
i
m
ep
r
o
v
i
d
eaq
u
a
n
t
i
t
a
t
i
v
e
c
h
a
r
a
c
t
e
r
i
z
a
t
i
o
no
ft
h
ed
e
s
i
g
np
r
o
b
l
e
m
.I
na
p
p
l
i
c
dm
a
t
h
e
m
a
t
i
c
a
lt
e
r
m
i
n
o
l
o
g
y
.d
e
s
i
g
n
v
a
r
i
a
b
l
e
ss
e
r
v
ea
st
h
eunknownso
ft
h
ep
r
o
b
l
e
mb
e
i
n
gs
o
l
v
c
d
.B
o
r
r
o
w
i
n
gana
n
a
10gy
fromt
h
ea
r
e
ao
fs
y
s
t
e
mdynami
a
ndc
o
n
t
r
o
lt
h
e
o
r
ye
ya
r
ee
q
u
i
va
1e
n
tt
od
e
f
i
n
i
n
g
t
h
es
t
a
t
eo
ft
h
es
y
s
t
c
mi
ni
sc
a
s
et
h
cs
t
a
t
co
fd
e
s
i
g
n
.T
y
p
i
c
a
l
l
yd
e
s
i
g
nv
a
r
i
a
b
l
e
sC
a
n
b
j
e
ct
'ss
i
z
es
u
c
ha
si
t
sl
e
n
g
t
ha
n
dh
e
i
g
ht
.I
no
t
h
e
r
bea
s
s
o
c
i
a
t
c
dw
i
t
hd
e
s
c
r
i
b
i
n
g o
c e
s
t
h
e
ymayr
e
p
r
e
s
e
n
1t
h
enumbero
fi
t
e
m
s
.Th
ec
h
o
i
c
eo
fd
e
s
i
g
nv
a
r
i
a
b
l
e
si
st
h
e
r
e
s
p
o
n
s
i
b
i
l
i
t
yo
fed
e
s
i
g
n
e
rg
u
i
d
c
dbyi
n
t
u
i
t
i
o
n
e
x
p
e
s
e
a
n
dk
n
o
w
l
e
d
g
e
. l
e
i
s
afundament
a
1r
e
q
u
i
r
e
m
e
n
tt
obemctbyi
ss
e
to
fd
e
s
i
g
nv
a
r
i
a
b
l
e
snamelye
ymust
bei
I
n
e
a
r
l
yi
n
d
e
p
e
n
d
e
n
.
tTh
i
smeansa
1youc
a
n
n
o
tc
s
t
a
b
l
i
s
ht
h
v
a
l
u
eo
fonco
fe
d
e
s
i
g
nv
a
r
i
a
b
l
e
sf
r
o
mev
a
l
u
e
so
fer
e
m
a
i
n
i
n
gv
a
r
i
a
b
l
e
st
h
r
o
u
g
hb
a
s
i
ca
r
i
t
h
m
e
t
i
c
i
nad
e
s
i
g
nh
a
v
i
n
gar
c
c
t
a
n
g
u
l
a
rc
r
o
s
s
(
s
c
a
l
i
n
go
ra
d
d
i
t
i
o
n
)o
p
e
r
a
t
i
o
n
s
.F
o
re
x
a
m
p
l
e
s
e
c
t
i
o
nyouc
a
n
n
o
th
a
v
eev
a
r
i
a
b
l
e
sr
e
p
r

e
n
t
i
n
ge
l
c
n
g

w
i
d
t
handa
r
e
a
.I
ft
h
e
p
r
e
s
c
r
i
b
e
det
h
i
r
di
sa
u
t
o
m
a
t
i
ca
1
l
ye
s
t
a
b
l
i
s
h
e
d
.I
ncomplexd
e
s
i
g
n
s
f
i
r
s
ttwoa
r
t
h
e
s
er
e
l
a
t
i
o
n
s
h
i
p
smayn
o
tb
ev
e
r
ya
p
p
a
r
e
n.
tN
e
v
e
r
t
h
e
l
e
s
s
t
h
ec
h
o
i
o
ft
h
es
e
to
f
d
e
s
i
g
nv
a
r
i
a
b
l
e
smustmeet c
r
i
t
e
r
i
o
no
fl
i
n
e
a
ri
n
d
e
p
e
n
d
e
n
c
ef
o
ra
p
p
l
y
i
n
gt
h
e
t
e
c
h
n
i
q
u
e
so
fo
p
t
i
m
i
z
a
t
i
o
n
. Fromap
r
a
c
t
i
c
n
lp
c
r
s
p
c
c
1
i
v
ct
h
cp
r
o
p
e
r
t
yo
fi
I
n
e
a
r
i
n
d
e
p
e
n
d
e
n
c
ei
d
e
n
t
i
f
i
e
saminimums
e
to
fv
a
r
i
a
b
l
e
st
h
a
tc
u
nc
o
m
p
l
c
t
e
l
yd
e
s
c
r
i
b
et
h
e

1
.
1

OPTIMIZATIONFUNDAMENTALS

d
e
s
i
g
n
.Thi
si
ss
i
g
n
i
f
i
c
a
n
tb
e
c
a
u
s
et
h
ee
f
f
o
r
ti
no
b
t
n
i
n
i
n
gt
h
es
o
l
u
t
i
o
nv
a
r
i
e
sa
sa
n
i
n
t
e
g
e
rpowero
ft
h
enumbero
fv
a
r
i
a
b
l
e
sa
n
dt
h
i
spoweri
s1
y
p
i
ca
11
yg
r
e
a
t
e
ra
n2
.
i
n
gt
h
er
e
q
u
i
r
e
m
e
n
te
n
s
u
r
e
sr
e
d
u
c
e
dd
i
f
f
i
c
u
l
t
yi
nm
a
t
h
e
m
a
t
i
c
a
l
l
ye
x
p
l
o
r
i
n
gt
h
c
M t
s
o
l
u
t
i
o
n
.
l
es
e
to
fd
e
s
i
g
nv
a
r
i
a
b
l
e
si
si
d
e
n
t
i
f
i
e
da
st
h
ed
e
s
i
g
nv
e
c
t
o
r
.1Isv t
o
rw
i
l
lb
e
c
o
n
s
i
d
e
r
ac
olumnv
e
c
t
o
ri
nt
h
i
sb
o
o
k
.I
nf
a
c
1

a
1
1v
e
c
t
o
r
sa
r
ecolumnv
e
c
t
o
r
s E
t
e
x
t
.Th
el
e
n
g
t
ho
fi
sv
e
c
t
o
r
whichi
sni
st
h
enumbero
fd
e
s
i
g
nv
a
r
i
l
e
si
oE
p
r
o
b
l
e
m
.Th
ed
e
s
i
g
nv
a
r
i
a
b
l
e
sc
a
nc
x
p
r
e
s
sd
i
f
f
e
r
e
n
td
i
m
e
n
s
i
o
n
a
lqu
t
i
t
i
e
si
n
p
r
o
b
l
e
m

b
u
ti
nt
h
em
a
l
h
e
m
a
t
i
c
a
lmodel

t
h
e
ya
r
ed
i
s
t
i
n
g
u
i
s
h
e
dby c
h
a
r
a
c
1
e
rx
.Al
I
o
ft
h
et
e
c
h
n
i
q
u
e
so
fo
p
t
i
m
i
z
a
t
i
o
ni
nt
h
i
sbookn
r
eb
a
s
c
donea
b
s
t
r
a
c
tm
a
t
h
e
m
a
t
i
c
a
l
es
u
b
s
c
r
i
p
tonxf
o
rcxamplcX3
r
c
p
r
e
s
e
n
t
st
h
et
h
i
r
dd
e
s
i
g
nv
a
r
i
a
b
l
e
w
h
i
c
h
model
.l
mayb
et
h
eh
e
i
g
h
to
fa
no
b
j
e
c
1i
nec
h
u
c
t
e
r
i
z
a
t
i
o
no
ft
h
ep
r
o
d
u
c
t
.T
h
i
sa
b
s
t
r
a
c
t
modeli
sne
s
s
a
r
yf
o
rm
a
t
h
e
m
a
t
i
c
a
lc
o
n
v
e
n
i
e
n
c
e
.T
h
i
sbookw
i
l
lr
e
f
e
rt
oed
e
s
i
g
n
v
a
r
i
a
b
l
e
si
noneo
f f
o
l
l
o
w
i
n
gw
a
y
s
:
(
1
)[
-refc i
n
gt
ot
h
ev
c
c
1
0
r0'
1d
s
i
g
nv
a
r
i
a
b
l
c
s

(
2
)X o
rx
r
e
f
e
r
r
i
n
gt
ot
h
cv
c
c
t
o
ra
g
a
i
no
m
i
l
l
i
n
g1
h
cs
q
u
a
r
eb
r
a
c
k
e
t
sf
o
r
c
o
n
v
c
n
i
c
n
c
ci
fa
p
p
r
o
p
r
i
a
t
e
x

jx
2
'
.
.
.
XIl]T_
i
n
d
i
c
a
t
i
n
gt
h
cv
c
c
l
o
rl
h
r
o
u
g
hi
t
se
l
e
m
e
n
t
s
.N
o
l
e e
(
3
)[
s
u
p
e
r
s
c
r
i
p
tt
r
a
n
s
p
o
s
i
t
i
o
nsymbolt
oi
d
e
n
t
i
f
yi
ta
sacolunmv
e
c
t
o
r
.
(
4
)X;.i
=1

2
....
n
r
e
f
e
r
r
i
n
gt
oa
l
lo
ft
h
ee
l
e
m
e
n
t
so
ft
h
ed
e
s
i
g
nv
c
t
o
r
.

1c
o
n
v
e
n
i
e
n
c
ei
se
x
t
c
n
d
c
dt
oa
l
lv
c
c
t
o
r
si
nt
h
eb
o
o
k
.
Thca
b
o
v
en
o
t
a
t
i
o
na
DeslgnParameters: I
ni
sbooke
s
ei
d
e
n
t
i
f
yc
o
n
s
t
a
n 1w
i
l
ln
o
tc
h
a
n
g
e
d
i
e
r
e
n
td
e
s
i
g
n
sa
r
ec
o
m
p
a
r
e
d
. Many 1
e
x
t
su
s
e t
e
r
md
e
s
i
g
np
a
r
a
m
e
t
e
r
s1
0
p
r
e
s
c
n
t d
c
s
i
g
nv
a
r
i
a
b
l
e
swed
e
f
i
n
e
de
a
r
l
i
e
randdon
o
tf
o
r
m
a
l
l
yr
e
c
o
g
n
i
z
ed
e
s
i
g
n
p
a
r
a
m
e
t
e
r
sa
sd
e
f
i
n
c
dh
c
.T
hep
r
i
n
c
i
p
n
lr
e
a
s
o
ni
st
h
a
tp
a
r
a
m
e
t
e
r
sh
a
v
enor
o
l
e1
0p
l
a
y
d
e
t
e
n
n
i
n
g o
p
t
i
m
a
1d
e
s
i
g
n
.Th
e
ya
r
es
i
g
n
i
f
i
c
a
n
ti
nt
h
ed
i
s
c
u
s
s
i
o
no
fm
o
d
e
l
i
n
g
i
s
s
u
e
s
.Exampleso
fp
a
r
a
m
e
t
e
r
si
n
c
l
u
d
em
a
t
c
r
ia
1p
r
o
p
c
y
a
p
p
l
i
e
dl
o
a
d
s
.a
n
dc
h
o
i
c
c
o
fs
h
a
p
e
.Th
ep
a
r
a
m
c
t
c
r
si
nt
h
ea
b
s
t
r
a
c
tm
a
t
h
e
m
a
t
i
c
a
lmodela
r
er
e
p
s
e
n
t
c
di
na
s
i
m
i
l
a
rf
o
r
ma
st
h
ed
e
s
i
g
nv
e
c
1
o
r

e
x
c
e
p
1 tweu
s
et
h
ec
h
a
r
a
c
t
c
rp
.Th
e
r
e
f
o
r
e
.[
P
]
P
[
P
"
P
2
'
.
..
P
q
1r
e
p
s
e
n
tep
a
r
a
m
e
t
e
r
so
ft
h
ep
r
o
b
l
e
m
.N
o
t
e e
l
e
n
g
o
f
t
h
e
p
a
r
n
e
r
v
e
c
t
o
ri
sq
.Excepti
ncd
i
s
c
u
s
s
i
o
no
fm
o
d
e
l
i
n
gepar
n
e
t
e
r
sw
i
l
ln
o
t
r
e
dt
o
a
st
h
e
ya
r
ep
r
i
m
a
r
i
l
yp
r
e
d
e
t
r
m
i
n
c
dc
o
n
s
t
a
n
t
si
n d
e
s
i
g
n
.
b
ee
x
p
l
i
c
i
t
l
yr
Des
'
nF
unctlons: D
e
s
i
g
nf
u
n
c
t
i
o
n
sd
e
f
i
n
em
e
a
n
i
n
g
f
u
li
n
f
o
r
m
a
t
i
o
na
b
o
u
tt
h
e
d
e
s
i
g
n
. Th
e
ya
r
ee
v
a
l
u
a
t
e
du
s
i
n
gt
h
ed
e
s
i
g
nv
a
r
i
a
b
l
e
sa
n
dd
e
s
i
g
np
a
r
a
m
e
t
e
r
s
d
i
s
c
u
s
s
e
de
a
r
l
i
e
r
.Theye
s
t
a
b
l
i
s
ht
h
em
a
t
h
e
m
a
t
i
c
a
lmodelo
ft
h
ed
e
s
i
g
np
r
o
b
l
e
m
.T
h
e
s
e
f
u
n
c
t
i
o
n
sc
a
nr
e
p
r
e
s
e
n
td
e
s
i
g
no
b
j
e
c
t
i
v
e
(
s
)a
n
d
l
o
rc
o
n
s
t
r
a
i
l
l
t
s
.Asinamei
m
p
l
i
e
s
d
e
s
i
g
n0
e
c
t
i
v
ed
r
i
v
e
st
h
es
e
a
r
c
hf
o
rt
h
eo
p
t
i
m
a
ld
e
s
i
g
n
.Thes
a
t
i
s
f
a
c
t
i
o
no
ft
h
e
t
r
a
i
l
l
t
se
s
t
a
b
l
i
s
h
e
st
h
ev
a
l
i
d
i
t
yo
ft
h
ed
e
s
i
g
n
.I
fn
o
te
x
p
l
i
c
it
1ys
t
a
t
e
d

ed
e
s
i
g
n
e
r
i
sr
e
s
p
o
n
s
i
b
l
cf
o
ri
d
n
t
i
f
y
i
n
gt
h
co
b
j
e
c
t
i
v
ca
n
dc
o
n
s
t
r
a
i
n
t
s
.M
i
n
i
m
i
z
et
h
emasso
ft
h
e

1
.
1 OPTIMIZATIONFUNDAMENTALS

INTRODUCTION

s
t
r
u
c
t
u
r
ew
i
l
lt
r
a
n
s
l
a
t
et
o o
b
j
e
c
t
i
v
e
U1ction.The sinthematerialmustbeless
t
h
a
nt
h
ey
ie
Ids
t
r
e
n
g
t
hwi
1
lt
r
a
n
s
l
a
t
et
oac
o
n
s
t
r
a
i
n
t
f
u
n
c
t
i
o
n
.I
nmanyp
r
o
b
l
e
m
s

i
ti
s
p
o
s
s
i
b
l
ef
o
rt
h
csamef
u
n
c
t
i
o
nt
os
w
i
t
c
hr
o
l
e
st
op
r
o
v
i
d
ed
i
f
f
c
r
c
n
td
e
s
i
g
ns
c
e
n
a
r
i
o
s
.

ConstraintFunctlons: As d
e
s
i
g
nf
u
n
c
t
i
o
n
st
h
e
s
ew
i
l
lb
ei
n
f
l
u
e
n
c
e
d byt
h
e
d
e
s
i
g
nv
a
r
i
a
b
l
e
s
.Th
ef
o
r
m
a
to
ft
h
e
s
ef
u
n
c
t
i
o
n
sr
e
q
u
i
r
e
sthemt
ob
ecomparedt
osome
ti
se
s
t
a
b
l
i
s
h
e
dbyd
e
s
i
g
nr
e
q
u
i
r
e
m
e
n
t

o
rt
h
ed
e
s
i
g
n
e
r
.
n
u
m
e
r
i
c
a
l
l
yl
i
m
i
t
i
n
gv
a
l
u
ea
Thi
sv
a
l
u
er
e
m
a
i
n
sc
o
n
s
t
a
n
td
u
r
i
n
gt
h
co
p
t
i
m
i
z
a
t
i
o
no
ft
h
ep
r
o
b
l
c
m
.Awcl
1d
e
s
c
r
i
b
c
d
d
c
s
i
g
np
r
o
b
l
e
mi
se
x
p
e
c
t
e
dt
oi
nc
1ude s
c
v
e
r
a
ls
u
c
hf
u
n
c
t
i
o
n
s
.w
h
i
c
hc
a
nb
e
r
e
p
r
e
s
e
n
t
e
da
sav
c
c
t
o
r
.Thec
o
m
p
a
r
i
s
o
ni
su
s
u
a
1ys
l
e
tupu
s
i
n
gt
h
et
h
r
e
cs
t
a
n
d
a
r
d
o
n
a
lo
p
e
r
a
t
o
r
s
:
=Sand
.C
o
n
s
i
d
e
ro
u
rf
i
r
s
te
x
a
m
p
l
e
.Le
t
f
i
m
J
(
X
)r
e
p
r
e
s
e
n
t
ef
u
n
c
t
i
o
n c
a
l
c
u
l
a
t
e
st
h
evolumeo
ft
h
enews
o
d
ac
a
nwea
r
ed
e
s
i
g
n
i
n
g
. l
e
c
o
n
s
t
r
n
tont
h
eds
i
g
nc
a
nb
ce
x
p
r
e
s
s
c
da
s

IIJ(X)=
500cmJ
I
nt
l
I
es
e
c
o
n
de
x
a
m
p
l
e
.
l
e
t
m2(X)b
t
h
ef
u
n
c
t
i
o
nt
h
a
tc
a
l
c
u
l
a
t
e
st
h
ed
e
f
l
e
c
t
i
o
no
ft
h
e
o
n
s
t
r
a
i
n
tc
a
nb
es
t
a
t
c
da
s
beamu
n
d
e
rt
h
ea
p
p
l
i
e
dl
o
a
d
.Thec

f
t
1
1
l2
(
X
):
S
;Imm

llY 2 44 yfJ

ObJectlveFunction
Thet
r
a
d
i
t
i
o
n
a
ld
e
s
i
g
no
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
e
mi
sd
c
f
i
n
d
ef
o
r
m
a
to
ft
h
i
ss
t
a
t
e
m
c
n
ti
su
s
u
a
l
l
yt
om
i
n
i
m
i
z
c
u
s
i
n
gas
i
n
g
l
eo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
n
. l
o
rmaximizesomeq
u
a
n
t
i
t
y i
sc
a
l
c
u
l
a
t
e
du
s
i
n
gsomcd
c
s
i
g
nf
u
n
c
t
i
o
n
.T
h
i
s
f
u
n
c
t
i
o
nm
u
s
td
e
p
e
n
de
x
p
l
i
c
i
t
l
yo
ri
m
p
l
i
c
i
t
l
yon t
h
ed
e
s
i
g
nv
a
r
i
a
b
l
e
s
.I
nt
h
e
l
i
t
e
r
a
t
u
r
ci
sp
r
o
b
l
c
mi
scxp
s
s
e
de
xc
1u
s
i
v
e
l
yw
i
t
h
o
u
tl
o
s
so
fg
e
n
er
a
I
i
t
ya
sa
minimump
r
o
b
l
e
m
.A maximump
r
o
b
l
e
mc
a
nb
r
c
c
a
s
ta
sam
i
n
i
m
i
z
a
t
i
o
np
r
o
b
l
c
m
u
s
i
n
gt
h
en
c
g
a
t
i
v
co
r r
e
c
i
p
r
o
c
a
lo
ft
h
cf
u
n
c
t
i
o
nu
s
e
df
o
rt
h
eo
b
j
e
c
t
i
v
cf
u
n
c
t
i
o
n
.l
n
t
h
ef
i
r
s
texamplei
n
d
u
c
e
de
a
r
l
i
e
r
co
e
c
t
i
v
ei
st
om
i
n
i
m
i
z
ec
o
s
t
.l
e
r
e
f
o
r
e
d
e
s
i
f
u
n
c
t
i
o
nr
e
p
r
e
s
gc
o
s
twi
l
Ibet
h
eo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
n
.I
n es
e
c
o
n
d s
e

eo
b
j
e
c
t
i
v
ei
st
om
i
n
i
m
i
z
em
a
s
s
.I
nt
h
et
h
i
r
dc
a
s
e
theobjectivei
st
omaximizc
machineu
t
i
l
i
z
a
t
i
o
n
.l
ea
r
o
fs
i
n
g
l
eo
b
j
e
c
t
i
v
ed
e
s
i
i
sc
o
n
s
i
d
e
r
e
dm
a
t
u
r
et
o
d
a
y
.
Nonews
o
l
u
t
i
o
nt
e
c
h
n
i
q
u
e
sf
o
rc
l
a
s
s
i
c
a
lp
r
o
b
l
c
m
sh
a
v
b
e
c
na
d
v
a
n
c
e
df
o
rsomct
i
m
e
now.T
o
d
a
y
.mucho
ft
h
eworki
na
p
p
l
i
e
do
p
t
i
m
i t
i
o
ni
sd
i
r
c
c
t
e
da
te
x
p
a
n
d
i
n
g
a
p
p
l
i
c
a
t
i
o
n
st
op
r
a
c
t
i
c
a
lp
r
o
b
l
e
m
s
.l
nmanyc
a
s
e
s
.t
h
i
sh
a
si
n
v
o
l
v
e
dc
r
c
a
t
i
v
eu
s
eo
ft
h
e
s
o
l
u
t
i
o
nt
e
c
h
n
i
q
u
e
s
.I
n a
b
s
t
r
a
c
tm
a
t
h
e
m
a
t
i
c
a
lm
o
d
e
l
.t
h
eo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
ni
s
r
e
p
r
o
s
e
n
t
e
dbyt
h
esymbolfToi
n
d
i
c
a
t
ei
t
sd
e
p
e
n
d
e
n
c
eont
h
ed
e
s
i
g
nv
a
r
i
a
b
l
e
s
.i
ti
s
.A m
q
u
e
n
t
l
yc
x
p
r
e
s
s
e
d f(x"x
2
.
...
x
orec
o
n
c
i
s
c p
r
c
s
e
n
t
a
t
i
o
ni
sj
{
X
)
.
f
n)
ti
sas
c
a
Ia
r(
n
o
ta
S
i
n
g
l
eo
b
j
e
c
t
i
v
ep
r
o
b
l
c
m
sh
a
v
eo
n
l
yonef
u
n
c
t
i
o
nd
c
n
o
t
e
dbyfI
V t
o
r
)
.Notca
ta
l
t
h
o
u
g
ht
h
eo
b
j
c
c
t
i
v
ef
u
n
c
t
i
o
nd
e
p
e
n
d
sonP(
p
a
r
a
m
e
t
e
rv
e
c
t
o
r
)
.i
t
1udedi
nt
h
ef
o
r
m
a
.
t
i
sn
o
te
x
p
l
i
c
i
t
l
yi
nc
M
u
l
t
i
o
b
j
e
c
t
i
v
ea
n
dm
u
l
t
i
d
i
s
c
i
p
l
i
n
a
r
yd
s
i
g
n
sa
r
ci
m
p
o
r
t
a
n
td
e
v
e
l
o
p
m
e
n
t
st
o
d
a
y
.
M
u
l
t
i
o
b
j
t
i
v
ed
e
s
i
g
n
o
rm
u
l
t
i
p
l
eo
b
j
e
c
t
iv
ed
e
s
i
g
n
.r
e
f
e
r
s

u
s
i
n
gs
e
v
c
r
a
Id
ie
r
e
n
td
e
s
i
g
n
f
u
n
c
t
i
o
n
st
oes
hf
o
rop
n
a
ld

i
g
n
.Gene
r
a
I
lye
ya
r
eexpecdbe
c
t
i
v
e
s
.Th
e
yc
o
u
l
da
l
s
ob
eo
p
c
r
a
t
i
n
go
b
j
e
c
t
i
v
e
s
.l
ec
u
r
r
e
n
ta
p
p
r
o
a
c
h
c
o
n
f
l
i
c
t
i
n
g0c
o
l
u
t
i
o
no
ft
h
e
s
ep
r
o
b
l
e
m
si
n
v
o
l
v
e
ss
t
a
n
d
a
r
do
p
t
i
m
i
z
a
t
i
o
np
r
o
c
e
d
a
p
p
l
i
c
dt
oa
t
oes
s
i
n
g
l
e(o
n
s
t
r
u
c
t
e
do
b
j
e
c
t
i
v
eo
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
c
mb
a
s
e
dont
h
ed
i
f
f
e
r
e
n
tm
u
l
t
i
p
l
e
o
b
j
e
c
t
i
v
e
s
.Apop
a
ra
p
p
r
o
a
c
hi
st
ou
s
eas
u
i
t
a
b
l
yw
e
i
g
h
d
l
i
n
r
c
o
m
b
i
n
a
t
i
o
no
f
m
u
l
t
i
p
l
co
b
j
e
c
t
i
v
e
s
.Ap
r
a
c
t
i
c
a
l
l
i
m
i
t
a
t
i
o
nw
i
i
sa
p
p
r
o
a
c
hi
s.
e
c
h
o
i
o
fw
e
i
g
h
t
su
s

i
nt
h
emode
.
lT
h
i
sa
p
p
r
o
a
c
hh
a
sn
o
tb
e
e
nc
m
b
r
a
c
e
dw
i
d
e
l
y
.Ana
l
t
e
m
a
t
i
v
ea
p
p
r
o
a
c
ho
f
e
c
t
i
v
e

a
n
ds
o
l
v
i
n
gas
i
n
g
l
co
b
j
e
c
t
i
v
cp
r
o
b
l
e
mw
i
t
ha
d
d
i
t
i
o
n
a
l
r
e
c
o
g
n
i
z
i
n
gap
r
c
n
c
o
n
s
t
r
a
i
n
t
sb
a
s
e
don ningo
ven
c
t
i
o
n
sc
a
nu
s
u
a
l
l
yg
e
n
e
a
c
c
e
p
l
e
s
o
l
u
t
i
o
n
.I
nm
u
l
t
i
o
b
j
e
c
t
i
v
ep
r
o
b
l
e
n

Thccons
i
n
tf
u
n
c
t
i
o
n
sc
a
nb
ec
1a
s
s
i
f
i
e
da
se
q
l
l
a
l
i
t
yc
o
n
s
t
r
a
i
n
t
s[
l
i
k
e
f
u
n
l
(
X
)a
b
o
v
e
J
o
ri
n
e
q
u
a
l
i
t
yc
o
n
s n
t
s[
l
i
k
e
f
u
n
2
(
X
)
]
.
fc
o
n
s
t
r
a
i
n
t
sa
r
e
P
r
o
b
l
e
m
sw
i
t
h
o
u
tc
o
n
s
t
r
a
i
n
t
sa
r
et
c
r
m
c
du
n
c
o
n
s
t
r
a
i
n
e
dp
r
o
b
l
e
m
s
.I
h
e
nm
e
e
t
i
n
g them i
s more pmount a
no
p
t
i
m
i
z
a
l
i
o
n
.C
o
n
s
t
r
a
i
n
t
p
r
e
s
e
n
tt
s
a
t
i
s
f
a
c
t
i
o
ni
sn
e
c
e
s
s
a
r
yb
e
f
o
r
ct
h
ed
c
s
i
g
ne
s
t
a
b
l
i
s
h
e
dby c
u
r
r
c
n
tv
a
l
u
eo
ft
h
e
d
e
s
i
v
a
r
i
a
b
l
e
si
sc
o
n
s
i
d
e
r
e
dv
a
l
i
danda
c
c
e
p
t
a
b
l
e
.I
fc
o
n
s
t
r
a
i
n
t
sa
r
en
o
ts
s
f
i
e

d
t
h
e
nt
h
e
r
ei
snos
o
l
u
t
i
o
n
.Af
e
a
s
i
b
l
ed
e
s
i
g
ni
sonei
nwhicha
l
lo
ft
h
ec
o
n
s
t
r
a
i
n
t
sa
r
e
s
a
t
i
s
f
i
e
d
.Ano
p
t
i
m
a
ls
o
l
u
t
i
o
ni
so
n
ea
th
a
smett
h
ed
e
s
i
g
no
b
j
e
c
t
i
v
e
.Ano
p
t
i
m
a
l
d
c
s
i
g
nm
u
s
tb
ef
c
a
s
i
b
l
c
.Thed
e
s
i
g
ns
p
a
c
cc
n
c
l
o
s
c
dbyt
h
cc
o
n
s
t
r
a
i
n
t
si
sc
a
l
l
e
dt
h
e
f
e
a
s
i
b
l
ed
o
m
a
i
n
.D
c
s
i
g
ns
p
a
c
ei
sd
e
s
c
r
i
b
e
dafewp
a
r
a
g
r
a
p
h
sb
c
l
o
w
.

q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
t
sa
r
em
a
t
h
e
m
a
t
i
c
a
l
l
yn
e
a
ta
n
de
a
s
yt
o
EqualltyConstralnts: E
h
a
n
d
l
c
.N
u
m
e
r
i
c
a
l
l
y
e
yJ
1
u
i
r
emoree
f
f
o
r
tt
os
a
t
i
s
f
y
.Theya
r
ea
l
s
omorer
e
s
t
r
i
c
t
i
v
e
ont
l
Iedesigntheyl
i
m
i
tt
h
er
e
g
i
o
nfromwhicht
h
es
o
l
u
t
i
o
nc
a
nbeo
b
t
a
i
n
e
d
.The
symbolr
e
p
r
e
s
e
n
t
i
n
ge
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
t
si
nt
h
ea
b
s
t
r
a
c
tmodeli
sh
.The
r
emayb
e
moret
h
a
nonee
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
ti
nt
h
ed
e
s
i
g
np
r
o
b
l
c
m
.A v
e
c
t
o
rr
e
p
r
e
s
e
n
t
a
t
i
o
n
f
o
r u
a
l
i
t
yc
o
n
s
t
r
a
i
n
t
si
si
n
t
r
o
d
u
c
e
dt
h
r
o
u
g
ht
h
ef
o
l
l
o
w
i
n
gr
e
p
r
e
s
e
n
t
a
t
i
o
n
.[
8
]
[
hl
h2....
h
]
.andhk:k=1
2
....
1a
r
ewayso
fi
d
e
n
t
i
f
y
i
n
gee
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
t
s
.l
e
d
e
p
e
n
d
e
n
c
eont
h
ed
e
s
i
g
nv
a
r
i
a
b
l
e
sXi
so
m
i
t
t
e
df
o
rc
o
n
v
c
n
i
e
n
c
e
.Notca
tt
h
el
e
n
g
t
h
o
ft
h
ev
e
c
t
o
ri
s1
.Ani
m
p
o
r
t
a
n
t onf
o
rd
i
s
t
i
n
g
u
i
s
h
i
n
gt
h
ee
q
u
a
l
i
t
ya
n
di
n
e
q
u
a
l
i
t
y
c
o
n
s
t
r
a
i
n
t
si
sa
tt
h
e
ya
r
cm
n
n
i
p
u
l
a
t
e
dd
i
f
f
e
r
e
n
t
l
yi
nt
h
es
e
n
r
c
hf
o
rt
h
eo
p
t
i
m
a
l
s
o
l
u
t
i
o
n
.Thenumberno
fd
e
s
i
g
nv
a
r
i
a
b
l
e
si
ncp
r
o
b
l
c
mm
u
s
tb
eg
r
e
a
t
c
r t
h
e
numbero
fe
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
t
s1
f
o
ro
p
t
i
m
i
z
a
t
i
o
nt
ot
a
k
ep
l
a
c
e
.l
fni
sc
q
u
a
lt
o.
1t
h
e
n
p
r
o
b
l
e
mw
i
l
lb
es
o
l
v
e
dw
i
t
h
o
u
tr
e
f
e
r
e
n
c
et
ot
h
eo
b
j
e
c
t
i
v
e
.l
nm
a
t
h
e
m
a
t
i
c
a
lt
e
r
m
s
fu
n
k
n
o
w
n
s
.I
fni
sl
e
s
st
h
a
n1

t
h
e
nyou
t
h
enumbero
fe
q
u
a
t
i
o
n
sm
a
t
c
h
e
senumbero
h
a
v
ea
no
v
e
r
d
e
t
e
r
m
i
n
e
ds
c
lo
fr
e
l
a
t
i
o
n
swhichc
o
u
l
dr
e
s
u
l
ti
na
ni
n
c
o
n
s
i
s
t
e
n
tp
r
o
b
l
c
m
d
e
f
i
n
i
t
i
o
n
.Thes
c
to
fc
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
t
sm
u
s
tb
el
i
n
e
a
r
l
yi
n
d
c
p
e
n
d
e
n.
tB
r
o
a
d
l
y
t
h
i
s

INTAODUCTION

1
.
1 OPTIMIZATIONFUNDAMENTALS 9

i
m
p
l
i
e
s y
o
uc
a
n
n
o
to
b
t
a
i
no
n
eo
ft
h
ec
o
n
s
a
i
n
t
sf
r
o
me
l
e
m
e
n
t
a
r
ya
r
i
t
h
m
e
t
i
c
ss
e
r
v
e
s1
0e
n
s
u
r
et
h
a
tt
h
em
a
t
h
e
m
a
t
i
c
a
l
o
p
e
r
a
t
i
o
n
so
nt
h
er
c
m
a
i
n
i
n
gc
o
n
s
t
r
a
i
n
t
s
.Thi
.T
h
e
s
et
c
c
h
n
i
q
u
e
sa
r
eb
a
s
e
do
nm
e
t
h
o
d
sf
r
o
mI
in
c
a
r
s
e
a
r
c
hf
o
rs
o
l
u
t
i
o
nw
i
l
ln
o
tf
a
il

t
h
e u
a
1i
t
yc
o
n
s
a
i
n
t
sa
r
e
a
l
g
e
b
r
a
.I
nt
h
es
t
a
n
d
a
r
df
o
r
m
a
tf
o
ro
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
e
m
s
w
r
i
t
t
e
nw
i
t
ha0o
nt
h
er
i
g
h
t
h
a
n
ds
i
d
e
.T
h
i
sm
e
a
n
s ee
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
ti
nt
h
c
s
s
e
da
s
f
i
r
s
te
x
a
m
p
l
ew
i
l
lb
ee
x
p

h
l
(
X
)
:
f
u
n
l
(
X
)-5
00 0

I
np
r
a
c
t
i
c
a
1p
r
o
b
l
e
m
s
e
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
t
sa
r
er
a
r
I
ye
n
c
o
u
n
t
c
r
e
d
.

8
1(
X
)
:
n2
(X)-1O

b
o
v
ed
e
f
i
n
i
t
i
o
n
sa
1l
o
wu
st
oa
s
s
e
m
b
l
eI
h
eg
e
n
e
r
a
l
TheStandardFormat:Thea
.
c
tm
a
t
h
e
r
n
a
t
i
c
a
lm
o
d
e
la
s
a
b
s
.
1
l

I
n
e
q
u
a
l
l
t
yC
o
n
s
t
r
a
i
n
t
s
:I
n
u
a
1i
t
yc
o
n
s
a
i
n
t
sa
p
p
r
moren
a
t
u
r
a
l
l
yi
np
r
o
b
l
c
m
ne
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
t
sa
l
s
op
r
o
v
i
d
emoref
l
e
x
i
b
i
l
i
t
yi
nd
c
s
i
g
ns
e
l
e
c
t
i
o
n
.
f
o
r
m
u
l
a
t
i
o
n
.I
'
a
i
n
t
si
nt
h
ea
b
s
t
r
a
c
tm
o
d
e
li
s8
.T
h
e
r
emay
Thes
y
m
b
o
lr
e
p
r
e
s
c
n
t
i
n
gi
n
e
q
u
a
l
i
t
yc
o
n
s
n
ci
n
u
a
1i
t
yc
o
n
s
t
r
a
i
n
ti
nt
h
ed
e
s
i
g
np
r
o
b
l
e
m
. The v
e
c
t
o
r
b
e more o
s
e
n
t
a
t
i
o
nf
o
ri
n
e
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
t
si
ss
i
m
i
l
a
rt
ow
h
a
tweh
a
v
s
e
e
nb
e
f
o
r
e
.Thu
s
r
e
p
r
[G)[
8
1
8
2..
.81
/
]
a
n
d
:j=1

2.
..
ma
r
cw
a
y
so
fi
d
e
n
t
i
f
y
i
n
gI
h
ei
n
e
q
u
a
l
i
t
y
c
o
n
s
a
i
n
t
s
.m r
e
p
r

e
n
t
senumbero
fi
n
e
q
u
a
1
i
t
yc
o
n
s
t
r
a
i
n
t
s
.AI
ld
e
s
i
g
nf
u
n
c
t
i
o
n
s
.8i
su
s
e
dt
o
e
x
p
l
i
c
i
t
l
yo
ri
m
p
l
i
c
i
t
l
yd
e
p
e
n
dont
h
ed
e
s
i
g
n(
o
ri
n
d
e
p
e
n
d
e
n
t
)v
a
r
i
a
b
l
eX
d
e
s
c
r
i
h
ob
o l
o
s
st
h
u
no
rc
q
u
a
l1
0)dg
r
e
a
l
c
r 0
1
"e
q
u
l
l
l1
0()c
o
n
s
t
r
a
l
n
t
s
.
n
ds
t
r
i
c
t
l
yI
c
s
st
h
(<)a
r
en
o
tu
s
e
dmuchi
no
p
t
i
m
i
z
a
t
i
o
n
S
t
r
i
c
t
l
yg
r
e
a
t
e
r(>)a
o
l
u
t
i
o
n
sa
r
e u
a
1
1
ye
x
p
e
c
t
e
dt
ol
i
ea
tec
o
n
s
t
r
a
i
n
tb
o
u
n
d
a
r
y
.I
nt
h
e
b
e
c
a
u
s
e s
a
l
lp
r
o
b
l
e
m
sa
r
ce
x
p
r
e
s
s
e
dw
i eSr
e
l
a
t
i
o
n
s
h
i
p
.M
o
r
e
o
v
e
r
t
h
e
s
t
a
n
d
a
r
df
o
r
m
a
t
i
g
ni
s
O
. l
ei
n
e
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
tf
r
o
mt
h
es
e
c
o
n
de
x
a
m
p
l
e
r
i
g
h
t
h
a
n
ds
i
d
eo
feSs
ftm2(X)i
ss
e
tu
pa
s

DesignSpace:led
e
s
i
g
ns
p
a
c
e
es
p
a
c
ea
twi

lb
es
e
a
r
c
h
e
df
o
ro
p
t
i
m
a
l
d
e
s
i
g
ni
st
h
eE
u
c
l
i
d
e
a
no
rC
a
r
t
e
s
i
a
nn
d
i
m
e
n
s
i
o
n
a
ls
p
a
c
eg
e
n
e
r
a
t
e
db
yt
h
en
i
n
d
e
p
e
n
d
e
n
td
e
s
i
g
nv
a
r
i
a
b
l
e
sX
.T
h
i
si
sag
e
n
e
r
a
1
i
z
a
t
i
o
no
ft
h
et
h
r
e
e
d
i
m
e
n
s
i
o
n
a
l
p
h
y
s
i
c
a
ls
p
a
c
ew
i
t
hw
h
i
c
h we a
r
ef
a
m
i
l
i
a
r
.F
o
rt
e
nd
s
i
g
nv
a
r
i
a
b
l
e
si
ti
sa
t
e
n
d
i
m
e
n
s
i
o
n
a
ls
p
a
c
e
.T
h
i
si
sn
o
te
a
s
yt
oi
m
a
g
i
n
e
.I
li
sa
1s
on
o
tc
a
s
yt
oe
x
p
r
e
s
st
h
i
s
i
n
f
o
r
m
a
t
i
o
n r
o
u
g
haf
i
g
u
r
eo
rg
r
a
p
hb
e
c
a
u
s
eo
ft
h
eI
im
i
t
a
t
i
o
no
ft
h
e
t
h
r
e
e
d
i
m
e
n
s
i
o
n
a
lw
o
r
l
d
.However
i
fed
:
s
i
g
nv
a
r
i
a
b
l
e
sa
r
ei
n
d
e
p
e
n
d
e
n
t

e
nt
h
e
l
1
d
i
m
e
n
s
i
o
n
a
lc
o
n
s
i
d
e
r
a
t
i
o
n
sa
r
em
e
r
ee
x
t
r
a
p
o
l
a
t
i
o
n
so
ft
h
ee
d
i
m
e
n
s
i
o
na
1r
e
a
l
i
t
y
.
Ofc
o
u
r
s
e

wec
a
n
n
o
tg
e
o
m
e
t
r
i
ca
1l
yd
e
f
i
n
t
h
e
mt
h
o
u
g
hwew
i
l
lb
ew
o
r
k
i
n
gw
i
t
ht
h
c
n
u
m
b
e
r
s
.Thes
i
d
ec
o
n
s
t
r
a
i
n
t
sI
i
m
i
tt
h
es
e
a
r
c
h g
i
o
n
i
m
p
l
y
i
n
gt
h
a
to
n
l
ys
o
l
u
t
i
o
n
sa
t
l
i
ew
i
t
h
i
n ac
r
t
a
i
nr
e
g
i
o
nw
i
l
lb
ea
c
c
e
p
t
a
b
l
e
.T
h
e
yd
e
f
i
n
ea
nn
d
i
m
e
n
s
i
o
n
a
l
r
e
c
t
a
n
g
u
l
a
rr
e
g
i
o
n(
h
y
p
e
r
c
u
b
e
)f
r
o
mw
h
i
c
hef
e
a
s
i
b
l
ea
n
do
p
t
i
m
a
ls
o
l
u
t
i
o
n
sm
u
s
tb
e
c
h
o
s
e
n
.Lat
e
r
wewi

ls
a
tt
h
em
a
t
h
e
m
a
t
i
c
a
1m
o
d
e
l
si
no
p
t
i
m
i
z
a
t
i
o
na
r
eu
s
u
a
l
l
y
d
e
s
c
r
i
b
e
db
yn
o
n
l
i
n
e
a
rr
e
l
a
t
i
o
n
s
h
i
p
s
.

1
1
es
o
l
u
t
i
o
n
st
os
u
c
hp
r
o
b
l
e
m
sc
a
n
n
o
tb
e
a
n
a
1
y
t
i
c
a
l
l
yp
r
e
d
i
c
t
e
d t
h
e
ya
r
et
y
p
i
c
a
1l
ygovemedb
yeu
n
d
e
r
l
y
i
n
gn
u
m
e
r
i
c
a
l
t
e
c
i
q
u
eu
s
e
dt
os
o
l
v
et
h
e
m
.I
ti
sn
e
c
c
s
s
a
r
yt
o(i
c
tt
h
es
o
l
u
t
i
o
n
st
oa
na
c
c
e
p
t
a
b
l
e
r
e
g
i
o
n
.Thes
i
d
ec
o
n
s
t
r
a
i
n
t
sp
r
o
v
i
d
es r
e
a
d
ym
e
c
h
a
n
i
s
mf
o
ri
m
p
l
e
m
e
n
t
i
n
gt
h
i
sl
i
m
it
.
C
a
r
em
u
s
tb
et
a
k
e
nt
h
a
tt
h
c
s
el
i
m
i
t
sa
r
en
o
ti
m
p
o
s
e
do
v
e
r
z
e
a
l
o
u
s
l
y
.Thc
r
em
u
s
tb
e
u
m
e
r
i
c
a
lt
e
c
h
n
i
q
u
e
s1
0o
p
e
r
a
t
e
.
r
e
a
s
o
n
a
b
l
es
p
a
c
ef
o
r n

(
x
.
x
2
.
.
.

xn)
M
i
n
i
m
i
z
e f

S
u
b
j
e
c
tt
o
: hl(x"X
2

XII)=O
h2
(
x
.
x
2
.
.
.
xn)=0

(
1
.2
)

h
/
(
X
I
'X
2

.
.
X
)=0
I
I

I
n c
a
s
eo
fi
n
e
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
t
sad
i
s
t
i
n
c
t
i
o
ni
smadca
st
ow
h
e
t
h
e
rt
h
ed
c
s
i
g
n
v
a
r
i
a
b
l
e
sl
i
eont
h
ec
o
n
s
t
r
a
i
n
tb
o
u
n
d
a
r
yo
ri
n i
n
t
e
r
i
o
ro
ft
h
er
e
g
i
o
nb
o
u
n
d
e
dbyE
c
o
n
s
t
r
a
i
nt
.I
ft
h
es
e
to
fd
s
i
g
nv
a
r
i
a
b
l
e
sl
i
eo
nt
h
eb
o
u
n
d
a
r
yo
fec
o
n
s
t
r
a
i
n
t
m
a
t
h
e
m
a
t
i
c
a
l
l
y
t
h
i
se
x
p
r
e
s
s
e
sef
a
c
tt
h
a
tc
o
n
s
t
r
a
i
n
ti
ss
a
t
i
s
f
i
e
dw
i
t
hs
t
r
i
c
te
q
u
a
l
i
t
y
i
s
g =O
.Thec
o
n
s
t
r
a
i
n
ta
c
t
sl
i
k
ea
ne
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
nt
.I
no
p
t
i
m
i
z
a
t
i
o
n
sp
a
r
t
i
c
u
l
a
rc
o
n
s
a
i
n
ti
sr
e
f
e
r
r
e
dt
oa
sa
na
c
t
i
v
ec
o
n
s
t
r
a
i
nt
.I
ft
h
es
e
t
t
e
r
m
i
n
o
l
o
g
yi
t
h
a
ti
s
t
h
e
yl
i
i
n
s
i
d
eer
e
g
i
o
no
ft
h
c
o
fd
e
s
i
g
nv
a
r
i
a
b
l
e
sdon
o
tl
i
eont
h
eb
o
u
n
d
a
r
y
t
h
e
ya
r
cc
o
n
s
i
d
e
di
n
a
c
t
i
v
ec
o
n
s
t
r
a
i
n
t
s
.M
a
t
h
e
m
a
t
i
ca
1l
y
ec
o
n
s
t
r
a
i
n
t
c
o
n
s
t
r
a
i
n
t
s
c
l
a
t
i
o
ng<O
.Ani
n
e
q
u
a
l
i
t
yc
o
n
s
U
'
a
i
n
tc
a
nt
h
e
r
c
f
o
r
eb
ee
i
t
h
e
ra
c
t
i
v
eo
r
s
a
t
i
s
f
i
e
ser
Jv
e
.
l
l
l
a
ct
SldeC
o
n
s
t
r
a
l
n
t
s
:S
i
d
ec
o
n
s
a
i
n
t
sa
r
ean
e
c
e
s
s
a
r
yp
a
r
to
ft
h
es
o
l
u
t
i
o
nt
e
c
h
n
i
q
u
e
s
e
s
c
ia
1
l
yn
u
m
e
r
i
c
a
lo
n
c
s
.Theye
x
p
r
e
s
st
h
er
a
n
g
ef
o
rt
h
ed
e
s
i
g
nv
a
r
i
a
b
l
c
s
.Eac
h
a
1u
e
sf
o
ri
t
sl
o
w
e
ra
n
du
p
p
e
rI
im
it
.Th
e
d
e
s
i
g
nv
a
r
i
a
b
l
em
u
s
tb
eb
o
u
n
db
yn
u
m
c
r
i
c
a
lv
d
e
s
i
g
n
e
rmakest
h
i
sc
h
o
i
c
eb
a
s
e
do
nh
i
sa
n
t
i
c
i
p
a
t
i
o
no
fa
na
c
c
e
p
t
a
b
l
ed
e
s
i
g
n
.

8
.
(
X
"X
2

.
.Xn)O

8
2
(
X
I
'X
2
Xn)S0

(
1
.
3
)

g
m
(
X
I
X
2

.
.Xn)SO

x
lSXjSxii 1

2
.
.
.
n

(
1
.
4
)

Thesamep
r
o
b
l
e
mc
a
nb
ee
x
p
r
e
s
s
e
dc
o
n
c
i
s
e
l
yu
s
i
n
gt
h
ef
o
l
l
o
w
i
n
gn
o
t
a
t
i
o
n
:
)
M
i
n
i
m
i
z
e f
(
x
.
X
2
..
.

X
n

(
1
.5)

S
u
b
j
e
c
tt
o
: h
k
(
x
t
X
2

Xn)=
0k=
1

2
.
.
.
I

(
1
.
6
)

(
X
"X
2
..

X
1

2
..
.
.
m
I)S0j=

(
1
.7)

10 I
N
T
R
O
O
U
C
T
1
0N

1
.
1O
P
T
I
M
I
Z
A
T
I
O
NF
U
N
O
A
M
E
N
T
A
l
S 1
1

xlsxjSx i=1

2
.
.
.
n

(
1
.8
)

E
x
p
l
o
i
t
i
n
gv
e
c
t
o
rn
o
t
a
t
i
o
nt
h
em
a
t
h
e
m
a
t
i
c
a
lmodeli
s
M
i
n
i
m
i
z
e j
(X)
[
X
]
.

(
1
.9
)

h(X
)
]
1=0
S
u
b
j
e
c
tt
o
: [

(
1
.
10
)

(X)
]
t
nS0

(
1
.1
1
)

X10wSXSX
u
p

(
1
.
12
)

Th
eabovem
a
t
h
e
m
a
t
i
c
a
l model e
x
p
r
e
s
s
e
s f
o
l
l
o
w
i
n
gs
t
a
n
d
a
r
df
o
r
m
a
to
ft
h
e
o
p
t
i
m
i
z
onp
r
o
b
l
e
me
x
p
r
e
s
s
e
di
nn
a
t
u
r
a
l
l
a
n
g
u
a
g
e
:
Minimizet
h
eo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
n

js
u
b
j
e
c
tt
o1
e
q
u
a

i
t
yc
o
n
s
t
r
a
i
n
t
s

mi
n
c
q
u
a
l
i
t
y
i
t
hend
e
s
i
g
nv
a
r
i
a
b
l
e
sI
y
i
n
gb
e
t
w
e
e
np
r

c
r
i
b
e
dl
o
w
e
rand
c
o
n
s
t
r
a
i
n
t
sw
u
p
p
e
rl
i
m
i
t
s
.

F
l
g
u
r
e1
.
1 E
x
a
m
p
l
e1
.
1-De
s
i
g
n0
1anewb
e
v
e
r
a
g
ec
a
n
.

p
J

:
<
:

l
et
e
c
h
n
i
q
u
c
si
ni
sbookw
i
l
la
p
p
l
yt
ot
h
cp
r
o
b
l
e
md
e
s
c
r
i
b
c
di
ncabovcf
o
r
m
a
t
.
Tos
o
l
v
ea
n
ys
p
e
c
i
f
i
cd
e
s
i
g
np
r
o
b
l
e
mi
ti
sr
e
q
u
i
r
c
d1
0r
e
f
u
n
n
u
l
a
t
ct
h
ep
r
o
b
l
c
mi
nt
h
e
e
t
h
o
d
sc
a
nb
ea
p
p
l
i
e
dd
i
r
e
c
t
l
y
.A
l
s
oi
nt
h
i
sbook
t
h
e
a
b
o
v
emanners
o m

t
e
c
h
n
i
q
u
e
sa
r
ed
e
v
e
l
o
p
e
dp
r
o
g
r
e
s
s
i
v
e
l
ybyc
o
n
s
i
d
e
r
i
n
gt
h
es
t
a
n
d
a
r
dmodelwi
d
ucede
l
e
m
e
n
t
s
.F
o
rexample

t
h
u
n
c
o
n
s
t
r
a
i
n
e
dp
r
o
b
l
e
mwi

lbee
x
p
l
o
df
i
r
st
.The
e
q
u
a
l
i
t
yc
o
n
s
i
n
t
sa
c
o
n
s
i
d
e
r
e
dn
e
x
t
followedbyt
h
ei
n
e
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
t
s
. d
f
i
n
a
l
l
yt
h
ec
o
m
p
l
e
t
emode
.
lT
h
i
sr
e
p
r
e
s
e
n
t
san
a
t
u
r
a
lp
r
o
g
r
e
s
s
i
o
na
sp
r
i
o
rknowledge
e
e
dt
ob
es
a
t
i
s
f
i
e
dbyt
h
es
o
l
u
t
i
o
ni
n
i
su
s
e
dt
od
e
v
e
l
o
pa
d
d
i
t
i
o
n
a
lc
o
n
d
i
t
i
o
n
s n
ei
n
s
t
a
n
c
e
s
.

1
.
1
.
2 MathematlcalModellng

I
nt
h
i
ss
e
c
t
i
o
net
h
r
e
ed
e
s
i
g
np
r
o
b
l
e
m
si
'O
d
u
c
e
de
a
r
l
i
e
rw
i
l
lb
ea
n
s
l
a
t
e
dt
oe
.Them
a
t
h
e
m
a
t
i
c
a
lmod
e
1w
i
l
lf
i
r
s
tb
ei
d
e
n
t
i
f
i
e
d
.Thes
e
c
o
n
dp
r
o
b
l
e
m
s
t
a
n
d
a
r
df
o
r
m
at
r
e
q
u
i
r
e
si
n
f
o
r
m
a
t
i
o
nf
r
o
mac
o
u
r
s
ei
nm
e
c
h
a
n
i
c
sa
n
dm
a
t
e
r
i
a
l
s
.T
h
i
ss
h
o
u
l
db
ew
i
t
h
i
n
.
ep
u
r
v
i
e
wo
fmoste
n
g
i
n
e
e
r
i
n
gs
t
u
d
e
n
t
s
.

Example1
.
1 Newconsumerr
e
s
e
a
r
c
hi
n
d
i
c
a
t
e
sa
tp
e
o
p
l
el
i
k
et
od
r
i
n
ka
b
o
u
t0
.
5
l
i
t
e
ro
fs
o
d
apopa
tat
i
m
ed
u
r
i
n
gt
h
esummerm
o
n
t
h
s
.Thef
a
b
r
i
c
a
t
i
o
nc
o
s
to
ft
h
e
r
e
d
e
s
i
g
n
e
ds
o
d
ac
a
ni
sp
r
o
p
o
r
t
i
o
n
a
l1
0t
h
es
u
r
f
a
c
ea
r
e
aa
n
dc
a
nbee
s
t
i
m
a
l
e
da
l$
1
.
0
0
p
e
rs
q
u
a
r
em
e
t
e
ro
fl
h
emat
r
i
a
lu
s
e
d
.Ac
i
r
c
u
l
a
rc
r
o
s
ss
e
c
t
i
o
ni
st
h
em
o
s
tp
l
a
u
s
i
b
l
e
t
l
1
eh
e
i
g
h
tm
u
s
t
g
i
v
e
nc
u
r
r
e
n
tl
o
o
l
i
n
ga
v
a
i
l
a
b
l
ef
o
rm
a
n
u
f
a
c
l
u
r
e
.F
o
ra
e
s
t
h
e
t
i
cr
e
a
s
o
n
s
l
1
ed
i
a
m
e
t
e
r
.S
t
u
d
i
e
si
n
d
i
c
a
t
ea
th
o
l
d
i
n
gc
o
m
f
o
r
tr
e
q
u
i
r
e
sa
bea
tl
e
a
s
tt
w
i
c
et
d
i
a
m
e
t
e
rb
e
t
w
e
e
n6a
n
d9cm.
F
i
g
u
r
c1
.
1showsas
k
e
t
c
ho
ft
l
1ec
a
n
.I
nm
o
s
tp
r
o
d
u
c
td
e
s
i
g
n
s
p
a
r
t
i
c
u
l
a
r
l
yi
n
e
r
i
n
g
t
h
emodeli
se
a
s
i
e
rt
od
e
v
e
l
o
pu
s
i
n
gaf
i
g
u
r
e
.Th
ed
i
a
m
e
t
e
rda
n
dt
h
e
e
n
g
i
n
r
es
u
f
f
i
c
i
e
n
tt
od
c
s
c
r
i
b
et
l
1es
o
d
ac
a
n
.Whata
b
o
u
tt
h
et
l
1i
c
k
n
e
s
sto
fe
h
e
i
g
h
tha

m
a
t
c
r
i
a
lo
ft
h
ec
a
n
?Whata
r
cl
h
ea
s
s
u
m
p
t
i
o
n
sf
o
rt
h
cd
c
s
i
g
np
r
o
b
l
c
m
?Oneo
ft
h
e
a
s
s
u
m
p
l
i
o
n
sc
o
u
l
dbel
h
a
tti
ss
m
a
l
lcnought
ob
ei
g
n
o
r
e
di
nt
h
cc
a
l
c
u
l
a
t
i
o
no
ft
h
c
a
n
.A
n
o
t
h
e
ra
s
s
u
m
p
t
i
o
ni
st
h
a
tem
a
t
e
r
i
a
l q
u
i
r
e
df
o
rl
h
ec
a
ni
s
v
o
l
u
m
eo
fs
o
d
ai
nec

.'
het
o
pa
n
db
o
t
l
o
mo
ft
h
cc
a
nw
i
l
lb
ef
i
t
t
e
dw
i
t
hc
n
dc
a
p
s
o
n
l
yt
h
ec
y
l
i
n
d
r
i
c
a
ls
u
r
f
a
c
atw
p
r
o
v
i
d
eemechanismbyw
h
i
c
ht
h
es
o
d
ac
a
nb
cpow
. i
sn
o
tp
a
r
to
fi
s
d
c
s
i
g
np
r
o
b
l
e
m
.I
nt
h
eta
t
t
e
m
p
ta
td
c
v
e
l
o
p
i
n
gt
h
em
a
t
h
e
m
a
t
i
c
a
lm
o
d
e
lwec
o
u
l
ds
t
a
r
t
o
u
tbyc
o
n
s
i
d
e
r
i
n
gequ
t
i
t
i
e
si
d
e
n
t
i
f
i
e
d o
v
ea
sd
e
s
1v
a
r
i
a
b
l
e
s
:

h
t
D
e
s
i
g
nv
a
r
i
a
b
l
e
s
:d
R
e
v
i
e
w
i
n
gt
h
es
t
a
l
e
m
e
n
to
ft
h
ed
e
s
i
g
np
r
o
b
l
e
m
o
n
eo
ft
h
ep
a
r
a
m
e
t
e
r
si
st
h
ec
o
s
to
f
sg
i
v
e
na
sl
(
tp
e
rs
q
U
8I c
e
n
t
i
m
e
t
e
r
.

b
i
si
si
d
e
n
l
i
f
i
e
d C
.
m
a
t
e
r
i
a
lp
e
ru
n
i
ta
r
e
a i
D
u
r
i
n
gt
h
es
e
a
r
c
hf
o
rt
h
eo
p
t
i
m
a
ls
o
l
u
t
i
o
ni
sq
u
a
n
t
i
t
yw
i
l
lb
eh
c
l
dc
o
n
s
t
a
n
ta
te
g
i
v
e
nv
a
l
u
e
. Note t
h
a
ti
ft
h
i
sv
a
l
u
cc
h
a
n
g
e
sl
h
e
n ec
o
s
to
ft
h
cc
a
nw
i
l
l
i
si
sw
h
a
twemeanbyad
e
s
i
g
np a
m
e
t
e
r
.T
y
p
i
c
a
l
l
y
c
o
r
r
e
s
p
o
n
d
i
n
g
l
yc
h
a
n
g
e
.Th
c
h
a
n
g
ei
np
a
r
a
m
e
t
e
r
swi

lc
a
u
s
et
h
es
o
l
u
t
i
o
nt
o c
o
m
p
u
t
e
d
.
D
e
s
i
g
np
a
r
a
m
e
t
e
r
:C
Thed
e
s
i
g
nf
u
n
c
t
i
o
n
sw
i
l
li
n
c
l
u
d
eec
o
m
p
u
t
a
t
i
o
no
ft
h
evolumee
n
c
l
o
s
e
dbyt
h
ec
a
n
evolumei
nt
h
ec
a
ni
s
c
f
2
h
1
4
.le
a
n
dt
h
es
u
r
f
a
c
ea
r
e
ao
ft
h
ec
y
l
i
n
d
r
i
c
a
ls
e
c
t
i
o
n
.Th
s
u
r
f
a
c
ea
r
e
ai
s1
t
d
h
.Th
ea
e
s
t
h
c
t
i
cc
o
n
s
t
r
a
i
n
tr
e
q
u
i
r
e
sa
th~ 2
d
.Thes
i
d
ec
o
n
s
a
i
n
t
s
i
a
m
e
t
e
ra
r
ep
r
e
s
c
r
i
b
i
nep
r
o
b
l
e
m
.F
o
rc
o
m
p
l
e
l
n
e
s
st
h
es
i
d
ec
o
n
s
t
r
a
i
n
t
s
on d
t
h
e
rv
a
r
i
a
b
l
e
sh
a
v
et
ob
ep
r
e
s
c
r
i
b
e
dbyt
h
ed
e
s
i
g
n
e
r
.Wec
a
nf
o
r
m
a
l
l
ys
e
tup
on o
t
h
eo
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
e
m
Minimize f(d

ht
)
: C7
r
d
1
z

(
1
.
13
)

2
S
u
b
j
e
c
l
l
O
: hM

h
t
)
: 7
r
d
/
1
/
4-500=0

(
1
.
14
)

INTRODUCTION

(
1
.
16
)

l
(
x
l
Xv:ru/4-500=0
S
u
b
j
e
c
l
l
o
: h

(
1
.
17
)
(
1
.
18
)

6SX)S9
; 5SX2S20

Thep
r
o
b
l
e
m p
r
e
s
e
n
l
e
dbyE
q
u
a
t
i
o
.
n
s(
1
.
16
)
(
1
.
18
)i
s m
a
t
h
e
m
a
t
i
c
a
lmode
f
o
.
r
t
h
d
e
s
i
g
np
r
o
b
l
e
me
x
p
r
e
s
s
e
di
nt
h
es
l
a
n
d
a
r
df
o
r
m
at
.F
o
rt
h
i
sp
r
o
b
l
e
mnoe
x
t
r
a
n
c
o
.
u
s
o
.
nwasn
dt
os
e
tupep
r
o
b
l
e
me
x
c
e
p
tf
o
rsomeg
e
o
m
e
t
r
i
c
a
lr
e
l
a
t
i
o
n
s
.
i
n
f
o
r
m
a
l
i

F
lg
u
r
e1
.
2 E
x
a
m
p
l
e1
.
2
-C
a
n
l
l
l
e
v
e
rb
e
a
m
.

l
n
t
u
i
t
i
v
e
l
y
el
a
r
g
e
rt
h
ev
a
l
u
eo
fF
.t
h
eg
r
e
a
t
e
ri
s c

o
s
8
S
e
c

l
l
i

on

a
l
a
r

=
a
n
ecs v
.
0h

a
n

d
l
ei
t
.1
1
c

a
nb
e
i

l
yc
o
n
c
l
u
d
e
da
t

i
fFw
a

t
e
la
ti
t
sl
o
w
e
rl
i
m
i
ta
8l
h
emasso
ft
h
ebeamw
i
l
lb
ed
i
r
e
c
l
l
yp
r
o
p
o
r
t
i
o
n
a
l1
0e
b s
a
r
e
ao
fc
r
o
S
88
c
c(
Iu
n
.Not
e
c
h
n
i
q
u
e
sa
r
er

u
i
r
e
df
o
.
rl
h
i
sc
o
n
c
l
u
s
i
o
.
n
.Byo
.
u
rd
e
f
i
n
i
t
i
o
n

Fi
sago
o
.
dc
a
n
d
i
d
a
t
ea
sap
a
r
n
e
t
e
rr
a
t
h
e
r a
sad
e
s
i
g
nv
a
r
i
a
b
l
e
.S
i
m
i
l
a
r
l
y
t
h
e

L eg
a
t
e
ri
st
h
emoment Fg
e
n
e
r
a
t
e
sa
b
o
u
tt
h
ec
a
n
t
i
l
e
v
e
r
e
d
l
a
r
g
e
rt
h
ev
a
l
u
eo
f
e
n
d
.l
i
sw
i
l
lr
e
q
u
i
r
ea
ni
n
c
r
e
a
s
ei
nea
r
e
ao
.
fc
r
o
s
s
s
e
c
t
i
o
n
.The
r
e
f
o
r
e
Li
sn
o
ta
o
.
red
e
s
i
g
nv
a
r
i
a
b
l
ee
i
t
h
e
ra
st
h
eb
e
s
tc
h
o
i
c
ef
o
.
ri
tw
i
l
lb
ci
l
o
.
wer
goodc
h
o
i
c
ef
i
I
mit
.Tor
e
p
r
e
s
c
n
tm
a
t
e
r
i
a
la
sad
e
s
i
g
nv
a
r
i
a
b
l
e
.wen
e
c
dt
ou
s
ei
t
sS
l
r
u
C
l
u
r
a
lp
r
o
p
c
r
t
i
e
s
o
r
m
u
l
a
l
i
o
no
f
t
h
ep
r
o
b
l
e
m
.T
y
p
i
c
a
l
l
y
.am
a
l
e
r
i
a
li
sc
h
a
r
a
c
l
e
r
i
z
e
dbyi
t
ss
p
e
c
c
i
nef

Ex
ample1
.2 A c
a
n
t
i
l
e
v
e
rbeamn
e
e
d
s1
0b
ed
e
s
i
g
n
e
dt
oc
r
yap
o
i
n
tl
o
a
dFa
lt
h
e

.
ft
h
ebeamo
.
fl
e
n
g
l
hL
.Th
ec
r
o
s
ss
e
c
t
i
o
no
ft
h
ebeamwi
I
Ib
ei
nt
h
es
h
a
p
eo
ft
h
e
endo
f
e
r
r
e
dt
oa
s 1

b m
)
. l
ebeams
h
o
u
l
dmeetp
r
e
s
c
r
i
b
e
df
a
i
l
u
r
ec
r
i
t
e
r
i
a
.
l
e
t
t
e
r1(
T
h
c
r
ei
sa
l
s
oai
Im
i
toni
t
sd
e
f
l
e
c
t
i
o
.
n
.A beamo
fminimummassi
s q
u
i
r
e
d1
0b
e
d

i
g
n
c
d
.
F
i
g
u
r
e1
.
2showsas
i
d
ev
i
e
wo
.
ft
h
ebeamc
a
r
r
y
i
n
gt
h
el
o
a
dandF
i
g
u
r
e1
.
3at
y
p
i
c
a
l
c
r
o
.
s
ss
n
ofth
b
e
n
.Th
eI
s
h
a
p
e
dc
r
o
s
ss
e
c
t
i
o
ni
ss
y
m
m
e
t
r
i
c
.Th
es
y
m
b
o
l
sd
t
w
.
b

a
n
dr
tco
.
r
r
e
s
p
o
n
dt
ot
h
edep

.
o
ft
h
ebcam
t
h
i
c
k
n
e
s
so
ft
h
eweb

w
i
d
t
ho
ft
h
f
l
a
n
g
e

r
a
n
dt
h
i
c
k
n
e
s
so
.
ft
h
ef
l
a
n
g
e

r
e
s
p
e
c
t
i
v
e
l
y
.Thes
q
u
a
n
t
i
t
i
e
sa
r
es
u
f
f
i
c
i
e
n
tt
od
e
f
i
n
eE
c
r
o
.
s
ss
e
c
t
i
o
n
. A handbo
o
.
ko
rt
e
x
t
b
o
.
ok on s
t
r
c
n
g
t
ho
fm
a
t
e
r
i
a
l
sc
a
na
i
d e
e
l
o
p
m
e
n
to
.
ft
h
d
e
s
i
g
nf
u
n
c
t
i
o
n
s
.Ani
m
p
o
r
t
a
n
ta
s
s
u
m
p
t
i
o
nf
o
rt
h
i
sp
r
o
b
l
e
mi
s
dv
wew
i
l
lbewo
.
r
k
i
n
gw
i
t
h
i
nt
h
ee
l
a
s
t
i
ci
Im
i
to
ft
h
cm
a
t
e
r
i
a
l
wheret
h
e
r
ei
sai
In
e
a
r
r
e
l
a
t
i
o
.
n
s
h
i
pb
e
t
w
e
e
nes
t
r
e
s
sandt
h
es
t
r
a
i
n
.Al
Io
.
ft
h
ev
a
r
i
a
b
l
e
si
d
e
n
t
i
f
i
e
di
nF
i
g
u
r
e
1
.
3w
i
l
ls
n
g
l
ya
f
f
e
c
tt
h
es
o
l
u
t
i
o
.
n
.Th
e
ya
r
ed
e
s
i
g
nv
a
r
i
a
b
l
e
s
.Thea
p
p
l
i
e
df
o
.
r
c
eF
w
i
l
la
l
s
od
i
r
e
c
t
l
ya
f
f
e
c
tt
h
ep
r
o
b
l
e
m
.Sow
i
l
li
t
sl
o
c
a
t
i
o
.
nL
.Ho
.
wa
b
o
u
t m
a
t
e
r
i
a
l
?
o
.
r1
.
0c
o
.
p
p
e
rf
o
rt
h
eb
e
a
m
.S
h
o
u
l
dFLa
n
dt
h
em
a
t
e
r
i
a
l
S
t
e
c
li
sd
c
f
i
n
i
t
e
l
ys
u
p
c
r
i
o
.
p
e
r
t
i
e
sb
ei
n
c
l
u
d
e
da
sd
e
s
i
g
nv
a
r
i
a
b
l
e
s
?
p
r

M
i
n
i
m
i
z
e f
(
X
I
X
2
)
:c:

13

I
V
1
-

(
1
.
15
)

I
n
t
u
i
t
i
v
e
l
yt
h
e
r
ci
ssomeconcemw
i
t
ht
h
ep
r
o
b
l
e
ma
se
x
p
r
e
s
s
e
dby E
q
u
a
t
i
o
n
s
(
1
.
13
(1.15
)e
v
e
nt
h
o
u
g
h d
e
s
c
r
i
p
t
i
o
ni
sv
a
l
i
d
.Howc
a
nt
h
ev
a
l
u
eoft
h
ed
e
s
i
g
n
s
t
a
b
1
is
h
e
d
?Thev
a
r
i
a
t
i
o
ni
ntd
o
e
sn
o
ta
'
e
c
tt
h
ed
e
s
i
g
n
.Changingt
h
e
v
a
r
i
a
b
l
etbee
v
a
l
u
eo
ftd
o
e
sn
o
tc
h
a
n
g
e

fhl

o
rg
lHence

i
tc
a
n
n
o
tb
ead
e
s
i
g
nv
a
r
i
a
b
l
e
.(
N
o
t
e
:t
h
e
.
)l
fi
sw
e
r
cap
r
a
c
t
i
c
a
ld
e
s
i
g
np
r
o
b
l
c
m
t
h
e
n
v
a
r
i
a
t
i
o
ni
ntmaya
f
f
e
c
tt
h
ev
a
l
u
eo
fc
c
a
n
sh
a
v
et
ob
ed
e
s
i
g
n
e
df
o
ri
m
p
a
c
ta
n
ds
t
a
c
k
i
n
gs
t
r
e
n
g
t
h
.I
n C etw
i
l
l
p
r
o
b
a
b
l
ybeac
r
i
t
i
c
a
ld
e
s
i
g
nv
a
r
i
a
b
l
e
.T
h
i
sw
i
l
lr
e
q
u
i
r
es
e
v
e
r
a
d
d
i
t
i
o
n
a
ls
c
l
u
r
a
l
c
o
n
s
t
r
a
i
n
t
si
nt
h
ep
r
o
b
l
e
m
.I
tc
o
u
l
ds
e
r
v
ea
sa
ni
n
t
e
r
e
s
t
i
n
ge
x
t
e
n
s
i
o
nt
ot
h
i
sp
r
o
b
l
e
m
f
o
rhomeworko
rp
r
o
j
e
c
tl
eo
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
e
ma
f
t
e
rd
r
o
p
p
i
n
gt de
x
p
r
e
s
s
i
n
g
[
d
.h
] [
X
I
X
2
]b
e
c
o
m
e
s
:

g
l
(
l

XU:2x
1-X2S0

6Sds9; 5s/
ts
20; O.OOIStsO.Ol

g
M
.
h
.
t
)
: 2d-hSO

OPTIMIZATIONFUNDAMENTALS

1
.
1

12

b
t
I~

'l

t
W

F
l
g
u
r
o1
.
3E
x
a
m
p
l
e1
.
2
-Cr
o
s
s
s
e
c
t
l
o
n
a
ls
h
a
p
e
.

INTRODUCTION

w
e
i
g
h
t y

i
t
sv
a
l
u
eo
f
t
h
emoduluso
fe
l
a
s
t
i
c
i
t
y
.E
.i
t
smoduluso
fr
i
g
i
d
i
t
y
.G
.i
t
sy
i
e
l
d
s
i
o
n
.

y
i
e
'
d
.i

y
i
e
l
di
ns
h

y
i
e
l
d
.i
t
su
l
t
i
m
a
t
es
t
r
e
n
g

l
i
m
i
ti
nt
e
n
s
i
o
na
n
dcompr
i
nt
e
n
s
i
o
n
.

u
l
t
.a
n
di
t
su
l
t
i
m
a
t
es
t
r
e
n
g
t
hi
ns
h
e
a
r
.

u
l
l
O
p
t
i
m
i l
i
o
nt
e
c
h
n
i
q
u
e
si
d
e
n
t
i
f
y
n
tva
1u
e
so
ft
h
ed
e
s
i
g
nv
a
r
i
a
b
l
e
s
.I
fE i
st
h
ev
a
r
i
a
b
l
eu
s
e
dt
o p
r
e
s
e
n
tE
d
i
f
f
e
m
a
t
e
r
i
a
ld
e
s
i
g
nv
a
r
i
a
b
l
ei
nap
r
o
b
l
e
m
.t
h
e
ni
ti
sp
o
s
s
i
b
l
et
h
es
o
l
u
t
i
o
nc
a
n q
u
i
r
ea
1
a
td
o
e
sn
o
te
x
i
st

a
ti
s
.ava
1ueo
fEs
t
i
l
lu
n
d
i
s
c
o
v
e
r
d
.F
o
ra
11i
n
t
e
n
t
sand
m
a
t
e
r
ia
p
p
o
s
e
s
.i
si
sbeyondt
h
es
c
o
p
eo
ft
h
ed
e
s
i
g
n
e
r
.A
g
a
i
n
.m
a
t
e
r
ia
1a
sap
a
r
a
m
e
t
e
r
makesal
o
to
f
.
s
e
n
s
e
andeo
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
e
ms
h
o
u
l
db
er
e
i
n
v
e
s
t
i
g
a
t
e
di
fF
o
rt
h
em
a
t
e
r
i
a
lc
h
a
n
g
e
s
.C
o
n
c
l
u
d
i
n
gf
r
o
mi
sd
i
s
c
u
s
s
i
o
nonm
o
d
e
l
i
n
g
.t
h
ef
o
l
l
o
w
i
n
g
i
Is
tc
a
nbee
s
t
a
b
l
i
s
h
e
df
o
rt
h
i
se
x
a
m
p
l
e
:

hFdf
m
JdJiJdh
hbMUFF

14

D
e
s
i
g
nv
a
r
i
a
b
l
e
s
:

d
t
b'
r
I
w'
f

i!i

I
nt
h
ed
e
v
l
o
p
m
e
n
to
ft
h
em
a
t
h
e
m
a
t
i
c
a
lmodels
t
a
n
d
a
r
dt
e
c
h
n
i
c
a
ld
e
f
i
n
i
t
i
o
n
s
u
s
e
d
. Mucho
ft
h
ei
n
f
o
r
m
a
t
i
o
nc
a
nb
eo
b
t
a
i
n
e
df
r
o
mam
e
c
h
a
n
i
c
a
le
n
g
i
n
e
e
r
i
n
g
h
a
n
d
b
o
o
k
.Th
f
i
r
s
td
e
s
i
g
nf
u
n
c
t
i
o
ni
s w
e
i
g
h
to
ft
h
eb
e
a
m
.T
h
i
si
st
h
ep
r
o
d
u
c
to
f
L
andAc
whereAci
sea
r
e
ao
fc
r
o
s
ss
e
c
t
i
o
n
.Themaximums
l
r
e
s
sduet
ob
e
n
d
i
n
g
y

"
"
n
dc
a
nb
ec
a
l
c
u
l
a
t
e
da
st
h
ep
r
o
d
u
c
to
fFLd
I
2
1
c

whereist
h
cmomenlo
fi
n
e
r
t
i
a
1a
n
g
e
.The
a
b
o
u
tt
h
cc
c
n
t
r
o
i
do
ft
h
ec
r
o
s
ss
e
c
t
i
o
na
l
o
n
gt
h
ea
x
i
sp
a
r
a
l
l
e
lt
ot
h
ef
m imums
h
e
a
rs
t
r
e
s
si
n c
r
o
s
ss
e
c
t
i
o
ni
se
x
p
r
e
s
s
e
da
sF
Q
c
l
l
c
1
w
whercQci
sf
i
r
s
t
i
dp
a
r
a
l
l
c
lt
ot
h
cf
l
a
n
g
e
.Themaximumd
e
f
l
e
c
t
i
o
n(~)
momento
fa
r
e
aa
b
o
u
tt
h
ecen
3
o
ft
h
ebeamwi
1
1bea
tt
h
eendo
ft
h
ebeamc
a
l
c
u
l
a
t
e
dbyt
h
ee
x
p
r
e
s
s
i
o
nFL
/
3
E
l
Ac
'
c
o
ft
h
ec
r
o
s
ss
e
c
t
i
o
no
ft
h
ebeami
s2
b
r
t
r+I
w(
d-2
t
f
)
.Thef
i
r
s
tmomcnto
fa
r
e
aQcc
a
n
d-t
r
)+0
.
5
1
(
0
.
5
d
-t
f
)
2
.
F
i
n
a
l
l
y
.

c
m
o
m
c
n
t
o
f
i
n
e
r
t
i
a

i
s
b
cc
a
l
c
u
l
a
t
e
da
s0
.
5
bf
t
f(
w
o
b
t
a
i
n
e
dfrom(
b
112) (
(
br-t
w)
(
d-2
t
f
)
3
/
1
2
)
.I
n f
o
l
l
o
w
i
n
g
f
o
rc
o
n
v
c
n
i
e
n
c
e
wew
i
l
lc
o
n
t
i
n
u
et
ou
s
eAe
Q
c
'andinstcado
fd
e
t
a
i
1
ingt
h
e
i
rd
e
p
e
n
d
e
n
c
eone
d
e
s
i
g
nv
a
1u
e
s
.A
s
s
o
c
i
a
t
i
n
gx w
i
t
hd
X
2w
i
t
ht
X
3w
i
t
hbr

andX
4w
i
t
hl
r

s
o atX=
w'
[
x
X
2

X
3
'x
4
1
ep
r
o
b
l
e
mi
ns
t
a
n
d
a
r
df
o
r
m
a
ti
s

T
h
i
si
sa
nexamplew
i
t
hf
o
u
rd
e
s
i
g
nv
a
r
i
a
b
l
e
s
t
h
r
e
ei
n
e
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
t
s
.a
n
d
noe
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
t
s
.O
t
h
c
rv
c
r
s
i
o
n
so
ft
h
i
sproblemc
a
ne
a
s
i
l
ybef
o
r
m
u
l
a
t
e
d
.
I
tcanbereducedt
oat
w
o
v
a
r
i
a
b
l
cp
r
o
b
l
c
m
.S
t
a
n
d
a
r
df
a
i
l
u
r
ec
r
i
t
e
r
i
aw
i
t
hr
e
s
p
e
c
t
t
ocombineds
t
r
e
s
s
c
so
rp
r
i
n
c
i
p
a
1s
t
r
c
s
s
e
scanbcf
o
r
m
u
l
a
t
e
d
.I
ft
h
ec
a
n
t
i
l
e
v
e
r
e
d
endi
sb
o
l
t
e
d
t
h
e
na
d
d
i
t
i
o
na
1d
e
s
i
g
nf
u
n
c
t
i
o
n
sr
e
g
a
r
d
i
n
gb
o
l
tf
a
i
l
u
r
en
e
e
dt
ob
e
n
e
d
.
e
x
a
r

ncwComponentP
1
a
c
e
m
e
n
tM
a
c
h
i
n
e
st
om
a
n
u
f
a
c
t
u
r
ed
ie
r
e
n
tk
i
n
d
so
fmo
e
o
a
r
d
s
f
o
ranewg
e
n
e
r
a
t
i
o
no
fp
e
r
s
o
n
a
lc
o
m
p
u
t
e
r
s
.T
h
r
e
em
o
d
e
l
so
ft
h
e
s
em
a
c
h
i
n
e
sa
r
eunder
a
1numberofo
p
e
r
a
t
o
r
sa
v
a
i
l
a
b
l
ei
s1
0
0b
e
c
a
u
s
cofel
o
c
a
ll
a
b
o
r
c
o
n
s
i
d
e
r
a
t
i
o
n
.Tot
.A f
l
o
o
rs
p
a
c
ec
o
n
s
t
r
a
i
n
tn
e
e
d
st
ob
es
a
t
i
s
f
i
e
db
e
c
a
u
s
eo
fed
i
f
f
e
r
e
n
t
market
d
i
m
e
n
s
i
o
n
so
f s
em
a
c
h
i
n
e
s
.A
d
d
i
t
i
o
na
1i
n
f
o
r
m
a
t
i
o
nr
e
l
a
t
i
n
gt
oe
a
c
ho
fe
m
a
c
h
i
n
e
si
sg
i
v
e
ni
nT
a
b
l
e1
.
1
. ecompanyw
i
s
h
e
st
od
e
t
e
r
m
i
n
ehowm yo
f
e
a
c
h
k
i
n
di
sa
p
p
r
o
p
r
i
a
t
et
omaximizct
h
enumbero
fb
o
a
r
d
sm
a
n
u
f
a
c
t
u
dp
e
r
d
a
y
.
l
ti
sd
i
f
f
i
c
u
1
tt
ou
s
eaf
i
g
u
r
ei
nt
h
i
sp
r
o
b
l
e
mt
os
c
tupo
u
rm
a
t
h
e
m
a
t
i
c
a
lmode.
lThe
sw
i
l
ls
e
r
v
ea
so
u
r
numbero
fm
a
c
h
i
n
e
so
fc
a
c
hmodeln
e
c
d
st
obed
e
t
e
r
m
i
n
e
d
.Thi
d
e
s
i
g
nv
a
r
i
a
b
l
e
s
.L
e
tX
Ir
e
p
r
e
s
e
n
tt
h
n
umbero
fComponentP
l
a
c
e
m
e
n
tM
a
c
h
i
n
e
so
f
ModelA
.S
i
m
i
l
a
r
l
y
X
2wi
1
lbea
s
s
o
c
i
a
t
c
dw
i
t
hModelB
andX
3w
i
t
hModelC
.

D i
g
l
lv
a
r
i
u
b
l
e
s
:x
X
2
X
3

Th
ei
n
f
o
n
n
a
t
i
o
ni
nT
a
b
l
e1
.
1i
su
s
e
dt
os
e
tupt
h
ed
e
s
i
g
nf
u
n
c
t
i
o
n
si
nt
e
r
m
so
fe
d
e
s
i
g
nv
a
r
i
a
b
l
e
s
.Ana
s
s
u
m
p
l
i
o
ni
smadet
h
a
ta
l
lm
a
c
h
i
n
e
sa
r
er
u
nf
o
rt
h
r
e
es
h
s
.The
c
o
s
to
fa
c
q
u
i
s
i
t
i
o
no
ft
h
em
a
c
h
i
n
e
si
st
h
esumo
fec
o
s
tp
e
rm
a
c
h
i
n
em
u
1
t
i
p
l
i
e
dby

n
umbero
fm
a
c
h
i
n
e
s(
g1
)
'Th m
a
c
h
i
n
e
smusts
a
t
i
s
f
yef
l
o
o
rs
p
a
c
cc
o
n
s
t
r
a
i
n
t(
8
2
)

1
.2
3
)
.Thec
o
n
s
t
r
a
i
n
tont
h
enumbero
f
whichi
sd
e
f
i
n
e
dw
i
t
h
o
u
td
e
t
a
i
l
si
nE
q
.(
o
p
e
m
t
o
r
si
st
h
r
e
et
i
m
c
s sumo
fep
r
o
d
u
c
to
ft
h
enumbero
fm
a
c
h
i
n
e
so
f c
h
g3
)
'Theu
t
i
l
i
z
a
t
i
o
no
fe
a
c
hm
a
c
h
i
n
ei
st
h
enumber
modela
n
dt
h
eo
p
e
r
a
t
o
r
sp
e
rs
h
i
f
t(
o
fb
o
a
r
d
sp
e
rh
o
u
rt
i
m
e
st
h
enumbero
fh
o
u
r
st
h
em
a
c
h
i
n
eo
p
e
r
a
t
e
sp
e
rd
a
y
.Thc
o
l
l
o
w
i
n
gf
o
r
m
:
o
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
e
mc
a
nb
ea
s
s
e
m
b
l
e
di
n f

f
t
X
)
:Y
L
A
C
Minimize .

(
1
.
19)
Ma
i
m
i
z
ef
(
X
)
:
1
8 x55xxl+18x50x~+2Ix50xx3

x
L

S
u
b
j
e
c
tt
o
: g(
X
)
:F /
2
1
c
-
y
i
e
'
d~ 0
i
e
g
2
(

)
:FQJ1

.
x
:
x
2-
1
:

15

Exam
p
l
e1
.3 MyPCCompanyh
a
sd
e
c
i
d
e
dt
oi
n
v
e
s
t$12m
i
l
l
i
o
ni
na
c
q
u
i
r
i
n
gs
e
v
e
r
a
1

G
.
Y
'e
l
d
'

1
:
l
d
'

u
1
1
'

1
:
E
y
ie

e
t

e
r
s
F.

L
.{
y

.
D
e
s
i
g
np

1
.
1 OPTIMIZA
TIONFUNDAMENTALS

o
r

(
1
.2
0
)

3
g
3

(
X
)
:FL
'/3El
c 3m ~O

1
8x50Xx2-2
1X50xX
3

Minimize f(X):-1
8x55xx

Su
c
tt
o
: g(
X
)
:400
OOOx +6
0
0
.
0
0
0
X
2+700
0
0
0
X
3~ 1
2
0 000

0
.
01~ x ~ 0
.
2
5
; 0
.
0
01~ X
2~ 0
.
0
5
;

0
.
01~ X
3~ 0
.
2
5
; 0
.
0
01~ X
4~ 0
.
0
5
sp
r
o
b
l
e
m
Fmustbcp
r
e
s
c
r
i
b
e
d(
1
0
000N)Lmustb
eg
i
v
c
n(
3m
)
.
Tos
o
l
v
ei
2;
M
a
t
c
r
i
a
1mustbcs
e
l
e
c
t
e
d(
s
t
e
e
1:=7860kglm3
;
O
"
y
i
e
l
d=250E+06N/m
y
i
c
l
d=1
4
5
2)
.A
E+06N/m
1
s
ot
h
emaximumdc
t
1e
c
t
i
o
ni
sp
r
c
s
c
r
i
b
c
d(amn~ =0
.
0
0
5m
)
.

(
1
.2
1
)

g
2
(
X
)
:3x X
2+X
3~ 30

(
1
.2
3
)

X
2+6
x
0
0
8
3
(
X
)
:3
x
I+6
3~ 1

(
1
.2
4
)

X
I~ 0; X
2~ 0; X
)~ 0

(
1
.2
2
)

61'.1.Y:.iK.~

Equ
a
t
i
o
n
s(
1
.21
)

(
1
.2
4
)e
x
p
r
e
s
s m
a
t
h
e
m
a
t
i
c
a
lm
o
d
e
lo
ft
h
ep
r
o
b
l
e
m
.N
o
t
e t
i
n

Eri-522JJyd

1
6 I
N
T
R
O
D
U
C
T
I
O
N

P
T
r
M
I
Z
A
T
I
O
NF
U
N
D
A
M
E
N
T
A
L
S 17
1
.
1O

t
h
i
sp
r
o
b
l
e
mt
h
e
r
ei
sn
op
r
o
d
u
c
tb
c
i
n
gd
c
s
i
g
n
e
d
.H
e
r
eas
t
r
a
t
e
g
yf
o
rp
l
a
c
i
n
go
r
d
e
rf
o
r
n
u
m
b
e
ro
fm
a
c
h
i
n
e
si
sb
e
i
n
gd
c
t
e
r
m
i
n
e
d
.E
q
u
a
t
i
o
n(
1
.
2
1
)i
l
l
u
s
t
h
et
r
a
n
s
l
a
t
i
o
n
o
ft
h
em i
m
i
z
a
t
i
o
no
b
j
e
c
t
i
v
ei
n
t
oa
n u
i
v
a
l
e
n
tm
i
n
i
m
i
z
i
n
go
n
e
.Thei
n
e
q
u
a
l
i
t
y
1.
2
2
)
(
1
.2
4
)a
r
ed
i
f
f
e
r
e
n
tf
r
o
mt
h
ep
r
e
v
i
o
u
st
w
oe
x
n
p
l
e
s
.
c
o
n
s
t
r
a
i
n
t
si
nE
q
u
a
t
i
o
n
s(
t
h
er
i
g
h
t
h
a
n
ds
i
d
eo
ft
h
eSo
p
e
r
a
t
o
ri
sn
o
n
z
e
r
o
.S
i
m
i
l
a
r
l
y
t
h
es
i
d
ec
o
n
s
t
r
a
i
n
t
s
He
r
ei
sas
i
g
n
i
f
i
c
a
n
l
a
r
eb
o
u
n
do
nt
h
el
o
w
e
rs
i
d
eo
n
l
y
.T
h
i
si
sd
o
n
ed
e
l
i
b
e
r
a
t
e
l
y
.The
sp
r
o
b
l
e
ma
n
dt
h
ep
r
e
v
i
o
u
st
w
o
.A
l
lo
ft
h
ed
e
s
i
g
nf
u
n
c
t
i
o
n
sh
e
r
e
d
i
f
f
e
r
e
n
c
eb
e
t
w
e
e
ni
1a
s
s
i
f
i
e
da
sal
i
n
e
a
rp
r
o
g
r
a
r
n
m
i
n
gp
r
o
b
l
e
m
.Thes
o
l
u
t
i
o
n
s
a
r
el
i
n
e
a
r
.T
h
i
sp
r
o
b
l
e
mi
sc
1
0t
h
e
s
et
y
p
c
so
fp
r
o
b
l
e
m
sa
r
ev
e
r
yd
i
f
f
e
r
e
n
tt
h
a
nt
h
cs
o
l
u
t
i
o
n1
0t
h
cf
i
r
s
tt
w
o
s

w
h
i
c
ha
r
er
e
c
o
g
n
i
z
e
da
sn
o
n
l
i
n
a
rp
r
o
g
r
a
m
m
i
n
gp
r
o
b
l
m
s
.
c
x
a
m
p
l

1
.
1
.
3 Natureo
fS
o
l
u
t
i
o
n
l
eec
x
a
r
n
p
I
s
i
ncp
r
e
v
i
o
u
ss
e
c
t
i
o
na
r
cu
s
e
dt
od
i
s
c
u
s
sen
a
t
u
r
eo
fs
o
l
u
t
i
o
n
s

o
n
l
i
n
r
p
r
o
g
r
a
m
m
i
n
g
t
h
eo
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
e
m
.E
x
a
r
n
p
l
e
s1
.
1a
n
d1
.
2d
c
s
c
r
i
b
een
w
h
i
c
hf
o
r
m
st
h
em
a
j
o
rp
a
r
to
fi
sb
o
o
k
.Exa
r
n
p
l
e1
.
3i
sal
i
n
c
a
rp
r
o
g
r
a
m
m
i
n
g
p
r
o
b
l
c
m

w
h
i
c
hi
sv
e
r
ys
i
g
n
i
f
i
c
a
n
ti
nd
c
c
i
s
i
o
ns
c
i
c
n
c
c
s
b
u
tr
a
i
np
r
o
d
u
c
td
e
s
i
g
n
.I
t
s
p
r
o
b
l
e
m
c
1u
s
i
o
nh
e
r
e
w
h
i
l
en
e
c
e
s
s
a
r
yf
o
rc
o
m
p
l
c
t
c
n
e
s
s
i
sa
l
s
oi
m
p
o
r
t
a
n
tf
o
ru
n
d
e
r
s
t
a
n
d
i
n
g
i
n
c
o
n
t
e
m
p
o
r
a
r
yo
p
t
i
m
i
z
a
t
i
o
nt
e
c
h
n
i
q
u
e
sf
o
rn
o
n
l
i
n
e
a
rp
r
o
g
r
a
m
m
i
n
gp
r
o
b
l
c
m
s
.E
x
a
r
n
p
l
e
s
1
.
1a
n
d1
.2a
d
e
a
l
tw
i
t
hf
t
r
S
t
.B
e
t
w
e
e
nt
h
el
w
o
wen
o
t
i
c
et
h
a
tE
x
a
r
n
p
l
e1
.
1i
sq
u
i
t
es
i
m
p
l
e

dt
h
ef
o
u
rv
a
r
i
a
b
l
e
si
nE
x
a
m
p
l
e1
.
2makei
td
i
f
f
i
c
u
l
tt
ou
s
e
r
e
l
a
t
i
v
et
oE
x
a
r
n
p
l
e1
.
2
.S
e
c
o
n
i
1
lu
s
t
r
a
t
i
o
nt
oe
s
t
a
b
l
i
s
hs
o
m
eo
ft
h
cd
i
s
c
u
s
s
i
o
n
.

e
p
r
e
s
e
n
t
a
t
i
o
n0
1E
x
a
m
p
l
e1
.
1
.
F
l
g
u
r
e1
.
4 G~hlI r

EZ

SOLUT10NTOEXAMPLE 1
.
1 Thcs
i
m
p
l
c
s
td
e
t
e
r
m
i
n
a
t
i
o
no
fn
o
n
l
i
n
e
a
r
i
t
yi
s
e
n
t
a
t
i
o
no
fcd
c
s
i
g
nf
u
n
c
t
i
o
n
si
n
v
o
l
v
e
di
nt
h
cp
r
o
b
l
e
m
.
t
h
r
o
u
g
hag
r
a
p
h
i
c
a
l P s
l
i
sc
a
nb
ed
o
n
c s
i
l
yf
o
rO
D
eo
rt
w
ov
a
r
i
a
b
l
e
s
.I
ft
h
ef
U
D
C
t
i
O
Dd
o
e
sn
o
tp
l
o
ta
sa
e
ni
ti
sn
o
n
l
i
n
c
a
r
.F
ig
u
1
.
4s
h
o
w
sat
h
r
e
e
d
i
m
c
n
s
i
o
n
a
lp
l
o
t
s
t
r
a
i
g
h
tl
i
n
eo
rap
l
a
n
e
d
c
t
a
i
l
c
d
o
fE
x
a
m
p
l
c1
.
1
.Thef
i
g
u
r
ei
so
b
t
a
i
n
e
du
s
i
n
gM
A
T
L
A
B
.C
h
a
p
t
e
r2w
i
l
lp
r
o
v
i
d
h
r
e
c
d
i
m
c
n
s
i
o
n
a
l
i
n
s
t
r
u
c
t
i
o
n
sf
o
rd
r
a
w
i
n
gs
u
c
hp
l
o
t
sf
o
rg
r
a
p
h
i
c
a
lo
p
t
i
m
i
z
a
t
i
o
n
.Thet
p
r
e
n
t
a
t
i
o
no
fE
x
a
r
n
p
l
e1
.
1i
nF
i
g
u
r
e1
.
4d
o
e
sn
o
tr
e
a
l
1
ye
n
h
a
n
c
eo
u
ru
n
d
e
r
s
t
a
n
d
i
n
g
.
5i
sa
na
l
t
e
m
a
t
er
e
p
s
c
n
t
a
t
i
o
no
ft
h
ep
r
o
b
l
e
m
.
o
ft
h
ep
r
o
b
l
e
mo
rt
h
es
o
l
u
t
i
o
n
.F
i
g
u
r
e1
n
ee
n
c
o
u
r
a
g
e
di
nt
h
i
sb
o
o
k
.Thch
o
r
i
z
o
n
t
a
l i
sr
e
p
s
c
n
t
st
h
ed
i
a
m
e
t
e
r
T
h
i
si
st
h o
a
n
d v
e
r
t
i
c
a
la
x
i
seh
e
i
g
h
t(
x
u
.1
nF
i
g
u
r
e1
.5t
h
ee
q
u
a
l
i
c
o
n
s
t
r
a
i
n
ti
s
(
X
I
)
r
o
fv
a
l
u
c
so
ni
sl
i
n
ew
i
l
lg
i
v
eav
o
l
u
m
eo
f5
0
0
m
a
r
k
c
da
p
p
r
o
p
r
i
a
t
e
l
ya
sh1
.Anyp
cm3 Thei
n
e
q
u
a
/
i
t
yc
o
n
s
t
r
a
i
n
ti
sshowna
sg1
.Th
ep
a
i
ro
fv
a
l
u
e
so
nt
h
i
sl
i
n
ee
x
a
c1
ty
e
t
i
cr
e
q
u
i
m
ent
.H
a
s
hl
i
n
e
so
nt
h
es
i
d
eo
ft
h
ei
n
e
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
t
s
a
t
i
s
f
i
e
st
h
ca
e
s
i
o
nf
o
rt
h
ed
e
s
i
g
nv
a
r
i
a
b
l
e
s
.l
ec
o
n
s
t
r
a
i
n
t
sa
r
ed
r
a
w
n
c
s
t
a
b
l
i
s
ht
h
ed
i
s
a
l
l
o
w
e
d g
i
s
.Thes
c
a
l
e
d0'
e
c
l
i
v
en
C
l
i
o
ni
sr
e
p
r
en
dt
h
r
o
u
g
hs
e
v
c
r
a
l
t
h
i
c
k
e
rf
o
remph
l
a
b
e
l
c
dc
o
n
t
o
u
r
s
.E
a
c
hc
o
n
t
o
u
ri
sa
s
s
o
c
i
a
t
e
dw
i
t
hax
e
dv
a
l
u
eo
ft
h
eo
b
j
c
c
t
i
v
e
er
a
n
g
eo
ft
h
et
w
oa
x
e
s
f
u
n
c
t
i
o
na
n
dt
h
e
s
ev
a
l
u
e
sa
r
eshowno
nt
h
ef
i
g
u
r
e
.l
l
S
t
r
a
i
n
t
s
.Theo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
n
fa
n
dt
h
ee
q
u
a
l
i
l
yc
o
n
s
t
r
a
i
n
th1
e
s
t
a
b
l
i
s
h
e
st
h
es
i
d
eC
O
l
'
cn
o
n
l
i
n
e
a
rs
i
n
c
ct
h
e
yd
on
o
tp
l
o
tI
I
Ss
t
r
a
i
g
h
tl
i
n
e
s
.R
c
f
e
r
r
i
n
gt
oE
q
u
a
t
i
o
n
s(
1
.
1
6
)a
n
d
t
h
i
si
ss
u
b
s
t
a
n
t
i
a
t
e
db
yt
h
epl'Od
u
C
I
So
ft
h
et
w
ou
n
k
n
o
w
n
s(
d
e
s
i
g
nv
a
r
i
l
l
b
l
e
s
)i
n
(
1
.
1
7
)

5 6

[~

bo
o
ft
h
e
s
cf
u
n
c
t
i
o
n
s
. l
e u
a
l
i
t
yc
o
n
s
t
r
a
i
n
tg
li
sl
i
n
c
a
r
.l
nE
q
u
a
t
i
o
n(
1
.
18
)i
si
s
e
v
i
d
e
n
t
b
ausethedesi variablesappcarbyemselveswithoutbeingraised1
0ap
o
w
c
r
1.Aswed
e
v
e
l
o
pet
e
c
h
n
i
q
u
e
sf
o
ra
p
p
l
y
i
n
go
p
t
i
m
i
z
a
t
i
o
n

i
sd
i
s
t
i
n
c
t
i
o
ni
s
o
i
m
p
o
r
t
a
n
tt
ok
e
e
pi
nm
i
n
d
.I
ti
sa
l
s
os
i
g
n
i
f
i
c
a
n
t g
r
a
p
h
i
c
a
lr
e
p
r
e
s
e
n
t
a
t
i
o
ni
s
e
ev
a
r
i
a
b
l
e
s
wen
e
e
daf
o
u
r
t
hd
i
m
e
n
s
i
o
n
t
y
p
i
c
a
l
l
yr
e
s
t
r
i
c
t
e
dt
ot
w
ov
a
r
i
a
b
l
e
s
.F
o
rr
l
Io
na
n
dt
h
r
e
e
d
i
m
e
n
s
i
o
n
a
lc
o
n
l
o
u
rp
l
o
t
sa
r
en
o
te
a
s
yt
o
1
0r
e
s
o
l
v
et
h
ei
n
f
o
r
m
a
i
I
Iu
s
t
r
a
t
e
.Thep
o
w
e
ro
fi
m
a
g
i
n
a
t
i
o
ni
sn
e
c
e
s
s
a
r
yt
oo
v
e
r
c
o
m
es
eh
u
r
d
l
e
s
.
1
p
r
o
b
l
e
mr
e
p
r
c
s
e
n
t
e
di
nF
i
g
u
r
e1
.
5p
r
o
v
i
d
c
s o
p
p
o
r
t
u
n
i
t
yt
oi
d
e
n
t
i
f
yt
h
eg
r
a
p
h
i
c
a
l

t
h
ef
e
a
s
i
b
l
er
e
g
i
o
ni
si
d
e
n
t
i
f
i
e
d
.I
nF
i
g
u
r
e1
.
5i
si
sE u
a
l
i
t
y
c
o
n
s
i
n
t
s
o
l
u
t
i
o
n
.F
i
r
s
t
i
n
.
tB
o
t
h
f
e
i
b
l
ea
n
do
p
l
i
m
a
lm
u
s
to
c
c
u
rf
r
o
mt
h
i
s :
g
i
o
n
.
a
b
o
v
et
h
ei
n
c
q
u
a
l
i
t
yc
o
n
s
An
ys
o
l
u
t
i
o
nf
r
o
mi
sr
e
g
i
o
ni
sa
na
c
c
e
p
l
a
b
l
e(
s
i
b
l
e
)s
o
l
u
t
i
o
n
.T
h
e
r
ea
r
eu
s
u
a
l
l
ya
i
n
f
i
n
i
l
es
o
l
u
t
i
o
n
s
.I
fo
p
t
i
m
i
z
a
t
i
o
ni
si
n
v
o
l
v
e
d
t
h
c
n
l
a
r
g
en
u
m
b
e
ro
fs
u
c
hs
o
l
u
t
i
o
n
s
t
h
e
s
cc
h
o
i
c
e
sm
u
s
tb
er
c
d
u
c
e
dt
ot
h
eb
e
s
to
n
ew
i
t
h s
p
c
c
tt
osomec
r
i
t
e
r
i
a
t
h
e
o
b
j
e
c
l
i
v
e
.1
nt
h
i
sp
r
o
b
l
e
m
l
f
i
g
u
r
e

t
h
es
m
a
l
l
c
s
tv
a
l
u
eo
f
f
i
sd
e
s
i
r
e
d
.Thel
o
w
e
s
tv
a
l
u
c
o
ffi
sj
u
s
tl
e
s
st
h
a
nt
h
ec
o
n
t
o
u
rv
a
l
u
eo
f3
0
8
.l
ti
sa
tei
n
t
e
r
s
e
c
t
i
o
no
ft
h
et
w
o
c
o
n
s
t
r
a
i
n
t
s
.W
h
i
l
et
h
ev
a
l
u
eo
ffn
e
e
d
st
ob
ec
a
l
c
u
l
a
t
e
dt
h
eo
p
t
i
m
a
lv
a
l
u
e
so
ft
h
e
r
e
a
df
r
o
mt
h
ef
i
g
u
r
e
a
r
ea
b
o
u
t
6
.
7
5a
n
d1
3

5
.r
e
s
p
e
c
t
i
v
e
l
y
.A
n
o
t
h
e
r
d
e
s
i
g
nv
a
r
i
a
b
l
e
s
s
i
g
n
i
f
i
c
a
n
ti
t
e
mo
fi
n
f
o
r
m
a
t
i
o
no
b
t
a
i
n
e
df
r
o
mF
i
g
u
r
e1
.
5i
s g
li
sa
na
c
t
i
v
e
tW
h
i
l
et
h
e
r
ea
r
ei
n
f
i
n
i
t
ef
e
a
s
i
b
l
es
o
l
u
t
i
o
n
s1
0t
h
cp
r
o
b
l
e
m
t
h
eo
p
t
i
m
a
l
c
o
n
s
t
r
a
i
n.
i
q
u
e
.
s
o
l
u
t
i
o
ni
sUlt

18

1
.
1

INTROOUCTION

20

8
3
(
X
)
:3
x
I+6
x
0
2~ 7
X1~O;

Jtub//drvdrfLb
ff/yJX

{EHU}-Z
34

5
6

OPTIMIZATIONFUNOAMENTALS

(
1
.2
8
)

X2~0

Ani
n
t
e
r
e
s
t
i
n
go
b
s
e
r
v
a
t
i
o
ni
nt
h
eaboves
e
to
fe
q
u
a
t
i
o
n
si
sa
tt
h
e
r
ea
r
eo
n
l
y
two d
e
s
i
g
nv
a
r
i
a
b
l
e
sb
u
tt
h
r
e
ec
o
n
s
t
r
a
i
n
t
s
.T
h
i
si
s av
a
l
i
dp
r
o
b
l
e
m
.I
ft
h
e
h
i
s would n
o
tb
e an a
c
c
e
p
t
a
b
l
e
c
o
n
s
t
r
a
i
n
t
s were e
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
t
sent
e
l
a
t
i
o
n
s
h
i
pb
e
t
w
e
e
nt
h
enumbero
f
problemd
e
f
i
n
i
t
i
o
na
si
twouldv
i
o
l
a
t
e er
v
a
r
i
a
b
l
e
s and t
h
e numbero
fe
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
t
s
.T
h
e
r
e
f
o
r
et
h
e numbero
f
i
n
e
q
u
a
l
i

c
o
n
s
t
r
a
i
n
t
sa
r
en
o
tr
e
l
a
t
e
dt
ot
h
enumb
r
o
fd
e
s
i
g
nv
a
r
i
a
b
l
e
su
s
e
d
.T
h
i
s
.6i
sas
c
a
l
e
dp
l
o
to
fc
f
a
c
ta
l
s
oa
p
p
l
i
e
st
on
o
n
l
i
n
e
a
rc
o
n
s
t
r
n
i
n
t
s
.F
i
g
u
r
e1
f
u
n
c
t
i
o
n
s(
1.
2
5
)(
1
.
2
8
)
.Choosingt
h
ef
i
r
s
tq
u
a
d
r
a
n
ts
a
t
i
s
f
i
e
st
h
es
i
d
ec
o
n
s
t
r
a
i
n
t
s
.

'
hel
i
n
e
a
r
i
t
yo
fa
l
lo
ft
h
ef
u
n
c
t
i
o
n
si
si
n
d
i
c
a
t
e
dbyt
h
es
t
r
a
i
g
h
t
l
i
n
ep
l
o
t
sont
h
e
f
i
g
u
r
e
.Again
s
e
v
e
r
a
lc
o
n
t
o
u
r
so
f
f
a
r
eshownt
oa
i
dt
h
er
e
c
o
g
n
i
t
i
o
no
f
t
h
es
o
l
u
t
i
o
n
.
A
l
lo
ft
h
ei
n
e
q
u
a
l
i
t
i
e
sa
r
edrawnw
i
t
hh
a
s
hmarksi
n
d
i
c
a
t
i
n
gt
h
en
o
n
f
e
a
s
i
b
1
es
i
d
e
.
tThef
e
a
s
i
b
l
er
e
g
i
o
ni
se
n
c
l
o
s
e
dbyt
h
etwoa
x
e
sa
l
o
n
gw
i
t
h
o
ft
h
ec
o
n
s
t
r
a
i
n
c
o
n
s
t
r
a
i
n
t
sg2andg3'Theo
b
j
e
c
t
i
v
ei
st
oi
n
c
r
e
a
s t
h
ev
a
l
u
eoff(whichi
si
n
d
i
c
a
t
e
d
byt
h
ec
o
n
t
o
u
rl
e
v
e
l
s
)w
i
t
h
o
u
tl
e
a
v
i
n
gt
h
ef
e
a
s
i
b
l
er
e
g
i
o
n
.Thes
o
l
u
t
i
o
nc
a
nb
e

H137
6
.
5

7
.
5

d
l
a
m
e
l
e
r(
c
m
)

8
.
5

Agure1
.
5C
o
n
l
o
u
rp
l
o
l0
1E
x
a
m
p
l
e1
.
1
.

SOLUTIONTOEXAMPLE1
.
3 I
nc
o
n
t
r
a
s
tt
oExample1
.
1
a
l
lo
ft
h
er
e
l
a
t
i
o
n
s
h
i
p
s
a
t
h
e
m
a
t
i
c
a
lmodele
x
p
r
e
s
s
e
dbyEq
u
a
t
i
o
n
s(
1
.
2
1
)
(
1
.2
4
)a
r
el
i
n
e
a
r
.The
i
n m
q
u
a
t
i
o
ni
sg
e
n
e
r
a
l
l
yu
s
e
dt
od
e
s
c
r
i
b
et
h
ee
q
u
i
v
a
l
e
n
c
eo
ftwoq
u
a
n
t
i
t
i
e
sone
i
t
h
e
r
worde
s
i
d
eo
ft
h
e=
s
i
g
n
.I
no
p
t
i
r
n
i
z
a
t
i
o
nwecomea
c
r
o
s
sm
o
s
t
l
yi
n
e
q
u
a
l
i
t
i
e
s
.eword
e
q
u
a
t
i
o
l
lw
i
l
lbe edt
oi
n
c
l
u
d
eb
o
t
hoft
h
e
s
es
i
t
u
a
t
i
o
n
s
.I
nExample1.1ena

o
fs
o
l
u
t
i
o
nwase
x
p
l
a
i
n
e
du
s
i
n
gF
i
g
u
1.5
.I
nExample1
.3
t
h
ep s
e
n
c
eo
ft
h
r

d
e
s
i
g
nv
a
r
i
a
b
l
e
sd
e
n
i
e
si
sa
p
p
r
o
a
c
h
.I
n
s
t
e
a
do
fd
e
s
i
g
n
i
n
ganewp
r
o
b
l
e
ms
u
p
p
o
s
e
t
h
ecompanyp s
i
d
e
n
th
a
sd i
d
e
dt
obuyf
i
v
em
a
c
h
i
n
e
so
ft
h
et
h
i
r
dt
y
p
e
.The
w
i
t
htwod
e
s
i
g
nv
a
r
i
a
b
l
e
sXla
n
dX2'i
sr
e
c
o
n
s
t
r
u
c
t
e
du
s
i
n
gt
h
e
m
a
t
h
e
m
a
t
i
c
a
lmodel
i
n
f
o
r
m
a
t
i
o
nt
h
a
tX3i
snowap
a
r
a
m
e
t
e
rw
i
t
ht
h
ev
a
l
u
eo
f5
.S
u
b
s
t
i
t
u
t
i
n
gf
o
rx3i
nt
h
e
modelf
o
rExample1
.3

andc
l
e
a
n
i
n
gupt
h
ef
i
r
s
tc
o
n
s i
n:
t

M i
m
i
z
e j
2
5
0
(
X
)
:9
9
0XXI+9 xX2+5

(
1
.2
5
)

S
u
b
j
e
c
tt
o
: g
.
4
x
l+0
.6
x
.
5
l):O
2~ 8

(
1
.2
6
)

g2(X):3
x
I-X2~

25

(
1
.2
7
)

1
9

1
0

1
S

n
u
m
b
e
r0
1M
o
d
e
lA

20

o
n
l
o
u
rp
l
o
l0
1m
o
d
l
l
l
e
dExa
m
p
l
e1
.
3
.
F
l
g
u
r
e1
.
6C

2
5

20

1
.
1 OPT1M1ZAT10NFUNDAMENTALS

lNTRODUCT10N

o
ft
h
es
e
a
r
c
hp
r
o
c
e
s
sf
o
rl
i
n
e
a
rp
r
o
b
l
c
m
sw
i
l
lbee
x
p
l
o
r
e
du
s
i
n
gEx
l
p
l
e1
.3and
.6
.
F
i
g
u
r
e1

s
p
o
t
t
c
da
tt
h
ei
n
t
e
r
s
e
c
t
i
o
no
fc
o
n
s
t
r
a
i
n
t
sg2 dg
3
.Th
ev
a
l
u
e
so
ft
h
ed
e
s
i
g
nv
a
r
i I
c
s
a
tt
h
es
o
l
u
t
i
o
nc
a
nb
er
e
a
do
f
ffromt
h
cp
l
ot
.I
ti
sn
o
to
v
e
ryet
.Beforeproceeding
f
u
r
t
h
e
ri
li
sn
e
c
c
s
s
a
r
yt
oacknowledgct
h
a
tas
o
l
u
t
i
o
no
f6
.
5machinesf
o
rModelB
i
su
n
a
c
c
e
p
t
a
b
l
e
.Thes
o
l
u
t
i
o
nmustb
ea
d
j
u
s
t
e
dt
ot
h
en
c
a
r
e
s
ti
n
t
e
g
e
rv
a
l
u
e
s
.I
n
a
C
l
u
a
lp
r
a
c
t
i
c
e
.t
h
i
sa
d
j
u
s
t
m
c
n
th
a
st
obcmadew
i
t
h
o
u
tv
i
o
l
a
t
i
n
ganyo
ft
h
e
c
o
n
s
t
r
a
i
n
t
s
.Then
e
e
df
o
ri
n
t
e
g
e
rv
a
r
i
a
b
l
e
si
sa
ni
m
p
o
r
t
a
n
tc
o
n
s
i
d
e
r
a
t
i
o
ni
nd
c
s
i
g
n
.
V a
b
l
e
s a
ta
r
q
u
i
r
e
dt
oh
a
v
eo
n
l
yi
n
t
e
g
e
rv
a
l
u
e
sb
e
l
o
n
gt
o et
y
p
eofv
a
r
i
a
b
l
e
s
'
hi
sbookmost
1yc
o
n
s
i
d
e
r
sc
o
n
t
i
n
u
o
u
sv
a
r
i
a
b
l
e
s
.Almosta
11
c
a
l
l
e
dd
i
s
c
r
e
t
ev
a
r
i
a
b
l
e
s
.
re
n
c
o
u
n
t
e
r
se
s
p
e
c
i
a
l
l
ya
c
a
d
e
m
i
c
a
l
l
yb
e
l
o
n
gt
oE
o
ft
h
em
a
t
h
e
m
a
t
i
c
sa
ne
n
g
i
n
domaino
fc
o
n
t
i
n
u
o
u
sv
a
r
i
a
b
l
e
s
.Thca
s
s
u
m
p
t
i
o
no
fc
o
n
t
i
n
u
i
t
yp
r
o
v
i
d
c
sv
e
r
yf
a
s
ta
n
d
aswew
i
l
l
c
f
f
i
c
i
c
n
tt
e
c
h
n
i
q
u
e
sf
o
rt
h
es
e
a
r
c
ho
fc
o
n
s
t
r
a
i
n
e
doptimums
o
l
u
t
i
o
n
s
tManyr
e
a
l
-i
If
cd
e
s
i
g
n
s
e
s
p
e
c
i
a
l
l
yt
h
o
s
et
h
a
tu
s
eo
f
f
t
h
e
s
h
e
l
f
d
e
v
e
l
o
pi
nt
h
i
st
e
x.
e
t
e
r
m
i
n
ed
i
s
c
r
e
t
em
o
d
e
l
s
.D
i
s
c
r
e
t
eprogramminga
n
d
m
a
t
e
r
i
a
l
sa
n
dc
o
m
p
o
n
e
n
t
sd
i
n
t
e
g
e
rprogramminga
r
ed
i
s
c
i
p
l
i
n
e
st
h
a
ta
d
d
r
e
s
st
h
e
s
em
o
d
e
l
s
.Then
a
t
u
r
eo
f
s
e
a
r
c
hm
e
t
h
o
d
sa
ti
sm
e
t
h
o
d
su
s
e
df
o
rs
e
a
r
c
h
i
n
geo
p
t
i
m
a
ls
o
l
u
t
i
o
ni
nd
i
s
c
r
c
t
c
ti
nc
o
n
t
i
n
u
o
u
sp
r
o
g
r
a
m
m
i
n
g
.Manyo
ft
h

p
r
o
g
r
a
r
n
m
i
n
gi
sv
e
r
yd
i
f
f
e
r
e
n
tf
r
o
ma
m
e
t
h
o
d
sa
r
eb
a
s
e
d on t
h
es
c
a
n
n
i
n
g and r
e
p
l
a
c
e
m
e
n
to
ft
h
es
o
l
u
t
i
o
n
st
h
r
o
u
g
h
d
i
s
c t
c
e
x
h
a
u
s
l
i
v
es
e
a
r
c
ho
ft
h
ed
e
s
i
g
ns
p
a
c
ei
nt
h
ed
e
s
i
g
nr
e
g
i
o
n
.I
nt
h
i
sbook
p
r
o
b
l
e
m
sa
r
ct
y
p
i
c
a
l
l
yh
a
n
d
l
e
da
si
ft
h
e
yw
e
r
ec
o
n
t
i
n
u
o
u
sp
r
o
b
l
e
m
s
.Thec
o
n
v
e
r
s
i
o
n
i
s
c
r
c
t
es
o
l
u
t
i
o
ni
st
h
ef
i
n
a
lp
a
r
to
ft
h
ed
e
s
i
g
ne
f
f
o
r
tandi
su
s
u
a
l
l
yp
e
r
f
o
r
m
e
d
t
oed
byt
h
ed
e
s
i
g
n
e
ro
u
t
s
i
d
o
fa
n
yo
p
t
i
m
i
z
a
t
i
o
nt
e
c
h
n
i
q
u
ewed
e
v
e
l
o
pi
nt
h
i
sb
o
o
k
.
i
s
c
u
s
s
e
sd
i
s
c
r
e
t
eo
p
t
i
m
i
z
a
t
i
o
nw
i
t
he
x
a
m
p
l
e
s
.
C
h
a
p
t
e
r8d

d
y
d
p
d
d

huuk xydyM
dr

i
tmustc
o
n
s
i
d
e
rt
h
ef
a
c
tt
h
a
ts
o
l
u
t
i
o
nc
o
u
l
dl
i
ei
n
s
i
d
e
s
o
l
v
e
d
.Forn
o
n
l
i
n
e
a
rp
r
o
b
l
e
m
s
.
5w
i
l
lp
r
o
v
i
d
eec
o
n
t
e
x
tf
o
r
t
h
ef
e
a
s
i
b
l
er
e
g
i
o
n
.Example 1
.
1t
h
r
o
l
l
g
hF
i
g
u
r
c1
d
i
s
c
u
s
s
i
n
gt
h
es
c
a
r
c
hp
r
o
c
e
d
u
r
ef
o
rn
o
n
l
i
n
c
a
rp
r
o
b
l
e
m
s
.Sim
1
i
a
r
I
y
t
h
ec
h
a
r
a
c
t
e
r
i
s
t
i
c
s

l
es
e
a
r
c
hf
o
rt
h
eo
p
t
i
m
a
ls
o
l
u
t
i
o
nw
i
l
ldependonen
a
t
u
r
eo
ft
h
ep
r
o
b
l
e
mb
e
i
n
g

1
.
1
.
4 CharacterlstlcsoftheSearchProcedure

yd
pd

i
n
e
a
rprogrammingp
r
o
b
l
e
mj
u
s
td
i
s
c
u
s
s
e
d
i
ts
h
o
u
l
da
l
s
ob
e
G
e
t
t
i
n
gb
a
c
kt
oel
n
o
t
e
dt
h
a
to
n
l
yt
h
eb
o
u
n
d
a
r
yo
ft
h
ef
e
a
s
i
b
l
er
e
g
i
o
nw
i
l
la
f
f
e
c
tt
h
es
o
l
u
t
i
o
n
.He
i
sa
sa
p
p
a
r
c
n
.
tNotet
h
a
tt
h
eo
b
jc
t
i
v
ef
u
n
c
t
i
o
nc
o
n
t
o
u
r
sa
r
s
t
r
a
i
g
h
t
wayt
h
a
tw
i
l
lmakei
l
i
n
e
s
.Ad
i
f
f
e
r
e
n
to
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
nw
i
l
lb
ed
i
s
p
l
a
y
e
dbyp
a
r
a
l
l
e
l
l
i
n
e
sw
i
t
had
i
f
f
e
r
e
n
t
in
e
sont
h
ef
i
g
u
r
ei
n
s
t
e
a
do
ft
h
ea
c
t
l
l
a
lo
b
j
e
c
t
i
v
e
.Notea
tt
h
e
s
l
o
p
e
.I
m
a
g
i
n
et
h
e
s
eI
s
o
l
u
t
i
o
na
l
w
a
y
sa
p
p
e
a
r
st
ol
i
ea
tt
hi
n
t
c
r
s
e
c
t
i
o
no
ft
h
ec
o
n
s
t
r
a
i
n
t
s
t
h
a
ti
s
t
h
c
o
r
n
e
r
s
s
i
g
ns
p
a
c
c
.S
o
l
u
t
i
o
nc
a
nn
e
v
c
rcom
f
romi
n
s
i
d
eo
ft
h
ef
c
a
s
i
b
l
e g
i
o
n
u
n
l
i
k
e
ofthed
o
l
u
t
i
o
n on t
h
e
i
nt
h
en
o
n
l
i
n
e
a
rprogramming p
r
o
b
l
e
mwhich accommodates s
b
o
u
n
d
a
r
ya
n
df
r
o
mw
i
t
h
i
nt
h
ef
e
a
s
i
b
l
er
e
g
i
o
n
.R
e
c
o
g
n
i
z
i
n
gt
h
cn
a
t
u
o
fs
o
l
u
t
i
o
n
si
s
i
m
p
o
r
t
a
n
ti
nd
e
v
e
l
o
p
i
n
gas
e
a
r
c
hp
r
o
c
e
d
u
r
e
.F
o
rexamplei
nl
i
n
e
a
rp
r
o
g
r
a
r
n
m
i
n
g
p
r
o
b
l
e
m
s
at
e
c
h
n
i
q
u
eemploycdt
os
e
a
r
c
hf
o
rt
h
eo
p
t
i
m
a
lds
i
g
nn
e
e
do
n
l
ys
e
a
r
c
ht
h
c
a
r
t
i
c
u
l
a
r
l
yt
h
ei
n
t
c
r
s
e
c
t
i
o
no
ft
h
eb
o
u
n
d
a
r
i
e
sw
h
e
r
e
a
s n
o
n
l
i
n
e
a
r
b
o
u
n
d
a
r
i
e
sp
t
h
es
e
a
r
c
hp
r
o
c
e
d
u
r
ec
a
n
n
o
ti
g
n
o
r
ei
n
t
e
r
i
o
rv
a
l
u
e
s
.
p
r
o
b
l
e
m
s

2
1

N
o
n
l
l
n
e
a
rProb/ems:Exceptf
o
raf
e
wc
1a
s
s
e
so
fp
r
o
b
l
c
m
s
s
o
l
u
t
i
o
n
st
on
o
n
1
i
n
e
a
r
p
r
o
b
l
e
m
sa
r
eo
b
t
a
i
n
e
dt
h
r
o
u
g
hn
u
m
e
r
i
c
a
la
n
a
l
y
s
i
s
. Through c
o
m
p
u
t

rc
o
d
e
n
u
m
e
r
i
c
a
1a
n
a
l
y
s
i
sbecomesn
u
m
c
r
i
c
a
lt
e
c
h
n
i
q
u
e
s
.Them
e
t
h
o
d
so
rt
e
c
h
n
i
q
u
e
sf
o
r
f
i
n
d
i
n
gt
h
es
o
l
u
t
i
o
nt
oo
p
t
i
m
i
z
a
t
i
o
np
r
o
b
1
e
m
sa
r
ec
a
l
l
e
ds
e
a
r
c
hm
e
t
h
o
d
s
.I
na
p
p
l
i
e
d
m
a
t
h
e
m
a
t
i
c
sa
n
dn
u
m
e
r
i
c
a
lt
h
n
i
q
u
e
st
h
e
s
e r
e
f
e
r
r
e
dt
oa
si
t
e
r
a
t
i
v
et
e
c
h
n
i
q
u
e
s
.
Im
p
l
i
c
it
1yt
h
i
smeanss
e
v
e
m
lt
r
i
e
sw
i
l
lben
e
c
e
s
s
a
r
yb
e
f
o
r
ct
h
es
o
l
u
t
i
o
nc
a
nb
e
o
b
t
a
i
n
e
d
.A
1
s
oi
m
p
l
i
e
di
st
h
a
tc
a
c
ht
r
yo
rs
e
a
r
c
hi
se
x
e
c
u
t
e
di
nac
o
n
s
i
s
t
e
n
tm
a
n
n
e
r
.
I
n
f
o
r
m
a
t
i
o
nf
r
o
m p
r
e
v
i
o
u
si
t
e
r
a
t
i
o
ni
su
t
i
l
i
z
e
di
nt
h
ec
o
m
p
u
t
a
t
i
o
no
fv
a
l
u
e
si
nt
h
e
essbywhichthesearchi
sc
a
r
r
i
e
do
u
t
p
r
e
s
e
n
ts
e
q
u
e
n
c
e
.T
h
i
sc
o
n
s
i
s
t
c
n
tmannero
rp
r
and s
o
l
u
t
i
o
nd
e
t
e
r
m
i
n
e
di
sc
a

1e
d a
l
g
o
r
i
t
h
m
.T
h
i
st
e
r
ma
1
s
or
e
f
e
r
st
ot
h
e
s
l
a
t
i
o
no
ft
h
ep
a
r
t
i
c
u
l
a
rs
e
a
r
c
hp
r
o
c
e
d
u
r
ei
n
t
oa
no
r
d
e
r
e
ds
u
e
n
c
eo
fs
t
e
p
b
y

s
t
e
p
t
r
ea
d
v
a
n
t
a
g
ei
sd
e
r
i
v
e
dm
o
s
t
l
yi
nt
h
ec
o
n
v
e
r
s
i
o
no
ft
h
e
s
es
t
e
p
s
t
h
r
o
u
g
h
a
c
t
i
o
n
s
.Th
somec
o
m
p
u
t
e
r1
a
n
g
u
a
g
ei
n
t
oc
o
d
ea
twi
1
le
x
e
c
u
t
eonec
o
m
p
u
t
c
r
.D
e
v
e
l
o
p
i
n
gE
a
1g
o
r
i
t
h
ma
n
dt
r
a
n
s
1
a
t
i
n
gi
ti
n
t
oc
o
d
et
h
a
twi
l
e
x
e
c
u
t
ei
nMATLABwi
l
beas
t
a
n
d
a
r
d
a
p
p
r
o
a
c
hi
ni
sb
o
o
k
.
Thes
e
a
r
c
hp
r
o
c
e
s
si
ss
t
a
r
t
e
dt
y
p
i
c
a
1ybyt
r
y
i
n
gt
og
u
e
s
st
hi
n
i
t
i
a
ld
e
s
i
g
ns
o
l
u
t
i
o
n
.
Thed
e
s
i
g
n
e
rc
a
nb
a
s
et
h
i
ss
e
l
e
c
t
i
o
nonh
i
se
x
p
e
r
i
e
n
c
e
.Ons
e
v
c
m
1o
c
c
a
s
i
o
n
st
h
e
s
u
c
c
e
s
so
fo
p
t
i
m
i
z
a
t
i
o
nmayh
i
n
g
eonh
i
sa
b
i
l
i
t
yt
oc
h
o
o
s
eagoodi
n
i
t
i
a
ls
o
l
u
t
i
o
n
.The
e
v
e
nwhenu
s
e
dc
o
n
s
i
s
t
e
nt
1
y
r
e
s
p
o
n
d
sd
i
f
f
e
r
e
n
t
1yt
ot
h
ep
a
r
t
i
c
u
l
a
r
s
e
a
r
c
hmethod
p
r
o
b
l
e
ma
ti
sb
e
i
n
gs
o
l
v
e
d
.Thed
e
g
r
e
eandt
y
p
eo
fn
o
n
l
i
n
e
a
r
i
t
ymayf
r
e
q
u
e
nt
1y
c
a
u
s
e e method t
of
a
i
la
s
s
u
r
n
i
n
gt
h
e
r
ea
r
e no i
n
c
o
n
s
i
s
t
e
n
c
i
e
si
np
r
o
b
l
e
m
fd
e
s
i
g
nv
a
r
i
a
b
l
e
senumbero
fc
o
n
s
i
n
t
sa
swe
1
f
o
r
m
u
1
a
t
i
o
n
.O
f
t
e
n numbero
a
s ab
a
di
n
i
t
i
a
lg
u
e
s
sp
1
a
y ap
a
r
ti
nu
n
e
x
p
e
c
t
e
db
e
h
a
v
i
o
ru
n
d
e
rs
t
a
n
d
a
r
d
i
m
D
l
e
m
e
n
t
a
t
i
o
no
ft
h
em
e
t
h
o
d
.T
h
i
sh
a
sl
e
dt
ot
h
ec
r
e
a
t
i
o
no
fs
e
v
e
m
ld
i
f
f
e
r
c
n
t
t
e
c
h
n
i
q
u
e
sf
o
ro
p
t
i
m
i
z
a
t
i
o
n
.P
r
a
c
t
i
c
a
1e
x
p
e
r
i
e
n
c
eb
a
sa
1
s
os
u
g
g
e
s
t
e
d a
tac
l
a
s
so
f
p
r
o
b
l
e
m
sr
e
s
p
o
n
dw
e
l
lt
oc
e
r
t
a
i
na
1
g
o
r
i
t
h
m
s
.
e
a
r
c
h
o
ri
t
m
t
i
v
em
e
t
h
o
d
s
.a
r
ec
u
r
r
e
nt
1yt
h
eo
n
l
ywayt
o
Them
c
t
h
o
d
sf
o
res

i
r
r
e
s
p
e
c
t
i
v
eo
ft
h
ed
i
s
c
i
p
l
i
n
ei
n
i
m
p
l
e
m
e
n
tn
u
m
e
r
i
c
a
ls
o
l
u
t
i
o
n
st
on
o
n
l
i
n
e
a
rp
r
o
b
l
e
m
s
w
h
i
c
ht
h
e
s
ep
r
o
b
l
e
m
sapp
r
.1 ep
r
o
b
l
e
m
sa
r
ep
1
e
n
t
i
f
u
1i
nd
e
s
i
g
no
p
t
i
m
i
z
a
t
i
o
n
.
o
p
t
i
m
a
lc
o
n
t
r
o
1
.s
t
r
u
c
t
u
r
a
la
n
dt
1u
i
dm
e
c
h
a
n
i
c
s
.quantump
h
y
s
i
c
s
.a
n
da
s
t
r
o
n
o
m
y
.
Wh
i
1
eo
p
t
i
m
i
z
a
t
i
o
nt
e
c
h
n
i
q
u
e
sa
r
ea
p
p
l
i
e
d a
l
lo
ft
h
en
o
t
e
dd
i
s
c
i
p
l
i
n
e
s
.i
n d
i
t
i
o
n
a
l
d
e
s
i
g
n

22

INTRODUCTION

1
.
1 OPTIMIZATIONFUNDAMENTALS

2
0
:
r
;
:
.

2
3

a
tad
i
s
t
a
n
s
S
.
.T
h
isp
o
i
n
ts
a
t
i
s
f
i
e
st
h
ec
o
n
s
t
r
a
i
n
th
.
.Ai
r
ds
Yw
o
u
l
db
et
o

T
hi
sl
i
e
so
nt
h
ec
o
n
s
t
r
a
i
n
tb
o
u
n
d
a
r
yo
fg
.
.l
ti
st
ob
en
o
t
e
dt
h
a
ta
l
lE
c
h
o
o
s
cX
.
c
h
o
i
c
e
sa
r
en
o
tf
e
a
s
i
b
l
e
.Whichp
o
i
n
ti
sc
h
o
s
e
ni
sd
c
c
i
d
e
db
yt
h
er
u
l
e
so
fap
a
r
t
i
c
u
l
a
r
sd
i
s
c
u
s
s
i
o
n
.
a
l
g
o
r
i
t
h
m
.Assumep
o
i
n
tX
)i
st
h
en
e
x
tp
o
i
n
ti
ni
o
i
n
t Anew hd
i
r
c
c
t
i
o
n~ i
sd
e
n
n
i
n
.
P
o
i
n
tX
)nowb
e
c
o
m
e
s newp
P
o
i
n
tX2i
s1
0
d
.Ano
t
h
S
u
e
n
c
eo
fc
a
l
c
u
l
a
t
i
o
n
slocat X3 le n
e
x
ti
t
e
r
a
t
i
o
ny
i
e
l
d
s
es
o
l
u
t
i
o
n
.Thee
s
s
e
n
c
eo
fu
p
d
a
t
i
n
ged
e
s
i
d
u
r
i
n
gs
u
c
s
s
i
v
ei
t
e
r
a
t
i
o
nl
i
e
si
nt
w
o
t
h
cs
e
a hd
i
r
e
c
t
i
o
l
l8
/a
n
ds
t
e
p
s
i
z
e
.l
Cc
h
a
n
g
ei
ned
e
s
i
g
n
m
a
j
o
rc
o
m
p
u
t
a
t
i
o
n
s

d
e
f
i
n
I
a
s.
6
X
w
h
i
c
hr
e
p
r
e
s
e
n
t
s v
e
c
l
o
rd
i
f
f
c
r
e
n
c
e

f
o
re
x
n
p
l
e

X2- X
.
i
s
v
e
c
t
o
r
u
e
n
c
eo
fa
c
t
i
v
i
t
ycanb
ei
d
e
n
t
i
f
i
e
d
:
o
b
t
a
i
n
e
da
s

S
/
.F
o
re
a
c
hi
t
e
r
a
t
i
o
nt
h
ef
o
l
l
o
w
i
n
gs

x
o

{ESEmgz

:
8
t
e
p0
S
t
e
p1
:

ChooseX
o
I
d
e
n
t
i
f
yS
D
e
t
e
r
m
i
n
e

AX=S

XM W=
Xo+.6X
S
e
tX
oXn
e
n
GoToStepO

7
.
5

d
l
a
m
o
l
e
r(
c
m
)

8
.
5

F
J
g
u
r
e1.7 At
y
p
i
c
a
ls
s
a
r
c
hm
e
t
h
o
d
.

x
o

i
n
i
t
i
a
ls
t
a
r
t
i
n
g e
s
si
sep
o
i
n
tm
a
r
k
e
d Remember whens
o
l
v
i
n
g
o
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
e
m
.suchaf
i
g
u
r
edoesn
o
t i
s
t
.I
fi
td
i
d

e
nwec
a
ne
a
s
i
l
yo
b
t
a
i
n
i
ss
t
a
t
e
m
e
n
ti
se
s
p
c
c
i
a
l
l
yt
r
u
ei
f e
r
ea
r
emoret
h
a
nt
w
o
t
h
es
o
l
u
t
i
o
nb
yi
n
s
p
e
c
t
i
o
n
.l
nt
h
ee
v
e
n
tt
h
a
r
et
w
ov
a
r
i
a
b
l
e
s

i
sn
or
e
a
s
o
nn
o
tt
oo
b
t
a
i
na
v
a
r
i
a
b
l
e
s
.I
g
r
a
p
h
i
c
a
ls
o
l
u
t
i
o
n(
C
h
a
p
t
e
r2
)
.Thcm
c
t
h
o
d
si
no
p
t
i
m
i
z
a
t
i
o
nr
e
l
yo
nt
h
cd
i
r
c
c
t
p
o
l
a
t
i
o
no
ft
h
ec
o
n
c
e
p
t
sa
n
di
d
e
a
sd
e
v
e
l
o df
o
rt
w
ov
a
r
i
a
b
l
e
st
oa
n
yn
u
m
b
e
ro
f
e
x
d
e
s
i
g
nv
a
r
i
a
b
l
e
s
.T
h
a
ti
st
h
er
e
a
s
o
nwew
i
l
lu
s
et
w
ov
a
r
i
a
b
l
e
st
od
i
s
c
u
s
sa
n
dd
e
s
i
g
n

u
s
i
n
gt
w
ov
a
r
i
a
b
l
e
s

c
o
n
c
e
p
t
sc
a
n
m
o
s
to
ft
h
ea
l
g
o
r
i
t
h
m
sa
n
dm
e
t
h
o
d
s
.M
o
r
e
o
v
e
r
du
s
i
n
gg
e
o
m
e
t
r
y
.Thed
e
s
i
g
na
tt
h
ep
o
i
n
tXoi
sn
o
tf
e
a
s
i
b
l
e
.l
ti
sn
o
t
a
l
s
ob
er
e
i
n
f
o
r
u
s
e
f
u
lt
od
i
s
c
u
s
sw
h
e
t
h
e
ri
ti
sa
no
p
t
i
m
a
ls
o
l
u
t
i
o
n
.A
tl
e
a
s
to
n
emorei
t
e
r
a
t
i
o
ni
s
n
e
c
e
s
s
a
r
yt
oo
b
nt
h
es
o
l
u
t
i
o
n
.C
o
n
s
i
d
e
ren
e
x
ts
o
l
u
t
i
o
na
tXI
.
M
o
s
tm
e
t
h
o
d
si
n
o
p
t
i
m
i
z
a
t
i
o
nw
i
l
la
t
t
e
m
p
tt
omovet
oXI
b
yf
i
r
s
ti
d
e
n
t
i
f
y
i
n
gas
e
a
r
c
hd
i
r
e
c
t
i
o
n
S1T
h
i
s
e
c
t
o
r(
s
a
m
ed
i
m
e
n
s
i
o
na
sd
e
s
i
g
nv
a
r
i
a
b
l
ev
c
c
t
o
rX)p
o
i
mX
oloXI
i
sav
e
oXI
c
a
l
li
tXo
X
.
.i
ssomem
u
l
t
i
p
l
eo
fS
I
's

S
IX10
n
c
e
a
c
t
u
a
ld
i
s
t
a
n
c
ef
r
o
m t

i
n
c
ei
ts
a
t
i
s E
a
g
a
i
ni
sn
o
tf
e
a
s
i
b
l
ea
l
t
b
o
u
g
hi
ti
sa
ni
m
p
r
o
v
e
m
e
n
to
n s
i
n
c
q
u
a
l
i
l
yc
o
n
s
t
r
a
i
n
lg
l(
ly
i
n
gi
nl
h
ei
n
l
e
r
i
o
ro
fec
o
n
s
t
r
a
i
n
t
)
.A
l
t
e
m
a
t
e
l
y
u
s
i
n
gl
h
e
s
a
m
es
c
a
l
'
c
hd
i
r
c
c
t
i
o
nS1
a
n
o
t
h
e
rt
e
c
h
n
i
q
ue
J
me
o
dmayc
h
o
o
s
ef
o
ri
t
sn
e
x
tp
o
i
n
t
X1
'

x
o

x
o

Thea
b
o
v
ei
st
h
eb
a
s
i
cs
t
r
u
c
t
u
r
eo
fat
y
p
i
c
a
lo
p
t
i
m
i
z
a
t
i
o
na
l
g
o
r
i
t
h
m
.Ac
o
m
p
l
e
t
e
e
ri
nw
h
i
c
hSa
n
darec
o
m
p
u
t
e
d
.I
lwi
l
1i
nc
Ju
d
et
e
s
t
s
o
n
ew
i
l
li
n
d
i
c
a
t
e m n
w
h
i
c
hw
i
l
ld
e
t
e
r
m
i
n
ci
f s
o
l
u
t
i
o
nh
a
sb
c
e
nr
e
a
c
h
e
d(
c
o
n
v
e
r
g
e
n
c
e
)o
ri
ft
h
e
e
x
p
l
o
r
c
d
p
r
o
c
e
d
u
r
en
e
e
d
st
ob
es
u
s
p
e
n
d
e
do
rr
e
s
t
a
r
t
e
d
.Thed
i
f
f
c
r
e
n
tt
e
c
h
n
i
q
u
e
st
ob
se
s
t
a
b
l
i
s
h
e
d
.T
a
k
eal
o
o
ka
tF
i
g
u
r
e1
.7a
g
a
i
n
.
i
nt
h
i
sb
o
o
km
o
s
t
l
yd
i
f
f
e
ri
newaySi
r
ei
n
f
i
n
i
t
ec
h
o
i
c
e
sf
o
rt
h
es
e
a
r
c
hd
i
r
e
c
t
i
o
n
.Tor
e
a
c
ht
h
es
o
l
u
t
i
o
ni
n
A
tXot
h
e
r
ea
o
n
a
b
l
et
i
m
ea
na
c
c
e
p
t
a
b
l
ea
l
g
o
r
i
t
h
mw
i
l
lt
r
yn
o
tt
os
e
a
r
c
ha
l
o
n
gd
i
r
e
c
t
i
o
n
s d
o
e
r
ea
r
emanyw
a
y
st
oa
c
c
o
m
p
l
i
s
hi
ss
t
r
a
t
e
g
y
.
n
o
ti
m
p
r
o
v
et
h
ec
u
r
r
e
n
td
e
s
i
g
n
.l

L
i
nearProgrammingProblem:I
tw
a
sp
r
e
v
i
o
u
s
l
ym
e
n
t
i
o
n
e
d t
h
es
o
l
u
t
i
o
nt
o
s
ep
r
o
b
l
e
m
sw
o
u
l
db
eo
nt
h
eb
o
u
n
d
a
r
yo
f f
e
a
s
i
b
l
e g
i
o
o(
a
l
s
oc
a
l
l
e
dt
h
e
j
i
b
l
ed
o
m
a
i
n
)
.I
nF
i
g
u
r
e1
.
6
e e
sa
n
dsomeo
ft
h
ec
o
n
s
i
n
t
sd
e
t
e
n
n
i
n
et
h
i
s
o
i
n
t
so
fi
n
t
e
r
s
c
c
t
i
o
no
ft
h
ec
o
n
s
t
r
a
i
n
L
s
r
e
g
i
o
n
.Thel
a
l
t
e
ri
sd
i
s
t
i
n
g
u
i
s
h
e
da
l
s
ob
yep

r
i
g
i
na
n
dei
n
t
e
r
s
e
c
t
i
o
no
ft
h
eb
i
n
d
i
n
g(
a
c
t
i
v
e
)c
o
o
s
t
r
a
i
n
t
s
amongt
h
e
m
s
e
l
v
e
seo
w
i
t
ht
h
ea
x
e
s
.T
h
e
s
ea
r
ev
e
r
t
i
c
e
so
rc
o
r
n
e
r
so
ft
h
eq
u
a
d
r
i
l
a
t
e
r
a
l
.E
s
s
e
n
t
i
a
l
l
yt
h
ed
e
s
i
g
n
i
m
p
r
o
v
e
sb
ym
o
v
i
n
gf
r
o
mo
n
eo
ft
h
e
s
ev
e
r
t
i
c
e
st
ot
h
en
e
x
to
n
et
h
r
o
u
g
hi
m
p
r
o
v
i
n
g
o
m
e
r
.T
h
i
sw
i
l
lb
ee
x
p
l
o
r
e
d
f
e
a
s
i
b
l
ed
e
s
i
g
n
s
.Tod
ot
h
a
ty
o
uh
a
v
et
os
t
a
r
ta
taf
e
a
s
i
b
lc
ti
st
ob
ee
m
p
h
a
s
i
z
e
d a
t
o
n
l
yw
i
t
h
m
o
r
ef
u
l
l
yi
nt
h
ec
h
a
p
t
e
ro
nl
i
n
e
a
rp
r
o
g
r
a
m
m
i
n
g
.I
I
Iu
s
t
r
a
t
i
o
no
ft
h
et
e
c
h
n
i
q
u
e i
s
t
w
ov
a
r
i
a
b
l
e
sc
a
nwea
c
t
u
a
l
l
ys
e
et
h
eg
e
o
m
e
t
r
i
ci
t
h
es
e
l
e
c
t
i
o
no
fv
e
r
t
i
c
e
sa
st
h
es
o
l
u
t
i
o
nc
o
n
v
e
r
g
e
s
.
T
h
i
sp
a
r
to
fC
h
a
p
t
e
r1h
a
sd
e
a
l
tw
i
t
ht
h
ei
n
t
r
o
d
u
c
t
i
o
no
ft
h
eo
p
t
i
o
np
r
o
b
l
e
m
.
Whi
l
ei
tw
a
sa
p
p
r
o
a
c
h
e
df
r
o
ma
ne
n
g
i
n
e
e
r
i
n
gd
e
s
i
g
np
e
r
s
p
c
c
t
i
v
e
t
h
em
a
t
h
e
m
a
t
i
c
a
l
m
o
d
e
li
na
b
s
t
r
a
c
tt
e
r
m
s
i
sn
o
ts
e
n
s
i
t
i
v
et
oa
n
yp
t
i
c
u
l
a
rd
i
s
c
i
p
l
i
n
e
.Af
e
wd
e
s
i
g
n
t
t
os
p
e
c
i
f
i
ce
x
a
m
p
l
e
sw
e
r
ee
x
a
m
i
n
e
d
.T
I
l
es
t
a
n
d
a
r
dm
a
t
h
e
m
a
t
i
c
a
l
r
u
l
e
sw
i
t
hr
e
s
p
Ja
s
s
e
so
fp
r
o
b
l
e
m
s
m
o
d
e
lf
o
ro
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
e
m
swase
s
t
a
b
l
i
s
h
e
d
.Twob
r
o
a
dc

.
1
"
"
'
:
.
.
.
.
.
.

1
:

n'

'

1
.
2

2
4 I
N
T
R
O
D
U
C
T
I
O
N

25

1
.
2 INTRODUC
T
l
ONTOMATLAB
MATLABi
si
n
t
r
o
d
u
c
c
db
yMathWorksI
n
c
.a
st
h
el
a
n
g
u
a
g
ef
o
rt
e
c
h
n
i
c
a
lc
o
m
p
u
t
i
n
g
.
n
eo
fi
t
sm
a
n
u
a
l
sMATLABi
n
t
e
g
r
a
t
e
s
B
o
r
r
o
w
i
n
gf
r
o
mt
h
ed
e
s
c
r
i
p
t
i
o
ni
n yo
mputationvisualizatio dprogrammingi
na
ne
a
s
y
t
ou
s
ee
n
v
i
r
o
n
m
e
n
tw
h
e
r
e
p
r
o
b
l
e
m
sa
n
ds
o
l
u
l
i
o
n
sa
r
ee
x
p
r
e
s
s
e
di
nf
a
r
n
i
l
i
a
rm
a
t
h
e
m
a
t
i
c
a
ln
o
t
a
t
i
o
n
.Thet
y
p
i
c
a
l
T
I
.
.
A
Di
n
c
l
u
d

u
s
e
sf
o
rMA

JA22// J1

.T
hegeome
i
c
a
lc
h
a
r
a
c
t
e
r
i
s
t
i
c
so
ft
h
es
o
l
u
t
i
o
n
l
i
n
e
a
ra
n
dn
o
n
l
i
n
e
a
r
.w
e
r
ei
n
t
r
o
d
u
c
f
o
rt
h
e
s
cp
r
o
b
l
e
m
sw
e
r
es
h
o
w
nr
o
u
g
h g
r
a
p
h
i
c
a
ld
e
s
c
r
i
p
t
i
o
no
fep
r
o
b
l
e
m
s
.An
o
v
e
r
v
i
e
wo
fes
e
a
r
c
ht
e
c
h
n
i
q
u
e
sf
o
ro
b
t
a
i
n
i
n
gt
h
es
o
l
u
t
i
o
nt
ot
h
e
s
ep
r
o
b
l
e
m
sw

I
Iu
s
t
r
a
t
e
d
.I
tw
a
sa
1
s
om
e
n
t
i
o
n
e
da
tes
o
l
u
t
i
o
n
sw
e
r
et
ob
eo
b
t
a
i
n
e
d
a
l
s
oi
n
u
m
e
r
i
c
a
l

.p
r
i
m
a
r
i
l
yt
h
r
o
u
g
hd
i
g
i
t
a
lc
o
m
p
u
t
a
t
i
o
n
.
sa
r
e
aa
r
ei
t
e
r
a
t
i
v
em
e
t
h
o
d
s
.N
u
m
e
r
i
c
a
ls
o
l
u
t
i
o
n
s
Then
u
m
e
r
i
c
a
lm
e
t
h
o
d
su
s
e
di
ni
l
g
o
r
i
t
h
mt
h
a
tw
i
l
ls
t
a
b
l
i
s
h i
t
e
r
a
t
i
v
es
e
to
f
i
n
v
o
l
v
et
w
oc
o
m
p
o
n
e
n
t
s
: a
c
a
l
c
u
l
a
t
i
o
n
s
.a
n
ds
e
c
o
n
d
.t
h
et
r
a
n
s
l
a
t
i
o
no
ft
h
ea
l
g
o
r
i
t
h
mi
n
t
oc
o
m
p
u
t
e
rc
o
d
e
s
.u
s
i
n
g
aprogrammingl
a
n
g
'
g
e
.T
h
i
sf
i
n
i
s
h
e
dc
o
d
ei
sr
e
f
e
r
r
e
dt
o s
o
f
t
w
a
r
e
.U
n
t
i
lr
e
c
e
n
t
l
y
r
l
a
n
g
u
a
g
eo
fc
h
o
i
c
ew
a
s
i
ne
n
g
i
n
e
e
r
i
n
ge
d
u
c
a
t
i
o
na
sw
e
l
la
sp
r
a
c
t
i
c
e
.t
h
ecompu
I
t
t
l
eP
a
s
c
a
la
n
dC
.Thec
o
m
p
u
t
a
t
i
o
n
a
le
n
v
i
r
o
n
m
e
n
ti
n
p
r
e
d
o
m
i
n
a
n
t
l
yF
O
R
T
R
A
Nw
i
t
hai
'amecomputingsystems.Today
w
h
i
c
hi
tw
a
sd
e
l
i
v
e
r
e
da
n
dp
r
a
c
t
i
c
e
dw emai
es
c
e
n
ei
sv
e
r
yd
i
f
f
e
r
e
n
t
.I
n
d
i
v
i
d
u
a
lp
e
r
s
o
n
a
lc
o
m
p
u
t
e
r
sC
s
)

r
u
n
n
i
n
gWindows
M
a
c
i
n
t
o
s
hLINUXo
ro
t
h
e
rf
l
a
v
o
r
so
fUNIXo
p
e
r
a
t
i
n
gs
y
s
t
e
m
sa
r
enowt
h
em
e
a
n
s
f
o
rt
e
c
h
n
i
c
a
lc
o
m
p
u
t
i
n
g
.The
yp
o
s
s
e
s
sm
o
r
ep
o
w
e
rt
h
a
nt
h
eo
l
dm
a
i
n
a
mec
o
m
p
u
t
e
r
s
o
w
e
ro
ft
h
eg
r
a
p
h
i
c
a
lu
s
e
ri
n
t
e
r
f
a
c
e(
G
U
I
)eway
a
n
dt
h
e
ya
r
ei
n
e
x
p
e
n
s
i
v
e
.Thep
eu
s
e
ri
n
t
e
r
a
c
t
sw
i
t
h
ec
o
m
p
u
t
e
ra
n
des
o
f
t
w
a
r
e
h r
e
v
o
l
u
t
i
o
n
i
z
e
dt
h
ec
r
e
a
t
i
o
n
o
fs
o
f
t
w
a
r
et
o
d
a
y
. New programming paradigmsa11
y 0'
e
c
lo
r
i
e
n
t
e
d
programming
h
a
v
el
e
dt
onewa
n
di
m
p
r
o
v
e
dw
a
y
st
od
e
v
e
l
o
ps
o
f
t
w
a
r
e
.T
o
d
a
y
c
n
g
i
n
c
c
r
i
n
gs
o
f
t
w
a
r
ci
sc
r
c
a
t
c
db
yB
u
c
hg
c
n
c
l
'
a
lp
r
o
g
r
a
m
m
i
n
gl
a
n
g
u
a
g
c
sI
lS
C
C++

J
a
v
a

P
a
s
c
a
l
a
n
dV
i
s
u
a
lB
a
s
i
c
.Manys
o
f
t
w
a
r
ev
e
n
d
o
r
sd
e
v
e
l
o
pt
h
e
i
rowne
x
t
e
n
s
i
ol
1s

p
r
o
v
i
d
i
n
gac
o
m
p
l
e
1
en
v
i
r
o
n
m
e
n
tf
o
rs
o
f
t
w
a
r
ed
e
vl
o
p
m
e
n
t
t
ot
h
es
t
a
n
d
a
r
dl
a
n
g
u
a
g
e
s
e
r
.T
h
i
sm
a
k
e
si
tv
e
r
yc
o
n
v
e
n
i
e
n
tf
o
ru
s
e
r
swhoa
r
en
o
tn
e
c
e
s
s
a
r
i
l
y
b
yeu
i
rj
o
bd
o
n
ei
nar
e
a
s
o
n
a
b
1
ys
h
o
r
tt
i
m
e
.
p
r
o
g
r
a
m
m
e
r
st
og
e
te
Wo
s

w
i t
h
e
i
rf
l
a
g
s
h
i
pp
r
o
d
u
c
tMA
T
L
A
Ba
10
n
gw
i
t
ha
Ones
u
c
hv
e
n
d
o
ri
sMa
td
e
l
i
v
e
rs
p
e
c
i
f
i
cd
i
s
c
i
p
l
i
n
e
o
r
i
e
n
t
e
d
c
o
m
p
l
e
m
e
n
t
a
r
yc
o
l
l
e
c
t
i
o
no
ft
o
o
l
b
o
x
e
s a
i
n
f
o
r
m
a
t
i
o
n
.I
ni
s
b
kMATLABi
su
s
e
de
x
c
l
u
s
i
v
e
l
yf
o
rn
u
m
e
r
i
c
a
ls
u
p
p
o
r
t
.Nop
r
i
o
r
i
t
hMATLABi
se
x
p
e
c
t
e
dt
h
o
u
g
hf
a
m
i
l
i
a
r
i
t
yw
i
t
ht
h
ePCi
sr
e
q
u
i
r
e
d
.L
ik
e
f
a
m
i
l
i
a
r
i
t
yw
e
f
f
e
c
t
i
v
e
n
e
s
sw
i
t
ht
h
i
snewp
r
o
g
r
a
m
m
i
n
gr
e
s
o
u
r
c
ec
o
m
e
s
a
l
lh
i
g
h
e
r
1
c
v
c
l
l
a
n
g
u
a
g
e
s

f
r
e
q
u
e
n
c
yo
fu
s
e
.o
r
i
g
i
n
a
lw
o
r
k
a
n
dac
o
n
s
i
s
t
e
n
tm
a
n
n
e
ro
fa
p
p
l
i
c
a
t
i
o
n
.
w
i
t
he
f
f
o
r
t
i
ti
sh
o
p
e
dt
h
e a
d
e
rw
i
l
lb
ev
e
r
yf
a
m
i
l
i
a
rw
i
t
ht
h
cu
Byt
h
ec
n
do
ft
h
i
sb
o
o
k

I
N
T
R
O
D
U
C
T
I
O
NTOM
A
1
1
.
A
I
I

Matha
n
dc
o
m
p
u
t
a
t
i
o
n
A
1
g
o
r
i
t
l
l
md
e
v
e
l
o
p
m
e
n
t
M
o
d
e
l
i
n
g
s
i
m
u
l
a
t
i
o
n
a
n
dp
r
o
t
o
t
y
p
i
n
g

e
x
p
l
o
r
a
t
i
o
n
.andvisualization
D
a
t
aa
n
a
l
y
s
i
s

Scientificandengineeringgraphics

A
p
p
l
i
c
a
t
i
o
nd
e
v
e
l
o
p
m
e
n
t

i
n
c
l
u
d
i
n
gGUIb
u
i
l
d
i
n
g
Theu
s
eofMA
T
I
.
.
A
Bh
a
se
x
p
l
o
d
e
dr
e
c
e
n
t
l
ya
n
ds
oh
a
v
ei

f
e
a
t
u
r
e
s
.Tbi
sc
o
n
t
i
n
u
o
u
s
T
I
.
.
A
Ba
ni
d
e
a
lv
e
h
i
c
l
ef
o
re
x
p
l
o
r
i
n
gp
r
o
b
l
e
m
si
na
l
ld
i
s
c
i
p
l
i
n
e
s
e
n
h
a
n
c
e
m
e
n
tm
a
k
e
sMA
.
tT
h
i
sa
b
i
1
it
yi
sm
u
l
t
i
p
l
i
e
db
ya
p
p
l
i
c
a
t
i
o
n
t
h
a
tm
a
n
i
p
u
1
a
t
em
a
t
h
e
m
a
t
i
c
a
lc
o
n
t
e
n
s
p
e
c
i
f
i
cs
o
l
u
t
i
o
n
st
h
r
o
u
g
ht
o
o
l
b
o
x
e
s
.Tol
e
a
r
nm
o
r
ea
b
o
u
tt
h
el
a
t
e
s
tMA
T
I
.
.
A
B
.i
t
s
f
e
a
t
u
r
e
st
h
et
o
o
l
b
o
x
e
sc
o
l
l
e
c
t
i
o
no
fu
s
e
r
s
u
p
p
o
r
t
e
da
r
c
h
i
v
e
si
n
f
o
r
m
a
t
i
o
na
b
o
u
t
U
s
e
n
e
ta
c
t
i
v
i
t
i
e
so
t
h
e
rf
o
r
u
m
si
n
f
o
r
m
a
t
i
o
na
b
o
u
tb
o
o
k
sa
n
d8
0o
nv
i
s
i
tMATLAI
la
1
l
w
w
w
.
m
a
1
h
w
o
r
k
s
.
c
o
m
.
h
u
p
:/

1
.
2
.
1 WhyMA
1
1
.
A
B
?
MATLABi
sas
t
a
n
d
a
r
dt
o
o
lf
o
ri
n
'
o
d
u
c
t
o
r
ya
n
da
d
v
a
n
c
e
dc
o
u
r
s
e
si
nm
a
t
h
e
m
a
t
i
c
s
e
n
g
i
n
e
e
r
i
n
g
a
n
ds
c
i
e
n
c
ei
nmanyu
n
i
v
e
r
s
i
t
i
e
sa
r
o
u
n
d w
o
r
l
d
.I
ni
n
d
u
s
t
r
y

i
ti
sat
o
o
1
h

d
e
v
e
l
o
p
m
e
n
t
a
n
da
n
a
l
y
s
i
s
.I
ni
sb
o
o
kM
A
T
L
A
B
'
Sb
a
s
i
ca
r
r
a
y
o
fc
h
o
i
c
ef
o
rr
e
s
e
l
e
m
e
n
ti
se
x
p
l
o
i
t
e
dt
om
a
n
i
p
u
l
a
t
ev
e
c
t
o
r
sa
n
dm
a
t
r
i
c
e
st
h
a
ta
r
en
a
t
u
r
a
lt
on
u
m
e
r
i
c
a
l
t
e
c
h
n
i
q
u
e
s
.I
nt
h
en
e
x
tc
h
a
p
t
e
r
i
t
sp
o
w
e
r
f
u
lv
i
s
u
a
l
i
z
a
t
i
o
nf
e
a
t
u
r
e
sa
r
eu
s
e
df
o
r
g
r
a
p
h
i
c
a
lo
p
t
i
m
i
z
a
t
i
o
n
.I
ner
e
s
to
ft
h
i
sb
o
o
kt
h
ei
n
t
e
r
a
c
t
i
v
ef
e
a
t
u
r
l
db
u
i
l
t
i
n
1
g
o
r
i
t
h
m
si
n
t
of
u
n
c
t
i
o
n
i
n
gc
o
d
e
i
na
s
u
p
p
o
r
to
fMATLABa
r
eu
t
i
l
i
z
e
dt
ot
r
a
n
s
l
a
t
ea
f
r
a
c
t
i
o
no
ft
h
et
i
m
en
e
e
d
e
di
no
t
h
e
rl
a
n
g
u
a
g
e
sl
i
k
eCo
rF
O
R
T
R
A
N
.Toward eend

we
w
i
l
li
n
o
d
u
c
et
h
eu
s
eo
fi
t
sO
p
t
i
m
i
t
i
o
nT
o
o
l
b
o
x
.
n
tc
o
v
e
r
a
g
eo
nt
h
e~hniques a
n
d
Mostb
o
o
k
so
no
p
t
i
m
i
z
a
t
i
o
np
r
o
v
i
d
ee
x
c
e
l
l
a
1
g
o
r
i
t
h
m
s
a
n
dsomea
1
s
op
r
o
v
i
d
eap
r
i
n
t
e
dv
e
r
s
i
o
no
ft
h
ec
o
d
e
.Thec
o
d
ei
s q
u
i
d
t
ob
eu
s
e
di
nmainfmmEorl
e
g
a
c
ye
n
v
i
r
o
n
m
e
n
t
st
h
a
tm
u
s
ts
u
p
p
o
r
tac
o
m
p
i
l
e
ri
nt
h
e
l
a
n
g
u
a
g
eec
o
d
ei
sw
r
i
t
t
e
n
w
h
i
c
hi
sp
r
e
d
o
m
i
n
a
n
t
l
yF
O
R
T
R
A
N
.Eachs
t
u
d
e
n
te
n
a
n
s
l
a
t
e
st
h
eo
r
i
n
t
e
dc
o
d
ei
n
t
oaf
o
r
r
ns
u
i
t
a
b
l
ef
o
rc
o
m
p
u
t
a
t
i
o
n
. l
en
e
x
ts
t
e
pi
st
o
e
b
u
gc
o
m
p
i
l
a
t
i
o
ne
r
r
o
r
sr
u
n p
r
o
g
r
a
ma
n
dt
r
o
u
b
l

h
o
o
t
c
o
m
p
i
l
et
h
ec
o
d
ed
u
l
t
sc
a
nb
eo
b
t
a
i
n
e
d
.Ani
m
p
o
r
t
a
n
tc
o
n
s
i
d
e
r
a
t
i
o
ni
s
e
x
e
c
u
t
i
o
nc
r
r
o
r
sb
e
f
o
r
eu
s
e
f
u
l s
t
h
es
t
u
d
e
n
ti
se
x
p
e
c
t
e
dt
ob
ec
o
n
v
e
r
s
a
n
tw
i t
h
ep
r
o
g
r
a
m
m
i
n
gl
a
n
g
u
a
g
e
.W
h
i
l
e
r
ss
e
v
e
r
a
ly
r
sa
g
oi
si
sn
ol
o
n
g
e
rt
r
u
e
.
F
O
R
T
R
A
Nw
a
sf
o
r
m
a
l
l
yt
a
u
g
h
tt
oe
n
g
i
n

1
.
2I
N
T
R
O
D
u
c
n
O
NT
OMATl.AB 27

2
6 I
N
T
R
O
D
U
C
T
I
O
N
Manye
l
e
c
t
r
i
c
a
le
n
g
i
n
e
e
r
i
n
gd
e
p
a
r
t
m
e
n
t
sh
a
v
es
t
a
r
t
e
dt
oe
m
b
r
a
c
eCa
n
dC
+
+
.PCsa
r
e
e
s
p
e
c
i
a
l
l
yt
h
o
s
ew
i
t
hGUI

b
u
ti
t
su
s
ef
o
r
g
r
e
a
tf
o
rt
h
eu
s
eo
fr
e
a
d
y
m
a
d
ea
p
p
l
i
c
a
t
i
o
n
s
p
r
o
g
r
a
m
m
i
n
gi
nF
O
R
T
R
A
Nh
a
sy
e
tt
ot
a
k
eo
f
f
.
T
I
.
.
A
Bs
u
f
f
e
r
sv
e
r
yf
e
wo
ft
h
e
s
ed
i
s
a
d
v
a
n
t
a
g
e
s
.I
t
sd
ei
sc
o
m
p
a
c
t di
t
i
s
i
n

i
t
i
v
e
MA
ot
r
a
d
i
t
i
o
n
a
l
t
ol
e
a
m
.Tomakep
r
o
b
l
e
m
sw
o
r
kr
e
q
u
i
r
e
sf
e
wl
i
n
e
so
fc
o
d
ec
o
m
p
a
r
e
dt
l
a
n
g
u
a
g
e
s
.I
n
s
t
e
a
do
fp
r
o
g
r
a
m
m
i
n
gf
r
o
meg
r
o
u
n
du
p

s
t
a
n
r
dMA
T
I
.
.
A
Bp
i
e
c
e
so
fc
o
d
e
c
a
n
b
ea
d
e
d
t
o
g
e
e
r

i
m
p
l
e
m
e
n
tt
h
enewa
l
g
o
r
i
t
h
m
.I
ta
v
o
i
d
st
h
es
n
d
a
r
d :
p
r
u
:

mpilationandlinks uencebyusing i
n
t
e
.
r
e
t
e
r
.Ano
t
h
e
ri
m
p
o
r
t
a
n
t a
s
o
ni
s
c
o
d
e
p
o
b
i
l
i
t
y
. l
ed
ew
r
i
t
t
e
nf
o
rePCv
e
r
s
i
o
nofMA
T
I
.
A
Bd
s
n
o
th
a
v
et
ob
ec
h
a
n
g
e
d
i
sw
i
l
ln
o
tb
et
r
u
ei
fs
y
s
t
e
m
d
e
p
e
n
d
e
n
tr
e
s
o
u
r
c
e
s
f
o
rt
h
eM
a
c
i
n
t
o
s
ho
rt
h
eU
N
I
Xs
y
s
t
c
m
s
.l
medinec
e
.Asa
ni
l
Iu
s
t
r
a
t
i
o
no
fMA
T
I
.
A
B'
Sl
e
a
m
i
n
ga
g
i
l
i
t
y

b
o
MA
TI.ABd
e
r
e
di
nt
h
i
sb
o
o

ks
o
m
e
t
h
i
n
gv
e
r
yd
i
f
f
i
c
t
a
omplishwi

a
p
p
l
i
e
do
p
t
i
m
i
z
a
t
i
o
na
r
ev
s
a
yF
O
R
T
R
A
Na
n
do
p
t
i
m
i
z
a
t
i
o
no
rC++ do
p
t
i
m
i
z
a
t
i
o
n
.F
l
e
r
m
o
r
e

I
A
'

.
.
A
B
l
i
k
eCo
r

i
sag
l
o
b
a
l
l
yr
e
l
e
v
a
n
tp
r
o
d
u
c
.
t
F
O
R
T
R
A
N
1
.
2
.
2 M
A
T
L
A
BI
n
s
t
a
l
l
a
l
i
o
nI
s
s
u
e
s
T
h
i
sb
o
o
kd
o
e
sn
o
tC
a
r
ei
fy
o
ua
r
er
u
n
n
i
n
gMATLABo
na
ni
n
d
i
v
i
d
u
a
lPCo
rMaca
n
e
t
w
o
r
k
e
dPCo
rMac
o
ri
n
d
i
v
i
d
u
a
lo
rn
e
t
w
o
r
k
e
dU
n
i
xw
o
r
k
s
t
a
t
i
o
J
ls
.MATLABt
a
k
c
s
c
a
r
eo
ft
h
i
si
s
s
u
.T
heb
o
o
ka
s
s
u
m
e
sy
o
uh
a
v
eaw
o
r
k
i
n
ga
n
da
p
p
r
o
p
r
i
a
t
e
l
yI
i
c
e
n
s
e
d
n
to
fy
o
u
.Ony
o
u
rp
e
r
s
o
n
a
lm
a
c
h
i
n
ef
o
r
c
o
p
yo
fMATLABi
nt
h
em
a
c
h
i
n
ei
no
i
n
s
t
a
l
l
i
n
gMATI
.AB
f
o
l
l
o
wt
h
ei
n
s
t
r
u
c
t
i
o
n
st
h
a
ta
c
c
o
m
p
a
n
idt
h
eb
o
x
.T
y
p
i
c
a
l
l
ye
s
y
s
t
e
ma
d
m
i
n
i
s
t
r
a
t
o
rh
a
n
d
l
e
sn
e
t
w
o
r
ki
n
s
t
a
l
l
a
t
i
o
n
s
.
T
t
.
A
BV
e
r
s
i
o
n5
.
2(
u
n
t
i
lC
h
a
p
t
e
r4
)a
n
dV
e
r
s
i
o
n5
.
3(
er
e
s
t
)
.
T
h
i
sb
o
o
ku
s
e
sMA
A
Bv
e
r
s
i
o
n5
.
0a
n
da
b
o
v
e
V
e
r
s
i
o
n6h
a
sb
e
e
ns
h
i
p
p
i
n
gs
i
n
c
et
h
ef
a
l
lo
f2
0
0
0
.MA
s
h
o
u
l
dw
o
r
k
.I
fMA
TI.ABh
a
sc
h
a
n
g
e
da
n
yb
u
i
l
t
i
ncommandsi
nn
e
w
e
rv
e
r
s
i
o
n
s
t
h
eo
n
e e
di
nt
h
eb
o
o
k
i
ts
h
o
u
l
dp
r
o
v
i
d
ei
n
f
o
r
m
a
t
i
o
no
nt
h
ea
v
a
i
l
a
b
i
l
i
t
yo
fn
e
w
e
r
o
rl
h
r
o
wa
ne
r
r
o
r
ar
a
r
ec
a
s
e
)
.TheMATLABc
o
d
ea
n
d
v
e
r
s
i
o
no
f commands(
e
x
a
m
p
l
e
sw
e
r
ec
r
e
a
t
e
do
na
nI
n
l
e
lP
e
n
t
i
u
mP
r
o2
0
0r
u
n
n
i
n
gt
h
eWindowsNT4
o
p
e
r
a
t
i
n
gs
y
s
l
e
m
.Thee
a
r
l
i
e
rc
o
d
e
sw
e
r
ea
l
s
ot
e
s
t
e
do
naDECAlp
h
as
e
r
v
e
rr
u
n
n
i
n
g
OSFlV
4
.TheMA
T
I
.
.
A
Bv
e
r
s
i
o
no
nt
h
es
e
r
v
e
rwas5
.1
.
I
no
r
g
a
n
i
z
a
t
i
o
n
sa
n
di
n
s
t
i
l
u
t
i
o
n
sMA
T
t
.
A
Bs
o
f
t
w
a
r
ei
su
s
u
a
l
l
yl
i
c
e
n
s
e
do
na
t
h
e1
a
t
e
s
tv
e
r
s
i
o
no
ft
h
es
o
f
t
w
a
r
ei
nt
h
e
s
e
s
u
b
s
c
r
i
p
t
i
o
nb
a
s
i
s
.U
s
e
r
st
h
e
r
e
f
o
r
eh
a
v
s
i
t
u
a
t
i
o
n
s
.I
ns
u
c
hac
a
s
et
h
enewv
e
r
s
i
o
n(
a
t t
i
m
et
h
eb
o
o
ki
sf
i
r
s
tp
u
b
l
i
s
h
e
d
)
w
i
l
l
e
b
eV
e
r
s
i
o
n6(
o
rR
e
l
e
a
s
eR1
2
)
.T
h
i
sbooki
sb
a
s
e
do
nV
e
r
s
i
o
n5(
R
e
l
e
a
s
eR
I
1
)
.l
r
o
g
r
a
m
m
i
n
gi
n
s
t
r
u
c
t
i
o
n
si
nt
h
eb
o
o
ks
h
o
u
1
de
x
e
c
u
t
ew
i
t
h
o
U
lp
r
o
b
l
e
m
c
o
d
e d p
nt
h
ePCp
1
a
t
f
o
r
m
t
h
ed
e
f
a
u
l
t
i
nt
h
enewv
e
r
s
i
o
n
.I
nt
h
enewR
e
l
e
a
s
eR12o
T
I
.
.
A
Bi
sac
o
m
p
o
s
i
t
eo
fs
e
v
e
r
a
lwindowsp
r
e
s
e
n
t
i
n
gd
i
f
f
e
r
e
n
tp
i
e
c
e
s
a
p
p
e
a
r
a
n
c
eo
fMA
b
ec
u
s
t
o
m
i
z
e
dt
h
r
o
u
g
ht
h
emenu
o
fi
n
f
o
r
m
a
t
i
o
n
.T
h
i
si
sshowni
nF
i
g
u
r
e1
.
8
.T
h
i
sc
b
yV
i
e
wD
e
s
k
t
o
pLa
y
o
u
tCommandWindowO
n
l
y

w
h
i
c
ha
l
s
ocoponds
b
r
e
v
i
o
u
s 1
e
a
s
e
.I
ti
sshowni
nF
i
g
u
r
e1
.
9
.l
i
sv
i
e
wi
ss
i
m
p
l
ea
n
dl
e
s
s
t
oep
d
i
s
t
r
a
c
t
i
n
gf
o
rg
e
t
t
i
n
gf
a
m
i
l
i
a
rw
i
MA
T
l
.
.
A
B
.N
o
t
eeMA
T
t
.
A
Bcommandp
r
o
m
p
t
>
.T
h
i
s windowd
e
l
i
v
e
r
ss
t
a
n
d
a
r
do
p
e
r
a
t
i
n
gs
y
s
t
e
mcommands f
o
rs
a
v
i
n
g
i
s>
p
r
i
n
t
i
n
go
p
e
n
i
n
gf
i
l
e
sa
n
ds
oo
nl
h
r
o
u
g
hmenusa
n
dacommandb
a
r
.I
no
r
d
e
rt
o

F
1g
u
r
e1
.
8 MATLABo
p
e
n
i
n
gv
i
e
w
.

h
'

m
.

c
r
e
a
t
ef
i
l
e
s(
o
rc
o
d
e
)a
n
o
t
h
e
rg
r
a
p
h
i
c
a
1windowi
sn
e
d
e
d
.T
h
i
si
seMATLABMf
i
l
e
e
d
i
t
o
r
/
d
e
b
u
g
g
e
rwindowa
n
di
sshowni
nF
i
g
u
1.10
.Thee
d
i
t
o
ru
s
e
sc
o
1
0
rf
o
r
e
n
h
a
n
c
i
n
gc
o
d
er
e
a
d
a
b
i
l
i
t
y
.E
a
c
ho
f
t
h
e
s
e
windowsspawna
d
d
i
t
i
o
n
a
1d
i
a
1
0
gb
o
x
e
s(
i
f
n
e
c

)i

o
r

t
i

n
1

g
f
i

e
a

t
u

r
e
s

d
u

r
i
i
r
w
o
r
k
s
p
a
c
eb
r
o
w
s
e
r
.
Whi
l
et
h
i
si
sn
o
ts
omucha
sa
ni
n
s
t
a
l
l
a
t
i
o
ni
s
s
u
e
i
two
u
1
db
ea
p
p
r
o
p
r
i
a
l
el
od
i
s
c
u
s
se
l
a
I
1
st
ow
r
i
a
n
d
e
x
e
c
MA
TI.ABC
Gen
y

e
r
ea
r
et
w
ow
a
y
st
o
way kp
T
I
.
A
B
:i
n
t
e
r
a
c
t
i
v
e
l
y
a
n
dt
h
r
o
'
p
t
s
.I
ni
n
t
e
r
a
c
t
i
v
em
o
d
e
.rnn1a
n
d
s
w
o
r
k MA
a
r
e
e
n
t
e
d
d
u
t
e
do
n
ea
tat
i
m
ei
nt
h
eMA
T
t
.
AB
c
O
rnn1
a
n
dw
i
n
d
o
w
.Once s
e
s
s
1
0
n
T
t
.
A
Be
x
i
t
e
d
t
h
e
r
ei
sn
oi
n
f
o
r
m
a
t
i
o
n1
t
a
i
n
e
do
n s
e
to
fcommands
i
so
v
e
ra
n
dMA
w
e
r
eu
s
e
da
n
dw
h
a
tw
o
r
k
e
d(
u
n
l
e
s
sac
h
o
i
c
ew
a
smadet
os
a
v
et
h
es
e
s
s
i
o
n
)
.L
a
r
g
e
e
a
u
e
n
s
o
fc
O
r
n
n
1
a
I
1
d
sa
r
eu
s
u
a
l
l
yd
i
c

u
1
u

.
ok
P

a
c
ko
f i

e
r
a
c
U
v
s
e
s
s
1
0
n
.
s
.

f
.

o
o

t
i

f
t
l
e
.MostMATLABf
i
l
l
l
e
dmf
t
l
e
s
.The
mi
sam
e
e
x
t
B
I
E
S
i
o
nu
s
u
a
l
l
ym
s
e
r
v
e
df
o
rM
A
I
L
A
B
.
A
n
yc
o
d
echangEsmemadet
h
m
u
s
h
g

E
ee
d
i
t
o
r
.t
h
ec
h
a
n
g
e
st
oef
i
l
es
a
v
e
d
a
n
def
i
l
c
e
x
e
c
u
di
nMA
T
t
.
A
Ba
g
a
i
n
.Exc
u

t
i

n
M
A
T
t
.
A

t
l

g
h
s

i
s

p
r

e
d

1
l

i
n
t

o
r
e

v
e
r
tt

oi
n
t
e
r
a
c
t
i
v
emodea
n
yt
i
m
e
.

f
i
l
ei
nMA
T
I
.
AB

T
h
is
i
sc
a
l
l
e
das
c
r
i
p
tm

ime
1
.
2
.
3 UsingM
A
T
L
A
BtheF
i
r
s
tT
MATw
i

lb
eu
s
e
di
n
t
e i
v
e
1
yi
n i
ss
e
c
t
i
o
n
. The a
u
.
t
h
o
rv
e
r
y n
y
n
d
sah
a
n
d
s
o
na
p
p
r
o
a
c
ht
ou
n
d
e
r
s
t
a
n
d
i
n
gMA
T
t
.
AB
Th
i
smeanstypmgt
t
1
e
r
e
c
ornn1e
c
o
d
e
sy
o
u
r
s
e
l
f
.I
ta
1
s
oi
n
c
1u
d
e
su
n
d
e
r
s
l
a
n
d
i
n
gt
h
es
y
n
t
a
xr
r
o
r
s

a
n
dd
e
b
u
g
g
i
n
ga
n
d

28 I
N
T
R
O
D
U
C
T
I
O
N

1
.
2

I
N
T
R
O
D
U
C
T
I
O
NTOMAl UB

29

r
i
n
gec
o
d
e
.Thea
u
t
h
o
ri
sn
o
ta
w
a
r
eo
fas
i
n
g
l
ee
x
a
m
p
l
ewhereprogramming
r
e
c
t
i
e
.
l
ni
sb
kMATLABcodesegmentsare
wasl
e
a
r
n
e
dw
i
t
h
o
u
tt
h
er
e
a
d
e
rt
y
p
i
n
g d
o
l
ds
t
y
l
el
Is
e
df
o
re
m
p
h
a
s
i
z
i
n
gcommandso
ro
t
h
e
rp
i
e
c
e
so
f
c
o
u
r
i
e
rf
o
n
tw
i
t
heb
u
g
g
e
s
t
i
o
n
so
re
x
e
r
c
i
s
e
s
. Th
e
i
n
f
o
r
m
a
t
i
o
n
.A
n
y
t
h
i
n
ge
l
s
ei
sr
e
c
o
m
m
e
n
d
a
t
i
o
n
ss
p
l
a
c
e
h
o
l
d
e
rf
o
rcommandi
si
ni
t
a
l
i
c
.
i
si
saC
l
a
s
s
t
o
o
mL
i
c
e
n
s
ef
O
ti
n
s
t
t
u
c
t
i
o
n
a
lu
s
eo
n
l
y
.
R
e
s
e
a
t
c
ha
n
dc
o
m
e
t
c
i
a
lu
s
ei
sp
t
o
h
i
b
1
t
e
d
.

BeforeWeS
t
a
r
t
:Thef
o
lowingp
u
n
c
t
u
a
t
i
o
n ds
p
e
c
i
a
lc
h
a
r
a
c
t
e
r
sa
r
ew
o
r
t
h
n
o
t
i
n
g
:

Toqets
t
a
t
t
e
ds
e
l
e
c
t"HAT
L
A
BH
e
l
p
"f
t
o
m.
t
h
eH
e
l
pm
e
n
u
.

F
l
g
u
r
e1
.
9 MATLABCommandw
l
n
d
o
w
.

h
.
J

>>

Ft

J-

D
t
.P
.V
e
n
k
a
t
a
J
;a
m
e
n
l

vl
pL
rEYHHEJr

m1z
a
t
i
o
n 11 ~IATtAB
O
p
t
i

'
i
IJ
o
h
n~iley

F
1g
u
r
e1
.
1
0 MaUabE
d
i
t
o
r
/
D
e
b
u
g
g
e
rw
i
n
d
o
w
.

>> help command_name

>> helpwin

>> helpdesk

A
lt
e
x
ta
f
t
e
rt
h
e% s
i
g
ni
sc
o
n
s
i
d
e
r
e
da
comment
. MATLAB i
g
n
o
r
e
sa
n
y
t
h
i
n
gt
o
t
h
er
i
g
h
to
ft
h
e% s
i
g
n
.Commentsa
r
e
u
s
e
dl
i
b
e
r
a
l
l
yi
nt
h
i
sbookt
oi
n
d
i
c
a
t
e
why'sandw
h
a
t
'
so
fcommando
rc
o
d
e
u
s
a
g
e
.Don
o
to
v
e
r
l
o
o
kt
h
e
m
.Youdon
o
t
however

p
r
o
b
a
b
l
y
h
a
v
et
ocodethemi
n
twoweeksl
a
t
e
ryouwi

ln
o
tremember
whyyoud
e
v
e
l
o
p
e
dt
h
a
tp
a
r
t
i
c
u
l
a
rp
i
e
c
e
o
fc
o
d
e
.
.
t
T
h
i
si
st
h
ed
e
f
a
u
l
tMATLABpromp
As
e
m
i
c
o
l
o
na
tt
h
ee
n
do
fal
i
n
ep
r
e
v
e
n
t
s
MA
A
Bfrome
c
h
o
i
n
gt
h
ei
n
f
o
r
m
a
t
i
o
nyou
e
n
t
e
ront
h
es
c
r
e
e
n
.
A commaw
i
l
lc
a
u
s
et
h
ei
n
f
o
r
m
a
t
i
o
nt
o
e
c
h
o
.D
e
f
a
u
l
ti
sacomma.
As
u
c
c
e
s
s
i
o
no
ft
h
r
e
ep
e
r
i
o
d
sa
tt
h
eendo
f
t
h
el
i
n
ei
n
f
o
r
m
sMATLAB codew
i
l
l
c
o
n
t
i
n
u
ethen
e
x
tl
i
n
e
.Youc
a
n
n
o
ts
p
l
i
t
av
a
r
i
a
b
l
e name a
c
r
o
s
s two l
i
n
e
s
.
Y
o
u
c
a
n
n
o
tc
o
n
t
i
n
u
eacommentona
n
o
t
h
e
r
I
i
n

.
Youc
a
n
s
t

o
'
pMATLABe
x
e
c
u
t
i
o
na
n
dg
e
t
b
a
c
k
t
h

c
omm!

mdprom

pt
b
yt
y
p
i
n
g^
c
(
C

l
C)
b
yh
0
1

d
i

n
gdown
t

g
e
e
r
.
Wi
l
Id
i
s
p
l
a
yi
n
f
o
r
m
a
t
i
o
nont
h
ev
a
r
i
o
u
s
wayst
h
ecommandc
a
nb
eu
s
e
d
.T
h
i
si
st
h
e
q
u
i
c
k
e
s
twayt
ou
s
eh
e
l
p
.
Opensah
e
l
pt
e
x
twindowt
h
a
tp
r
o
v
i
d
e
s
morei
n
f
o
r
m
a
t
i
o
nona
l
lo
ft
h
eMATLAB
r
e
s
o
u
r
c
e
si
n
s
t
a
l
l
e
dony
o
u
rs
y
s
t
e
m
.
P
r
o
v
i
d
e
sh
e
l
pu
s
i
n
gab
r
o
w
s
e
rwindow.

1
.
2I
N
T
A
O
D
U
C
T
lO
NT
OM
A
TlA
D

30 I
N
T
A
O
D
U
C
T
I
O
N
Fort
h
o
s
eo
fyouprogrammingf
o
rt
h
ef
i
r
s
tt
i
m
e

t
h
ee
q
u
a
lt
os
i
g
n(
=
)h
a
sav
e
r
ys
p
e
c
i
a
l
i
n
c
Ju
d
i
n
gMATLAB.I
ti
sc
a
l
l
e
dt
h
ea
s
s
i
g
n
m
e
n
t
meaningi
na
l
lprogrammingl
a
n
g
u
a
g
e
s
ev
a
r
i
a
b
l
eont
h
cI
c
f
t
h
a
n
ds
i
d
eo
ft
h
es
i
g
ni
sa
s
s
i
g
n
e
dt
h
ev
a
l
u
eo
fe
o
p
e
r
a
l
o
r
. l
q
l
l
a
llOo
p
e
r
a
t
i
o
ni
nMATLABi
su
s
u
a
l
l
ya
c
c
o
m
p
l
i
s
h
e
dby
r
i
g
h
t
h
a
n
ds
i
d
e
.Thca
c
t
u
a
le
ad
o
u
b
l
ec
q
u
a
lt
os
i
g
n(
=
=
)
.

J
1u
s
ci
tal
o
ti
nt
h
en
e
x
tc
h
a
p
t
e
rf
o
r
r
e
l
e
v
a
n
c
ei
nm
a
t
h
e
m
a
t
i
c
so
ra
n
y
w
h
e
r
ee
l
s
e
.Wewi
g
Cl
Ie
r
a
t
i
n
gd
a
t
af
o
rg
r
a
p
h
i
c
a
lop
z
a
t
i
o
n
.

AnI
n
t
e
r
a
c
t
l
v
eSe
s
s
l
o
n
: S MA
A
B.On PCp
r
c
s
st
h
eS
t
a
r
tb
u
t
t
o
n
.P
r
e
s
st
h
e
r
c
s
st
h
cMATLAB'
g
r
a
mG
r
o
u
p
.P
r
e
s
sM A
TLAB.TheMATLAB
Programb
u
t
t
o
n
.P
Co
mmandwindowo
p
e
n
supa
n
da
f
t
e
rsomci
n
t
r
o
d
u
c
t
o
r
ym
e
s
s
a
g
e
sE w
i
l
lb
eaMATLAB
p
r
o
m
p
t
i
n
d
i
db
yd
o
u
b
l
ef
o
r
w
a
r
da
r
r
o
w
.l
i
si
s i
n
d
i
c
a
t
i
o
n MATLABi
sr
e
a
d
yt
o
aeptyour
mmands.Ont
h
en
e
t
w
o
r
k
e
ds
t
i
o
n
s

t
y
p
i
n
gma
t
J
a
ba
tt
h
ewindowp
r
o
m
p
t
a
sd
e
f
m
e
di
ne
s
h
o
u
l
ddot
h
es
a
m
e
.N
o
t
et
h
a
tt
h
eMATLABp
r
o
g
r
a
mm
u
s
tb
ei
nt
h
ep
a
.
I
o
g
i
nor.h
f
i
l
e
.As
kt
h
el
a
ba
s
s
i
s
t
a
n
tf
o
rh
e
l
p
.Th
e s
t
a
n
d
a
r
dUNIXp
r
o
c
c
d

t
r
na
tt
h
ee
n
do
fel
i
n
co
rb
e
f
o
ecommen
.
tTh
ecommcntsr
c
l
a
t
ct
ow
h
a
ti
s
b
e
i
n
gt
y
p
e
d dw
h
a
ti
sd
i
s
p
l
a
y
e
di
nt
h
eMATLABcommandw
i
n
d
o
w
.F
I
1
f
r
c
c
y
o
u
r
ownv
a
r
i
a
t
i
o
s
S
.I
nf
a
c
t t
ou
n
d
e
d
a
n
d
r
e
i
n
f
o
c
ommands
i
ti
s commended
y
o
umakeupy
o
u
rowne
x
a
m
p
l
e
so
f
t
e
n
.

SomeA
d
d
i
t
i
o
n
s
lFeatures
MA
A
Bi
sc
a
s
es
e
n
s
i
t
i
v
e
.Anai
sd
i
f
f
e
r
e
n
tt
h
a
nA
.T
h
i
si
sd
i
f
f
i
c
u
l
tf
o
r
p
e
r
s
o
n
swhoa
r
cu
s
c
dt
oF
O
R
T
R
A
N
.A
l
lb
u
i
l
t
i
nMATLABcommandsa
r
ci
n
l
o
w
e
r
c
a
s
e
.
MATLABd
o
e
sn
o
tn
e
c
dat
y
p
ed
e
f
i
n
i
t
i
o
no
rad
i
m
e
n
s
i
o
ns
t
a
t
e
m
e
n
tt
oi
n
t
r
o
d
u
c
e
ta
u
t
o
m
a
t
i
c
a
l
l
yc
r
t
e
soneonf
i
r
s
te
n
c
o
u
n
t
e
ri
nc
o
d
e
.Th
et
y
p
ei
s
v
a
r
i
a
b
l
e
s
.I
a
s
s
i
g
n
e
di
nc
o
n
t
e
x
t

4.0; e 5.0
1.0; b
2.0; c 3.0d
why did only c and e echo on the screen?
lists all the variables in t
he workspace
>> who
g
ives the value stored in a
>> a
>>A 1.5
;
v
ariable A
% case matters
>> aA
>>one a; two b; three C;
>> a
ssigningvalues to new variables
>>four d; five e
; six pi; valueof pi available
>> A1
[
ab c ;
'd e f
] A1 is a 2 by 3 matrix
>> % space or comma separates columns
>> % semi-colon separates rows
>>a

t
a
r
tw
i
t
hal
e
t
t
e
ra
n
dc
o
n
t
a
i
nupt
o3
1c
b
a
r
a
c
t
c
r
s(
o
n
l
yI
c
t
t
r
s
V
a
r
i
a
b
l namess
d
i
g
i
t
s
andu
n
d
e
r
s
c
o
r
c
)
.

>>

MATLABu
s
c
ssomeb
u
i
l
t
i
nv
a
r
i
a
b
l
en
a
m
e
s
.Avoidu
s
i
n
gb
u
i
l
t
i
nv
a
r
i
a
b
l
enam.
A
l
lnumbersa
r
cs
t
o
r
c
di
n
t
e
r
n
a
l
l
yu
s
i
n
gt
h
el
o
n
gf
o
r
m
a
ts
p
c
c
i
f
i
c
db
yIEEE
f
1o
a
t
i
n
g
p
o
i
n
ts
t
a
n
d
a
r
d
.The
s
enumbersh
a
v
er
o
u
g
h
l
yap
r
e
c
i
s
i
o
no
f1
6d
e
c
i
m
a
l
d
i
g
i
t
s
.Theyr
a
n
g
er
o
u
g
h
l
yb
e
t
w
e
e
n1
0
E
3
0
8a
n
d1
0
E
+
3
0
8
.Howeve
y
are
d
i
s
p
l
a
y
e
dd
i
f
f
e
r
c
n
t
l
yd
e
p
e
n
d
i
n
go
nt
h
ec
o
n
t
e
xt
.
MATLABu
s
e
sc
o
n
v
e
n
t
i
o
n
a
ld
e
c
i
m
a
ln
o
t
a
t
i
o
nf
o
rnumbersu
s
i
n
gd
e
c
i
m
a
l
p
o
i
n
t
s and l
e
a
d
i
n
gs
i
g
n
so
p
t
i
o
n
a
l
l
yf
o
r example1
-9+9.00
.
0
0
1
9
9
.
9
9
9
9
.
S
c
i
e
n
t
i
f
i
cn
o
t
a
t
i
o
ni
sc
x
p
r
c
s
s
e
dw
i
t
ht
h
el
e
t
t
e
re

f
o
rc
x
a
m
p
l
e
2
.
0
e
0
31
.
0
7
e
2
3

1
.7
3
2
e
+
0
3
.

>> A1(2
2
)
>>
>> size(
A
1
l

l
m
a
g
i
n
a
r
ynumbersu
s
ee
i
t
h
e
rio
rjas
u
f
f
i
x
f
o
re
x
a
m
p
l
el
i
3.
l4
j
3
e
5.
i

I
nt
h
ec
a
s
eo
fa
r
r
a
y
s
e
a
c
ho
ft
h
e
s
eo
p
e
r
a
t
o
r
sc
a
nb
eu
s
e
dw
i
a
p
e
r
i
o
dp
r
e
f
i
x
e
dt
oe
o
p
e
r
a
t
o
r

f
o
rc
x
a
m
p
l
e
(.
*
)
o
r(^)o
r(1). h
a
sas
p
e
c
i
a
lmeaningi
nMATLAB.I
t
imp
i
1
e
sl
e
m
e
n
t
b
y
e
l
e
m
e
n
to
p
e
r
a
t
i
o
n
.I
ti
su
s
c
f
u
lf
o
rq
u
i
c
kc
o
m
p
u
t
a
t
i
o
n
.I
th
a
sno

"

accesses the Matrix element on the

% second row and second column

gives you the size of the matrix


(rowcolu
mIls
l
What should happen here?
>>AA1 size(A1l
> > f
romprevious satement the size of A1
> > c
ontains two numbers organized as a row
> > m
atrix. This is assigned to AA1
AA1 i
s a one by two matrix
>> size(AA1l
t
his transposes the matrix A1
>> A1'
the transpose of matrix A1
>>B1 A1'
> > i
s assigned to B1. B1 is a three by two matrix
>> C1 "
' A1 *B1 SinceA1 and B1 are matrices this
> > i
8 a matrix multiplication
> > S
hould this multiplication be allowed?
> > c
onsidermatrix multiplication in C or
>

Operators:Thef
o
l
l
o
w
i
n
ga
r
ct
h
ea
r
i
t
h
m
e
t
i
co
p
e
r
a
t
o
r
si
nMATLAB.
A
d
d
i
t
i
o
n(wbena
d
d
i
n
gm
a
t
r
i
c
c
s
/
a
r
r
a
y
ss
i
z
em
u
s
tm
a
t
c
h
)
Sub
a
c
t
i
o
n(
s
a
m
ea
sa
b
o
v
e
)
* Multiplication (the subscripts ofarrays must be consistent when
m
u
l
t
i
p
l
y
i
n
gt
h
e
m
)
/ D
iv
i
s
i
o
n
Power
Complcxc
o
u
g
a
t
t
r
a
n
s
p
o
s
e(
0 a
yt
r
a
n
s
p
o
s
e
)

3
1

>

3
2 I
N
T
R
O
D
U
C
T
I
O
N

% Fortran
thepower of MATLAB derives from its
> > a
bilityto handle matrices like numbers
>>C2 81 * A1 How about this?
>> C1 * C2
What about t
his?
% read the error message
>>
> > i
t is quite informative
>>D1
[
1 2]' D1 is a column vector
>>C3
[
C
1 D1] C1 is augmented by an extra column
>>C3
[
C
3 C2(3
:)] Note
is an assignment
> > m
eans do the right hand side and overwrite the
> > o
ld information in C3 with the result
> > o
f the right
> > h
and side calculation
> > On t
he right you are adding a row to current
> > m
atrix C3. This row has the value of the third
> > r
ow of C2 -Notice the procedure of
> > i
dentifying the third row. The colon
>> % represents all the columns
permissible multiplication
>>C4 C2 * C3
> > N
ote the presence of a scaling factor
> > i
n the displayed output
>>CS C2 * C3
s
eems to multiplyl
> > I
s there a difference between C4 and CS?
>> % The * represents the product of each element
> > o
f C2 multiplied with the corresponding
> > e
lement of C3
>>C6 inverse(C2)
% find the inverse of C2
> > a
pparently inverse is not a command in MATLAB
> > i
f command name is known it is easy to obain
> > h
elp
>> lookfor inverse this command will find all files
> > w
here it comes across the word "inverse" in
> > t
he initial comment lines
> > T
he command we need appears to be INV which
> > s
ays Inverse of a Matrix
>> % The actual command is in lower case. To find
> > o
ut how to use it -Now
% shows how to use the command
>> help inv
>> inv(C2)
i
nverse of C2
>> for i "
' 1:20
f(i)
i
"
'
2
;
end
> > T
his is an example of a for loop

1
.
2

>>

> >

> >

> >

I
N
T
R
O
D
U
C
n
O
NT
OMAn.AD

3
3

the index ranges from 1 to 20 in steps of


1(default)
> > t
he loop is terminated with "end"
> > t
heprompt does not appear until "end" is
> > e
ntered
>> plot(sin(0.01*f)
cos(0.03 f
))
>> xlabel('sin(O.Ol*f)') strings appear in single
> > q
uotes
>> ylabel('cos(0.03*f)')
>> legend ('Example')
>> title ('A Plot Example')
>> grid
> > T
heprevious set of commands will create plot
> > l
abel axeswrite a legendtitle and grid the
> > p
lot
>> exit finishedwih MATLAB
T
h
i
sc
o
m
p
l
e
t
e
st
h
ef
i
r
s
ts
e
s
s
i
o
nw
i
t
hM
A
T
L
A
B
.A
d
d
i
t
i
o
na
1commandsa
n
df
e
a
t
u
r
e
sw
i
l
l
bee
n
c
o
u
n
t
e
r
e
dt
h
r
o
u
g
h
o
u
tt
h
eb
o
o
k
.I
ni
ss
e
s
s
i
o
n
i
ti
se
v
i
d
e
n
t MA
.AB
a
ows
l
e
a
s
ym
a
n
i
p
u
l
a
t
i
o
no
fm
a
t
r
i
c
e
s
d
e
f
i
n
i
t
e
l
yi
nr
e
l
a
t
i
o
nt
oo
t
h
e
rp
r
o
g
r
a
m
m
i
n
gl
a
n
g
u
a
g
e
s
.
P
l
o
t
t
i
n
gi
sn
o
td
i
f
f
i
c
u
l
te
i
t
h
e
r
.The
s
ea
d
v
a
n
t
a
g
e
sa
r
q
u
i
t
es
u
b
s
t
a
n
t
i
a
li
nt
h
es
u
b
j
e
c
to
f
o
p
t
i
m
i
z
a
t
i
o
n
.1
n n
e
x
ts
e
s
s
i
o
nwewi
l
1u
s
et
h
ee
d
i
t
o
r10I1c
c
o
m
p
l
i
s
hesame
t
h
r
o
u
g
hs
c
r
i
p
t
s
.T
h
i
ss
e
s
s
i
o
ni
n
t
r
o
d
u
c
e
d
MATLABCommandwindowa
n
dW
o
r
k
s
p
a
c
e

Variableassignment

B
a
s
i
cm
a
t
r
i
xo
p
e
r
a
t
i
o
n
s
A
c
c
e
s
s
i
n
gr
o
w
sa
n
dc
o
l
u
m
n
s
S
u
p
p
r
e
s
s
i
n
ge
c
h
o
e
s

i
n
v
e
r
s
ecommands
who
.
*m
u
l
t
i
p
l
i
c
a
t
i
o
n
B
a
s
i
cp
l
o
t
t
i
n
gcommands
1
.
2
.
4 UsingtheE
d
i
t
o
r
I
nt
h
i
ss
e
c
t
i
o
n

wew
i
l
lu
s
et
h
ee
d
i
t
o
rt
o
t
e dr
u
naMATLABas
c
r
i
p
tf
i
l
e
.Norma
1
l
y
e
e
d
i
t
o
ri
su
s
e
dt
og
e
n
e
r
a
t
et
w
ok
i
n
d
sofMA
.A
Bf
i
l
e
s
.1 ef
i
l
e
sa
r
et
e
r
m
e
ds
c
r
i
p
t
f
t
l
e
s

a
n
d
f
u
n
c
t
i
o
n
f
t
l
e
s
.A
I
o
u
g
hb
o
o
f
t
h
s
ef
i
I
e
sc
o
n
nM
ATLABcommandsl
i
k
ee
o
n
e
s
weh
a
v
eu
s
e
d

t
h
es
e
c
o
n
dtypeo
ff
i
I
e
sn
e
e
d
st
ob
eo
r
g
a
n
i
z
e
di
nas
p
e
c
i
f
i
e
df
o
r
m
a
.
tBo

f
i
l
et
y
ss
h
o
u
l
dh
a
v
et
h
ee
x
t
e
n
s
i
o
n.
m
.A
l
t
h
o
u
g
ht
h
e
s
ef
i
l
e
sa
r
eASCIIt
e
x
tf
i
I

e
g
e
n
e
r
i
c.ms
h
o
u
l
dbe e
d
b
e
c
a
e
MATLABs
h
e
sf
o
ri
s
e
x
s
i
o
n
.

h
i
se
x
t
e
n
s
i
o
ni
s

q
uetoMA
.AB.1cs
c
r
i
p
tf
i
l
ec
o
n
t
a
i
n
saI
i
s
to
fMA
T
L
A
Bcommands a
r
e
e
x
e
d
i
ns
e
q
u
e
n
c
e
.Th
i
si
sd
i
f
f
e
r
e
n
tmt
h
ei
n
t
e
r
a
c
t
i
v
es
e
s
s
i
o
no
ft
h
ep
r
e
v
i
o
u
ss
e
c
t
i
o
nw
h
e
r
e

34

1
.
2I
N
T
R
O
D
U
C
T
I
O
NTOM
A
T
L
A
D 35

I
N
T
R
O
D
U
C
T
I
O
N

M
A
1.ABndedt
o chconunandi
m
m
e
d
i
a
t
e
l
y
.Th
es
c
r
i
p
tf
i
l
ei
smoreu
s
e
f
u
lwhen
t
h
e
r
ea
r
m
anyconun
d
s n
e
e
dt
ob
eex
u
t
e
dt
oa
c
mplishsomeobjectivelike
r
u
n
n
i
n
ga
no
p
t
i
m
i
z
a
t
i
o
nt
e
c
h
n
i
q
u
e
.I
ti
si
m
p
o
r
t
a
n
tt
oremembert
h
a
tMATLABa
I
l
o
w
syou
t
os
w
i
t
c
hb
a
c
k
i
n
t
e
r
a
c
t
i
v
emodea
ta
n
yt
i
m
ebyj
u
s
tt
y
p
i
n
gcommandsi
neworkspace
windowl
i
k
ei
nep
v
i
o
u
ss
e
c
t
i
o
n
.
MATLABp
r
o
v
i
d
e
sane
d
i
t
o
rf
o
rc
r
e
a
t
i
n
gt
h
e
s
ef
i
l
e
sont
h
ePCp
l
a
t
f
o
r
m
.T
h
i
se
d
i
t
o
r
u
r
r
e
n
tva
Iu
e
s
u
s
e
sc
o
l
o
rt
oi
d
e
n
t
i
f
yMATLABs
t
a
t
e
m
e
n
t
sande
l
e
m
e
n
t
s
.I
tp
r
o
v
i
d
e
sec
r
ea
v
a
i
l
a
b
l
ei
nt
h
ew
o
r
k
s
p
a
c
e
)whent
h
emousei
so
v
e
r
ofthev
a
r
i
a
b
l
e
s(
a
f
t
e
r eya
t
h
ev
a
r
i
a
b
l
enamei
nt
h
ee
d
i
t
o
r
.Th
e
l
t
wowayst
oa
c
c
e
s
see
d
i
t
o
rt
h
r
o
u
g
ht
h
e
MATLABCommandwindowont
h
eP
C
.5
t
a
r
tMATLAB.T
h
i
sw
i
l
lopenaMATLAB
Commando
rWorkspacewindow.I
nt
h
i
swindowt
h
ee
d
i
t
o
rc
a
nb
es
t
a
r
t
e
dbyu
s
i
n
g
t
h
emenuo
ret
o
o
l
b
a
r
.On t
h
eF
i
l
emenuc
l
i
c
konNewandc
h
o
o
s
eMf
i
Ic
.
A
l
t
e
m
a
t
e
l
y
c
l
i
c
kont
h
eI
e
f
t
m
o
s
ti
c
o
nont
h
et
o
o
l
b
a
r(
et
o
o
l
t
i
pr
e
a
d
sNewF
i
l
e
)
.The
i
c
o
nf
o
rt
h
ee
d
i
t
o
rc
a
na
I
s
obep
l
a
c
e
dont
h
ed
e
s
k
t
o
p
i
nwhichc
a
s
et
h
ee
d
i
t
o
rc
a
nbe

aMA
TlABCommandw
indoww
i
l
l
s
t
a
r
t
e
dbyd
o
u
b
l
e
c
l
i
c
k
i
n
gt
h
ei
c
o
n
.I
nt
h
i
se
v
e
n
t
ee
d
i
t
o
rp
r
o
v
i
d
e
si
t
sownwindowf
o
re
n
t
e
r
i
n
g s
c
r
i
p
ts
t
a
t
e
m
e
n
t
s
.
n
o
tbeo
p
e
n
e
d
.Th
Att
h
i
sp
o
i
n
twea
r
er
e
a
d
yt
ou
s
et
h
ee
d
i
t
o
r
.Makes
u
r
eyou ad commentsand
u
n
d
e
r
s
t
a
n
dthem
a
st
h
e
s
ep
r
o
c
e
d
u sw
i
l
lbeu
s
e
do
f
t
e
n
.Usingt
h
ee
d
i
t
o
ri
m
p
l
i
e
swe
w
i
l
lbew
o
r
k
i
n
gw
i
t
hm
f
i
l
e
s
.
Thecommandsa
r
es
amea
si
n1
h
i
n
t
e
r
a
c
t
i
v
es
e
s
s
i
o
ne
x
c
e
p
tt
h
e
r
ei
snoMATLAB
I
lhavet
os
a
v
e
promptp
r
e
f
i
x
i
n
gt
h
ee
x
p
r
e
s
s
i
o
n
s
.Toe
x
e
c
u
t
et
h
e
s
ecommandsyouwi
t
u
sc
a
l
lt
h
ef
i
l
e
s
c
r
i
p
t
l
.
m
.The.me
x
t
e
n
s
i
o
nn
e
e
dn
o
tbet
y
p
e
di
f
y
o
u
themt
o
a
f
i
l.Le
a
r
eu
s
i
n
g MATLABe
d
i
t
o
r
.Youc
a
ns
a
v
et
h
ef
i
l
eu
s
i
n
gt
h
eSaveo
rSaveAs
commandfrommoste
d
i
t
o
r
s
.I
ti
si
m
p
o
r
t
t
t
oknowt
h
ef
u
l
lpa
t
oi
sf
i
l
e
.Le
tu
s
assumet
h
ep
a
t
hf
o
ref
i
l
ei
sC:OpCbookChl
s
c
r
i
p
t
l
.
m
.Note epa
h
e
r
ei
s
s
p
e
c
i
f
i
e
da
saPCp
a
t
hd
e
s
c
r
i
p
t
i
o
n
.Th
er
e
a
s
o
nwen
e
e
di
si
n
f
o
r
m
a
t
i
o
ni
st
oi
n
f
o
r
m
MATLABwheret
of
i
n
def
i
l
e
.Wedoi
si
neMATLABCommandwindow.
p
e
n
.OnPCswes
t
a
r
t
T
h
i
si
m
p
l
i
e
swes
h
o
u
l
dhaveeMATLABCommandwindowo
MATLABt
h
r
o
u
g
ht
h
ei
c
o
n
s
.I
nt
h
i
sbookwew
i
l
la
c
c
o
m
p
l
i
s
hmostt
a
s
k
sbyt
y
p
i
n
gi
n
r
ea
r
ea
l
t
e
m
a
t
e
t
h
eCommandwindowo
rt
h
r
o
u
g
hp
r
o
g
r
a
m
m
i
n
g
.l
nmanyi
n
s
t
a
n
c
e
se
waysu
s
i
n
gmenui
t
e
m
so
ret
o
o
l
b
a
r
.T
h
i
si
sl
e
f
tt
ot
h
er
e
a
d
e
rt
oe
x
p
e
r
i
m
e
n
ta
n
d
d
i
s
c
o
v
e
r
.I
nt
h
eMATLABCommandwindowu
s
ea
ddpathcommandt
oi
n
f
o
r
m
MATLABo
ft
h
el
o
c
a
t
i
o
noft
h
ef
i
l
e
:
>> addpath C:Opt_bookCh1

Thes
c
r
i
p
tt
h
8
1w
i
l
lbec
r
e
a
t
e
dands
a
v
e
di
nSCT
t
l
.
mcanber
u
nbyt
y
p
i
n
g(
n
o
t
et
h
a
t
ee
x
t
e
n
s
i
o
ni
so
m
i
t
t
e
d
)
>> scriptl

Tou
n
d
e
r
s
t
a
n
danda
s
s
o
c
i
a
t
ew
i
t
ht
h
ep
r
o
g
r
a
r
n
m
i
n
gc
o
n
c
e
p
t
sembeddedi
n
.t
h
es
c
r
i
p
t

p
a
r
t
i
c
u
l
a
r
l
yf
o
rp
e
r
s
o
n
sw
i
t
hI
im
i
t
e
dp
r
o
g
r
a
r
n
m
i
n
ge
x
p
e
r
i
e
n
c
e
i
ti
srecommendedt
o
r
u
nt
h
es
c
r
i
p
ta
f
t
e
rab
l
o
c
ko
fs
t
a
t
e
m
e
n
t
shaveb
e
e
nw
r
i
t
t
e
nr
a
e
rt
h
a
nt
y
p
i
n
gef
i
l
e

i
ni
t
se
n
t
i
r
e
t
y
.P
r
i
o
rt
or
u
n
n
i
n
gt
h
es
c
r
i
p
tyouw
i
l
lhavet
os
a
v
et
h
es
c
r
i
p
te
a
c
ht
i
m
e50
t
h
ec
h
a
n
g
e
si
nfiJea
r
er
e
c
o
r
d
e
dand t
h
ec
h
a
n
g
e
sa
r
ec
u
r
r
e
nt
.Another
recommendationi
st
od
e
l
i
b
e
r
a
t
e
l
ym
i
s
'
p
esomes
t
a
t
e
m
e
n
t
sa
n
da
t
t
e
m
p
tt
odebugt
h
e
u
r
i
n
g u
t
i
o
n
.
e
r
r
o
rg
e
n
e
r
a
t
e
dbyeMATLAsdebuggerd

C
r
e
a
t
i
n
gtheScr

tM

.
f
l
l
e(Thef
o
l
l
o
w
i
n
gw
i
l
lbet
y
p
e
d
l
s
a
v
i
naf
i
l

.
)
example of using script
A1
[
1 2 31;
[
4 5 61;
A2
the commands not terminated with semi-colon will
display information on the screen
A = [A1; A2]
B
[
A
1
' A2']
C A*B

now re-create the matrix and perform matrix


multiplication as in other programming languages
example of for loop
variable i ranges from 1 to 3 in
for i
1
3
steps of 1 (default)
i
)
i;
a1(
1
loops must be closed with end
end
a1

for i
6:-1:4 note loop is decremented
i-3)
i;
filling vector from rear
a2(1
end
a2
creatingmatrix A and B (called AA and BB here)
for i
1:3
i
)
a1(
1
i
); assign a1 to AA.
AA(l
AA(2
i
)
a2(
1
i
);
1
)
a1(
1
i
);
BB(
i
2) a2(
1
i
);
BB(
i
end
the same can be accomplished by AA (
1:)
a1
without the for loop
AA
print the value of AA in the window
BB
who
list all the variables in the workspace
consider code for Matrix multiplication
% which Matlab does so easily

36 I
N
T
R
O
D
U
C
T
I
O
N

1
.
2

multiply two matrices (column of first marix must


match row of second matrix)
size of AA
szAA size(AA)
szBB size(
B
B
);
if (szAA(l
2
)
szBB(l
l
)
)
only in c
olumn of AA match the rows of BB
for i ~ l:szAA(l
l
)
1:szBB(l
2)
for j
CC(
i
j)
0.
0; initializevalue to zero
for k 1:szAA(l
2)
j) c CC(i
j) + AA(i
k)*BB(k
j);
CC(i
k -l
oop
end
end j - loop
end i - loop
end if - loop
CC
N
ote the power of MAT~B derives from its ability to
h
andle matrices very easily
t
his completes the script session

I
N
T
R
O
D
U
C
T
I
O
NTOMA AD

37

C
l
e
a
rs
t
a
l
e
m
e
n
t

1
.
2
.
5 CreatlngaCodeSnlppet
I
nt
h
i
ss
e
c
t
i
o
n
wew
i
l
le
x
a
m
i
n
et
h
e0e
rt
y
p
co
fm
f
i
l w
h
i
c
hi
sc
al
1e
dt
h
e
f
u
n
c
t
i
o
n
o
rt
h
o
s
ef
a
m
i
l
i
a
rw
i
l
ho
t
h
e
rprogrammingl
a
n
g
u
a
g
e
sI
ik
eCJ
a
v
ao
r
m
f
i
l
e
.F
T
R
A
Ne
s
ef
i
l
e
sr
e
p
r
e
s
e
n
tf
u
n
c
t
i
o
n
so
rs
u
b
r
o
u
t
i
n
e
s
.Theya
r
ep
r
i
m
a
r
i
l
yu
s
c
dt
'
O
F'OR
h
a
n
d
l
es
'
O
mes
p
e
c
i
f
i
ct
y
p
e
so
fc
a
l
c
u
l
a
t
i
o
n
s
.Thc
ya
l
s
op
r
'
Ov
i
d
eawayf
o
rt
h
em
o
d
u
l
a
r
o
d
er
e
u
s
e
.T
h
e
s
ec
o
d
em
o
d
u
l
e
sa
r

u
s
e
dbyb
e
i
n
g
d
e
v
e
l
o
p
m
e
n
to
fc
o
d
ea
sw
e
l
l c
s
a
yt
h
r
o
u
g
has
c
r
i
p
tf
i
l
ewel
o
o
k
e
da
t
c
a
l
l
c
do
rr
e
f
c
r
r
e
di
no
t
h
e
rs
e
c
t
i
o
n
so
ft
h
ec
o
d
c
am/
c
o
d
e
.The
e
a
r
l
i
e
r
.Thec
o
d
et
h
a
tc
a
l
l
st
h
cf
u
n
c
t
i
o
nm
f
i
l
ei
sc
a
l
l
e
dt
h
ec
a
l
l
i
n
gp
r
o
g
r
'
O
p
i
n
gt
h
ef
u
n
c
t
i
o
nm
f
i
l
ca
r
c(
1
)w
h
a
ti
n
p
u
1i
sn
e
c
e
s
s
a
r
y
e
s
s
e
n
t
i
a
lp
a
r
a
m
c
t
c
r
si
nd
c
v
e
l
f
o
rt
h
ec
a
I
c
u
l
a
t
i
o
n
s(
2
)w
h
a
ts
p
e
c
i
f
i
cc
a
l
c
u
l
a
t
i
o
n
smustt
a
k
ep
l
a
c
ea
n
d(
3
)w
h
a
t
'
O
g
r
a
m
.MATLABr
e
q
u
i
r
e
st
h
eS
l
r
u
c
t
u
r
e
i
n
f
o
r
m
a
t
i
o
nm
u
s
tber
e
t
u
m
e
dt
ot
h
ec
a
l
l
i
n
gp
r
u
n
c
t
i
o
nmf
i
l
et
of
o
l
l
o
wap
r
e
s
c
r
i
b
e
df
o
r
m
a
t
.
o
fef
Wew
i
l
lu
s
e1
h
ee
d
i
1
0
r1
0d
e
v
e
l
'
OP
af
u
n
c
1
i
o
nm
f
i
l
et
h
a
1wi
I
1p
c
r
f
o
r
map
o
l
y
n
o
m
i
a
l
tI
Ir
e
q
u
i
r
e
sas
e
to
f d
a
t
a
r
e
p
r
e
s
e
n
1
i
n
gac
u
r
v
et
h
a
tn
e
e
d
st
ob
ef
i
1
1
0
g
e
1
h
e
r
c
u
r
v
e
f
i.
h
eo
r
d
e
ro
ft
h
ep
o
l
y
n
o
m
i
a
lt
ob
ef
it
.T
h
i
se
x
e
r
c
i
s
ci
sc
a
l
l
c
dc
u
r
v
ef
i
u
i
n
g
.I
n
w
i t
n
c
o
n
s
t
r
a
i
n
e
d
C
h
a
p
t
c
r6 s
u
c
hap
r
o
b
l
e
mw
i
l
lb
ei
d
e
n
1
i
f
i
c
da
s ap
r
o
b
l
e
mi
nu
o
rnowt
h
ec
a
l
c
u
l
a
1
i
o
n
sn
e
c
e
s
s
a
r
yt
oa
c
c
o
m
p
l
i
s
ht
h
ee
x
e
r
c
i
s
ea
r
e
o
p
t
i
m
i
z
a
t
i
o
n
.F
c
o
n
s
i
d
e
r
e
dk
n
o
w
n
.I
ti
n
v
o
l
v
e
ss
o
l
v
i
n
gal
i
n
e
a
re
q
u
a
t
i
o
nw
i
t
henomtalm
a
t
r
l
xa
n
da
dv
c
c
t
o
ro
b
t
a
i
n
e
du
s
i
n
gt
h
ed
a
t
ap
o
i
n
t
s
.
'
heo
u
t
p
u
tf
r
o
mt
h
em
f
i
l
ew
i
l
lb
e
r
i
g
h
t
h
e
coe
i
c
i
e
n
t
sr
e
p
r
e
s
e
n
t
i
n
gep
o
l
y
n
o
r
n
i
a
l
.
t
h
ef
i
r
s
tl
i
n
eo
ft
h
i
sf
i
l
em
u
s
tb
ef
o
r
m
a
t
t
e
da
s
B
e
f
o
r
ewes
t
a
r
tt
od
e
v
e
l
o
pt
h
ec
o
d
e
yM
A
T
L
A
B
.I
nt
h
ef
i
r
s
tl
i
n
e

t
h
ef
i
r
s
tw'Or
ds
t
a
r
t
i
r
i
gumef
i
r
s
tc
o
l
u
m
ni
s
s
p
e
c
i
f
i
e
db
cw
o
r
d
f
u
n
c
t
i
o
n
.I
ti
sf
o
l
l
o
w
e
db
yt
h
es
e
to
frelump
a
r
a
m
e
t
e
r
s(
r
e
t
u
m
v
a
l
)
.N
e
x
t

en
ame(
m
y
p
o
l
y
f
i
t
)'Oft
h
ef
u
n
c
t
i
o
nw
i
t
ht
h
ep
a
r
a
m
e
t
e
r
sp
a
s
s
e
dt
ot
h
cf
u
n
c
t
i
o
nw
i
t
h
i
n
s
e
s(X

YN
)
. l
ef
i
l
em
u
s
tb
es
a
v
e
da
sname.m(
m
y
p
o
l
y
f
i
t
.
m
)
.Thecomme

p e
b
e
1
w
e
e
nt
h
ef
i
r
s
1l
i
n
ea
n
dt
h
ef
i
r
s
1e
x
e
c
u
t
a
b
l
cs
t
a
t
e
m
e
n
twi
I
1a
p
p
c
i
fy
o
u1
y
p
eh
e
l
p
l
Ia
me (
h
e
l
pm
y
p
o
l
y
f
i
t
)i
nt
h
c Command w
i
n
d
o
w
. The r
e
a
s
o
nf
o
rt
h
cn
a
r
n
e
.
tm i
st
h
a
tMATLABh
a
sabui
1
ti
nf
u
n
cl
Io
np
o
l
y
f
i
t
.Opent
h
ee
d
i
t
o
r1
0c
r
e
a
t
e
m
y
p
o
l
y
f
i
t
h
ef
i
l
ec
o
n
t
a
i
n
i
n
gt
h
ef
o
l
l
o
w
i
n
gi
n
f
o
r
m
a
t
i
o
n
:

S
a
v
et
h
ea
b
o
v
ef
i
)e(
s
c
r
i
p
t
l
.m
)
.Addt
h
ed
i
r
e
c
t
o
r
yt
ot
h
eMATLABp
a
t
h i
n
d
i
c
a
t
e
d
b
e
f
o
r
e
.Runt
h
es
c
r
i
p
tf
i
l
ebyt
y
p
i
n
gs
c
r
i
p
t
Ja
tt
h
ecommandpromp
.
tThccommands
r
o
m
p
ti
nt
h
cCommand
s
h
o
u
l
da
l
le
x
e
c
u
t
ea
n
dy
o
us
h
o
u
l
df
m
a
l
l
ys
e
et
h
eMATLABp
w
i
n
d
o
w
.
N
o
t
e
: You c
a
na
l
w
a
y
sr
c
v
e
r
tt
ot
h
ei
n
t
c
r
a
c
t
i
v
e mode b
yd
i
r
e
c
t
l
ye
n
t
e
r
i
n
g
f
t
e
rt
h
epromp
.
tI
nt
h
eCommandw
i
n
d
o
w
:
commandsineMATLABwindowa
>> who
>> clearC

% discards thevariableC fromtheworkspace


use with caution. Values cannot be recovered
>> help clear
>> exit

> >

Thi
ss
e
s
s
i
o
ni
l
1u
s
t
r
a
t
e
d
:

function returnval
Usco
f
t
h
e
d
i
t
o
r

Creatingascript

ip
t
R
u
n
n
i
n
gas
cI

E
r
r
o
rd
e
b
u
g
g
i
n
g(recommcndcda
c
t
i
v
i
t
y
)

Programmingconcepts

. Lo
opc
o
n
s
t
r
u
c
t
s
i
fa
n
df
o
rl
o
o
p
s
Loopv
a
r
i
a
b
l
ea
n
di
n
c
r
e
m
e
n
t
s

Arrayaccess

mypolyfit(XYN)

These comments will appear when the user types


helpmypolyfit in the Command window
T
his space is intended to inform the user how to
i
nteract with the programwhat it does
w
hat are the input and output parameters

Least square error fit of polynomial of order N

xy - Data found in XY

returns the vector of coefficients starting from


%
the constant term

..

\U! t:J.\i. ~~tn::~!;:)_.~r..i:1 1


:
'
.

~.
:
i '~ S~-:t'~!{j!':it~\.J\".'l\"~' ~
j'J

~ f:~' . j~ ;';~E

1
.
2I
N
T
R
O
O
U
C
T
I
O
NTOMAl1.AB

l
N
T
R
O
D
u
c
n
O
N

38

end
e
nd i 100p
i
f the x-points are distinct then inverse is not a
prob1em

or i

1:N+1
a(i) = 0.0; initia1ize the coefficient to zero
end

X
*

x
*
x

2J

n
U

==

..

oX2

10 J )

>>coeff mypolyfit(XY1
3)
>> ac
ubic po1ynomia1 was de1iberate1y created to
>> % check the results. You shou1d get back
>> t
he coefficients you used to generate the curve
>> t
his is a good test of the program
>> L
et us create another examp1e
1
)
XYl(:
1
); same first co1umn
>> XY2(:
>> XY2(:
2
)
20 + exp(XY2(:
1
))
>> n
ote the power of Mat1ab - a vector fi11
>> an
ew set of xy-data
>>coeff1 mypolyfit(XY23) cubic po1y. again
>> helpmypoly
it
>> y
ou shou1d see the statements you set up
>> I
n the next exercise we wi11 create a script
>> f
i1e that wi11 run a program for po1ynomia1
>> c
urve-fittingwe wi11 save XY2 so that it can
>> b
eused again

for j 1:N+1;
if j >= i ca1cu1atingupper diagona1 terms
j1
);
power
(i-1) + (
A(i
j
)
0.0
i
nitialize
for k l:NDATA; sumover data points
j
) = A(i
j
) + XY(k
l)^power;
A(i
end k loop
end
c
10se if statement
A
i
) = A(i
j
) exp10itingMatrix symmetry
end
e
nd j 100p

1
1

E
= 4

b(i)
end

100p over al1 data points


v
ariab1e is "11" (
e
1
)(
e)
2)*XY(11
1)^(i-1);
b(i) + XY(11
100p 11

ryy
a
o
xxx

for i 1:N+1;
b(i)
0.0;
for 11 l:NDATA;

usmuadu

The processing starts here.


The coefficients are obtained as solution to the
Linear A1gebra prob1em [
A
][
c
]
[
b
]
Matrix [
A
] is the Norma1 Matrix which contains
the sum of the powers of the independent variab1e
[
A
] is symmetric

y
p
o
l
y
f
u
.
m
.Tou
s
et
h
ef
u
n
c
t
i
o
nwew
i
l
ln
e
e
dt
oc
r
e
a
t
esomex
yd
a
t
a
S
a
v
et
h
ef
i
l
ea
sm
dt
h
c
nc
a
l
lt
h
em
y
p
o
l
y
f
i
tf
u
n
c
t
i
o
n
.S MA
1LABi
ncd
i
r
e
c
t
o
r
ytm
y
p
o
l
y
f
i
t
.m
r
e
s
i
d
e
si
n
.o
ra
d
dt
h
ed
i
r
e
c
t
o
r
yt
ot
h
ep
a
t
h
.I
nt
h
eM A11.ABCommandw
i
n
d
o
w
:The
i
st
y
p
e
di
nt
h
ecommandw
i
n
d
o
w
.
f
o
l
1owingc
o
d

n
e

fprintf'(
Too few data points for fit
')
returnva1 a;
reurn
end

inv(A)*b';

if NDATA < 2*N

returnva1

sz size(
X
Y
);
NDATA sZ(l
l); number of data points - rows 0
xymatrix
if NDATA == 0
fprintf('There is no data to fit');
reurnva1 a;
z
erova1ue returned
return;
% return back to ca11ing program
end

3
9

save C:Opt_bookChlXY2.dat -ascii doub1e


this wi11 save the fi1e as an ascii text fi1e
withdoub1e precision va1ues

T
h
i
sc
o
n
c
1u
d
e
see
x
e
r
c
i
s
ew
h
e
r
eac
o
d
es
p
W w
r
i
t
t
e
nto l
c
a
t
eE
C
fficientsofthepolynooualusedt
of
i
tac
u
r
v
t
osome d
a
t
a
. et
y
p
eo
ff
i
l
ci
st
h
c

40

I
N
T
R
O
D
U
C
T
I
O
N

f
i
m
c
t
i
o
nmf
i
l
e
.I
tn
beu
s
e
di
nac
e
r
t
a
i
nw
a
y
.l
ec
o
d
ewas t
e
du
s
i
n
gac
u
b
i
c
p
o
l
y
n
o
m
i
a
l
.N
o
n
p
o
l
y
n
o
m
i
a
ld
a
t
aw
c
r
ea
1s
ot
e
s
t
e
d
.Th
ed
a
t
aw
e
r
es
a
v
e
df
o
rl
a
t
e
ru
s
e
.

1
.
2I
N
T
R
O
D
U
C
T
I
O
NTOMATLAB

1
.
2
.
6 Creatlngaprogram
h i
ss
e
c
t
i
o
nap
r
o
g
r
a
m w
i
l
l ad d
a

a
tc
u
r
v
e
f
i
tt
h
ed
a
t
au
s
i
n
gap
o
l
y
n
o
m
i
a
l
a
n
d
o
r
i
g
i
n
a
la
n
df
i
d
da
p
r
u
l
l
yw
i
l
lb
ed
e
v
e
l
o
p
e
d
.Th
e t
aw
i
l
lb
er
e
a
d
u
s
i
n
gaU
s
e
rIn(UI)windowa
n
dad
i
a
l
o
gb
o
x
.Th
ec
o
d
ew
i
l
lb
ed
e
v
e
l
o
p
e
d a
sg
i
v
e
su
sa
no
p
p
o
r
t
u
n
i
t
yt
or
e
v
i
s
i
tm
o
s
to
ft
h
ec
o
d
eweu
s
e
d r
l
i
e
rf
o
r
s
c
r
i
p
tf
i
l
c
.Thi
i
n
f
o
r
c
e
m
e
n
.
tWcw
i
l
la
l
s
ou
s
et
h
cf
u
n
c
t
i
o
nm-f
i
I
ec
r
e
a
t
e
di
n p
r
e
v
i
o
u
se
x
e
r
c
i
s
e
.

mpare

eval(10adpathfi1e);
eva1uates the string enc10sed -which inc1udes the
M
at1ab command 10ad. This wi11 import the xy-data
t
he data wi11 be avai1ab1e in the workspace as a
v
ariab1ewith the same name as the fi1ename
w
ithout the extension (this assumes you se1ected
t
he xy-data using the fi1e se1ection box)

b
a
s
i
cm
e
t
h
o
di
st
op
r
o
m
p
tu
s
e
r
sf
o
ri
n
f
o
r
m
a
t
i
o
na
t p
r
o
m
p
ti
nt
h
eCommandw
i
n
d
o
w
.
i
seq
u
i
c
k
e
s
.
tT
h
isi
sp
r
o
b
a
b
l
yw
h
a
ty
o
uw
i
l
lu
s
ewhend
e
v
e
l
o
p
i
n
gt
h
ec
o
d
e
.Once

t
h
ed
eh
a
sb
e
e
nt
e
s
t
e
d
d
e
p
e
n
c
l
i
n
go
nu
s
e
f
u
l
n
e
s
si
tm
i
g
h
tb
er
e
l
e
v
a
n
t
c
o
n
s
i
d
e
ru
s
i
n
g
I
e
m
e
n
t
si
I
k
ei
n
p
u
tb
o
x
e
sa
n
df
i
l
es
e
l
e
c
t
i
o
nb
o
x
e
s
.l
i
sb
o
o
k
mores
o
p
h
i
s
t
i
c
a
t
e
dc
u
s
t
o
me
i
l
et
h
ei
n
p
u
te
l
e
m
e
n
t
s
w
i
l
lc
o
n
t
i
n
u
et
ou
s
et
h
e
s
ee
l
e
m
e
n
t
st
h
r
o
u
g
h
o
u
ta
sa
p
p
r
o
p
r
i
a
t
e
.Wh
u
s
i
ni
sc
o
d
ea
r
cncwcommands
t
h
er
e
s
to
ft
h
ep
r
o
g
r
a
mw
i
l
lm
o
s
t
l
yu
s
ecommands
t
h
a
v
cb
c
c
ni
n
t
r
o
d
u
d
r
l
i
e
r
.I
ns
u
e
n
t
i
a
lo
r
d
e
rt
h
ee
v
c
n
l
si
n sp
r
o
g
r
a
ma
r
e
:(
1
)t
o
a
t
as
a
v
e
de
a
r
l
i
e
ru
s
i
n
gaf
i
l
cs
c
l
e
c
t
i
o
nb
o
x
(
2
)t
or
e
a
deo
r
d
e
ro
ff
i
tu
s
i
n
g
r
e
a
dt
h
e d
a
nI
np
u
td
i
a
l
o
gb
o
x(
3
)t
ou
s
et
h
em
y
p
o
l
y
f
i
tf
u
n
c
t
i
o
nd
c
v
e
l
o
p
e
di
nt
h
el
a
s
ts
e
c
t
i
o
nt
o
o
b
t
a
t
h
e
c
f
f
i
c
i
e
n
t
s
(
4
)t
oo
b
t
a
i
nec
i
n
a
o
ft
h
ef
i
t
t
c
dc
v
e

(
5
)t
og
r
a
p
r
u
c
a
l
l
y
r
i
g
i
n
a
l df
i
t
t
d
a

a
ta
n
d(
6
)
tonef
i
t
t
e
da
uracyone
f
i
m
c
o
m
p
a
r
e o
i
t
s
e
l
n
ews
c
r
i
p
tf
i
l
ew
i
l
lb
ec
a
l
l
e
dprog
.
.
.
J
1
j
m
.

newname strrep(fi1e'.dat'");
newname i
s a string variab1e which contains the
s
tring file stripped of the .da extension. Thi
i
s a string rep1acement command
x

S
t
a
r
tt
h
ec
d
i
t
o
rt
oc
r
e
a
t
et
h
ef
i
l
cc
a
l
l
e
dp
'

the stri variab1epath wi11 have the path


information the defau1t directory pointed wi11
differdepending on the p1atform

if isstr(file) if a fi1e is se1ected


10adpathfi1e
['load'
path fi1e];
1
0adpathfi1e is a st ingvariab1e concatenated
with three s
trings "10ad
" path and fi1e
n
ote the space after 10ad is important

l
e
r
s
e
v
e
r
a
lw
a
y
sf
o
ry
o
uo
rt
h
eu
s
e
r
st
oi
n
t
e
r
a
c
tw
i
t
ht
h
ec
o
d
ey
o
ud
e
v
e
l
o
p
.1

eval(newname); assigns the imported data to x


just for convenience. The above step is not
necessary

program for fitting a po1ynomia1 curve to xy data


from
App1iedOptimization using Mat1ab
Dr. P.Venkataraman

1)); numberof data points


NDATA length(x(:
clear path 10adpathfi1e newname
g
et rid of these variab1es to recover memory
end

Chapter 1Section 1.2.6


The program 100ks for a fi1e with two co1umn ascii
data with extension .dat. The order of the curve is
obtained from user. The origina1 and fitted data
are compared with re1evant information disp1ayed on
the same figure. The program demonstrates the use
of the fi1e se1ection boxan input dia10g box
creating specia1 text strings and disp1aying them

Note: if a fi1e is not se1ected nothing is being


done
U
se of an input dia10g box to get the order
o
f po1ynomia1 to be fitted
PROMPT {'Enter the Order of the Curve'};
P
ROMPT is a stringArray with one e
1
e
r

n
ote the curlybrackets
TITLE 'Order of the Po1ynomia1 to be Fitted';
as
tring variab1e
LINENO 1
; a data variab1e

[fi1e
path] =uigetfi1e('*.dat'
'A11 Fi1es'200
200);
uigetfi1e opens a fi1e se1ection box
c
heckout he1p uigetfi1e
t
he string variable fi1e will hold the filename

4
1

getva1

inputdlg(PROMPTTTLELINENO);

42 I
N
T
R
O
D
U
C
T
I
O
N

TlA9
1
.
2I
N
T
R
O
D
U
C
T
I
O
NT
OM
A

the input dialog captures the user input in getval


geval is a string Array
check help inputdlg for more information

l});
no str2num(getval{l
t
he string is converted to a number- the order
clear PROMPT TTLE LNENO

deletingvariables

% call function mypolyfit and obtain the coefficients


coeff mypolyfit(x
no);

generate the fitted curve and obtain the squared


e
rror
err2 0.0i
'
for i 1:
NDATA for each data point
for j l:no + 1
a(l
j
) = x(i
l)"(j-l);
end
y(i)
a*coeff; he data for he fitedcurve
err2 err2 + (x(i
2
) -y(i))*(x(i
2)-y(i));
he square error
end

p
lotting
plot (
x(:
1
)
x(:
2
)

ro'
x(:
1
)
y

b-');
% original data are red o's
f
itteddata is blue solid line

x
')i
xlabel(
ylabel(

y
');
strorder setstr(num2str(no));
c
onvert the order of curve o a string
s
ame as getval if you have not cleared it
s
etstrassigns the string to strorder
titlestr
['polynomial curvefit or order'
srorder 'of file'file)i
t
he three dots at the end are continuation marks
t
he title will have the order and the file name
title(titlestr)
legend('original data''fitted data');

errstrl num2str(err2);
errstr2
['squared er or
gtext(errstr2);

43

this places the string errstr2 which is obtained


bycombining the string 'squared error' with
t
he string representing the value of the error
w
herever the mouse is clicked on the plot.
m
oving the mouse over the figure you should
s
ee location cross-hairs
clear strorder titlestr errstrl errstr2 a y x i j
clearNDATA no coeff XY2 file err2 getval
grid

This finishes the exercise

b
yf
i
r
s
tr
u
n
n
i
n
gMATLABi
n d
i
r
e
c
t
o
r
yw
h
e
r
et
h
e
s
ee
sa
r
e
.o
r
Runt
h
ep
r
o
g
r
a
a
d
d
i
n
gt
h
ep
t
o1
ate AtecommandprompttypeproLP.The
p
r
o
g
r
a
ms
h
o
u
l
dE
X
E
c
u
t
er
e
q
u
i
r
i
n
gu
s
e
ri
n
p
u
tt
h
r
o
u
g
ht
h
et
i
l
es
e
l
e
c
t
i
o
nbOLinput

a
n
df
i
na
l
1yd
i
s
p
l
a
y
i
n
gF
i
g
u
r
e1
.
1
1
.l
ea
p
p
e
a
n
c
emayb
es
l
i
g
h
t
l
yd
i
f
f
e
r
e
n
t
d
i
a
l
o
g
d
e
p
e
n
d
i
n
go
n p
l
a
'
o
r
mMATLABi
sb
e
i
n
gr
u
n
.
MT
h
i
sf
i
n
i
s
h
e
st
h
eMATLABs
e
c
t
i
o
no
ft
h
ec
h
a
p
t
e
r
.
T
l
ms
e
c
t
i
o
nl
I
a
si
n
t
r
o
d
u
c
e
d
A
T
L
A
Bi
nar
o
b
u
s
tm
a
n
n
e
r
.A b
r
o
a
dmng
o
fprogmmminge
x
p
e
r
i
e
n
c
eh
a
sb
e
e
n
m
l
t
l
a
t
e
di
nt
h
i
sc
h
a
p
t
e
r
A
Hm wcommandsh
a
v
eheni
d
e
r
u
i
n
c
dw
i
t
hab
d
e
f
o
m
m
e
n
t
s
.I
ti
si
m
p
o
r
t
a
n
t youu
s
et
h
eO
P
P
Ol
t
u
n
i
t
yt
ot
y
p
ei
n
e
x
p
l
a
n
a
t
i
o
ni
n c
EC
o
d
Ey
o
u
r
s
e
l
f
;
T
h
a
ti
st
h
eo
n
l
ywayt
h
eu
s
eofMAT1ABw
i
l
lbecomef
a
m
i
l
i
a
r
.
T
h
e
p
r
a
c
t
i
c
ea
1s
ow
i
l
l
l
e
a
dt
of
e
w
e
rs
y
n
t
a
xe
r
r
o
T
hew
r
i
t
i
n
go
fc
o
d
ew
i
l
ls
i
g
n
i
f
i
c
a
n
t
l
y
i
m
p
r
o
v
ey
o
u
ra
b
i
1
it
yt
od
e
b
u
ga
n
dt
r
o
u
b
l
e
s
h
o
o
t
.Wh
i
l
ei
sc
h
a
p
t
e
re
m
p
l
o
y
e
da

I
to
fo
r
d
e
r3o
ff
i
l
e
X
Y
2
.
d
a
t
P
o
l
y
n
o
m
i
a
lc
u
r
v
ef

errstrl]
x
F
l
g
u
r
e1
.
1
1O
r
i
g
l
n
a
la
n
df
l
l
l
e
dd
a
l
a
.

i44uu1uikJt:i

2
J
!
.
:;
l
;
'
f
l
.

44 INTRODUCTION
s
e
o
a
r
a
t
es
e
c
t
i
o
nont
h
eu
s
eo
fMATLABo
u
to
fn
s
i
t
y
sub uentchapmMe
i
z
e
dbyacompl
r
a
t
i
o
no
f
t
h
eu
s
m

ofMA.

T
T
L
A
B
CMa
o
p
t
i
m
i

z
a
t
i

o
n
.

u
r
u
u
z
;
z
t
m
:; l
r

;
:
:
;
t
;
PROBLEMS

m
a
t
h
e
m
a
t
i
c
a
lmodeJ
.
1
.
2 I
d
e
n
t
i
f
ys
e
v
e
r
a
lp
o
s
s
i
b
l
eo
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
e
m
sr
e
l
a
t
e
dana
i
r
c
r
a
f
t
.F
o
re
a
c
h
i
s
c
i
p
l
i
n
e
s w
i
l
lh
e
l
pe
s
t
a
b
l
i
s
ht
h
em
a
t
h
e
m
a
t
i
c
a
l
problemi
d
e
n
t
i
f
ya
l
led
mode
l
.
1
.
3 I

d
n
t
s
e

v p
o
s
s
i
b
l
e
0p
t
i
m
i
z
a
p
m

b
l
i

d
e
a
1
t
hed
i

s
c
i

p
l
i
n

GRAPHICALOPTIMIZATION

w
i
l
l

pe
s
i

s
l
h ma

l
a
n
a
t

modeJ
.
1

4 I
d
e
n
t
i
f
ys
e
v
e
r
a
1p
o
s
s
i
b
lo
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
e
m
sr
e
l
a
t
e
dt
oam
i
c
r
o
s
y
s
t
e
r
nu
s
e
d
f
o
rc
o
n
t
r
o
l
.
F
o
r
B
a
c
hp
r
o
b
l
e
mi
d
e
n
t
i
f
ya
l
lt
h
E
d
i
s
c
I
P
l
i
n
e
st
h
a
tw
i
l
lh
e
l
pe
s
t
a
b
l
i
s
h
em
a
t
h
e
m
a
t
i
c
a
lmode
.
l

o
v
e
r
h
a
n
g
i
n
gt
r
a
f
f
i
cl
i
g
h
t
.
1
.
8 D
e
f
i
n
eep
r
o
b
l
e
mandi
d
e
n
t
i
f
yam
a
t
h
e
m
a
t
i
ca
1modelf
o
rs
c
h
e
d
u
l
i
n
ga
n
d
o
o
t
i
m
i
z
a
t
i
o
no
ft
h
ed
a
i
l
yr
o
u
t
i
n
ea
c
t
i
v
i
t
y
.

P U 1
2

1
.
5 D
e
f
i
n
eap
r
o
b
l
e
mwit
1
Ir
e
s
p
e
c
tt
oy
o
u
ri
n
v
e
s
t
m
e
n
ti
nt
h
es
t
o
c
km
a
r
k
e
t
.D
e
s
c
r
i
b
e
en
a
t
u
r
e
o
f
t
h
e
m
a
t
h
e
m
a
t
i
c
a
lmodd1
.
6 D
e
f
i
n
e p
r
o
b
l
e
ma
n
de
s
t
a
b
l
i
s
h m
a
t
h
e
m
a
t
i
ca
1modelf
o
rt
h
eI
b
e
a
m
h
o
l
d
i
n
gupa
ni
n
d
e
p
e
n
d
e
n
ts
i
n
g
l
e
f
a
m
i
l
yhome.
1
.
7 D
e
f
i
n
et
h
ep
r
o
b
l
e
ma
n
di
d
e
n
t
i
f
yt
h
em
a
t
h
e
m
a
t
i
c
a
lmodel f
o
ra
noptimum

D
ne

ep
emf
o
ral
a
m
i
n
g
i
v
e
nasp
e
:
e
i

c
pump.
1

10 D
e
f
i
n
eac
h
e
m
i
c
a
le
n
g
i
n
e
e
r
i
n
gp
r
o
b
l
e
mt
omixv i
o
u
sm
i
x
t
u
r
e
so
f1
im
i
t
e
d
p
e
c
i
f
i
e
ddm
a
n
d
s
.
a
v
a
i
l
a
b
i
l
i
t
yt
omakes
p
e
c
i
f
i
e
dcompoundst
om ts
i

19U

1
m

T
h
i
sbooki
n
c
l
u
d
e
sal
a
r
g
enumbero
fcxamplesw
i
t
htwov
a
r
i
a
b
l
e
s
.T
w
o
v
a
r
i
a
b
l
e
p
r
o
b
l
e
m
sc
a
nb
ed
i
s
p
l
a
y
e
dg
r
a
p
h
i
c
a
l
l
yandt
h s
o
l
u
t
i
o
no
b
t
a
i
n
e
dbyi
n
s
p
e
c
t
i
o
n
.
Along w
i
t
ht
h
eu
s
e
f
u
l
n
e
s
so
fo
b
t
a
i
n
i
n
gt
h
es
o
l
u
t
i
o
nw
i
t
h
o
u
ta
p
p
l
y
i
n
g
a
t
i
c
a
lc
o
n
d
i
t
i
o
n
st
h
eg
r
a
p
h
i
c
a
lr
e
p
r
e
s
e
n
t
a
t
i
o
no
ft
h
eproblemp
r
o
v
i
d
e
sa
n
mathm
o
p
p
o
r
t
u
n
i
t
yt
oe
x
p
l
o
r
et
h
eg
e
o
m
e
t
r
yo
fmanyo
ft
h
en
u
m
e
r
i
c
a
lt
e
c
h
n
i
q
u
e
swe
examinel
a
t
e
r
.T
h
i
si
sn
e
c
e
s
s
a
r
yf
o
ri
n
t
u
i
t
i
v
e
l
yu
n
d
e
r
s
t
a
n
d
i
n
gt
h
ea
l
g
o
r
i
t
h
mandi
t
s
p
r
o
g
r
e
s
st
o
w
a
r
dt
h
es
o
l
u
t
i
o
n
.I
np
r
a
c
t
i
c
a
ls
i
t
u
a
t
i
o
n
swhichu
s
u
a
l
l
yi
n
v
o
l
v
eo
v
e
r
t
h
e
r
ei
snoo
p
p
o
r
t
u
n
i
t
yf
o
rn
e
a
tg
r
a
p
h
i
c
a
lc
o
r
r
e
l
a
t
i
o
nb
e
c
a
u
s
eoft
h
e
twov
a
r
i
a
b
l
e
s
I
im
i
t
a
t
i
o
n
so
fg
r
a
p
h
i
c
a
lr
e
p
r
e
s
e
n
t
a
t
i
o
n
.I
twi
l
1ben
e
c
e
s
s
a
r
yt
or
e
v
i
e
wt
h
enumbers
t
od
e
c
i
d
eonc
o
n
v
e
r
g
e
n
c
eo
rt
h
el
a
c
kt
h
e
r
e
o
f
.Whath
a
p
p
e
n
sw
i
t
hmanyv
a
r
i
a
b
l
e
s
i
sa
ne
x
t
e
n
s
i
o
noft
h
eg
e
o
m
e
t
r
i
cf
e
a
t
u
r
e
st
h
a
ta
r
eo
b
s
e
r
v
e
di
np
r
o
b
l
e
m
si
n
v
o
l
v
i
n
g
t
i
a
lt
o
o
lf
o
rt
h
e
s
ep
r
o
b
l
e
m
sw
i
t
h
twov
a
r
i
a
b
l
e
s
.A goodi
m
a
g
i
n
a
t
i
o
ni
sa
ne
s
sn
manyv
a
r
i
a
b
l
e
s
.
r
e
s
e
n
t
a
t
i
o
no
ft
h
eo
p
t
i
m
i
z
a
t
i
o
n
C
o
n
t
o
u
rp
l
o
t
sp
r
o
v
i
d
et
h
eb
e
s
tg
r
a
p
h
i
c
a
l p
p
r
o
b
l
e
mi
ntwov
a
r
i
a
b
l
e
s
.Thep
o
i
n
t
so
na
n
yc
o
n
t
o
u
r(
o
rc
u
r
v
e
)h
a
v
et
h
esamev
a
l
u
e
o
ft
h
ef
u
n
c
t
i
o
n
.S
e
v
e
r
a
ls
o
w
arep
a
c
k
a
g
e
sa
r
ea
v
a
i
l
a
b
l
et
oc
r
e
a
t
ea
n
dd
i
s
p
l
a
yt
h
e
s
e
p
l
o
t
s
.M
A
T
L
A
B
.M
a
t
h
e
m
a
t
i
c

a MapleandMathcada
r
eamongeg
e
n
e
r
a
l
p
u
q
J
O
s
e
a
ndrawc
o
n
t
o
u
rp
l
o
t
s
.I
nt
h
i
sbook
wew
i
l
lu
s
eMATLABf
o
r
s
o
f
t
w
a
r
ep
a
c
k
a
g
e
s c
1o
p
t
i
m
i
z
a
t
i
o
n
.
g
r
a
p
h
i
ca

2
.
1 PROBLEMDEFINITION

Thes
t
a
n
d
a
r
df
o
r
m
a
tf
o
ro
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
e
m
swase
s
t
a
b
l
i
s
h
e
di
nC
h
a
p
t
e
r1
.l
ti
s
d
u
c
h
e
r
ef
o
rc
o
n
v
e
n
i
e
n
c
e
:
r
e
i
no

2
.
1P
A
O
B
L
E
MD
E
F
I
Nf
T
lO
N 4
7

4
6 G
R
A
P
H
I
C
A
LO
P
T
I
M
'
Z
A
T
I
O
N

f
l
.
x

JX
2

.

Xn)

M
i
n
i
m
i
z
e

I
I
u
s
t
r
a
t
e
st
h
eg
r
a
p
h
i
c
a
ls
o
l
u
t
i
o
nt
oi
s
p
r
o
b
l
e
m
.l
ef
i
g
u
r
ea
l
s
od
i
s
p
l
a
y
s
F
i
g
u
r
e2
.
1i
e
l
a
t
e
dt
ot
h
ei
n
e
q
u
a
l
i
t
yc
o
n
s
a
i
n
t
sp
l
a
c
e
do
ni
ta
f
t
e
rt
h
e
a
d
d
i
t
i
o
n
a
li
n
f
o
r
m
a
t
i
o
nr
p
r
i
n
c
i
p
a
lg
r
a
p
h
i
ci
n
f
o
r
m
a
t
i
o
nwasg
e
n
e
r
a
t
e
da
n
dd
i
s
p
l
a
y
e
d
.

(
2.
1
)

h2
(
x

X
2

.
.
x
)=O

S
u
b
j
e
c
tt
o
: h(
x
"X
2

Xn)=0
h
t
(
X
I
X
2

.
..X

)=0

2
.
1
.
2 Formatf
o
rt
h
eG
r
a
p
h
l
c
a
lD
l
s
p
l
a
y

(
2
.
2
)

Th
eg
r
a
p
h
i
c
a
ls
o
l
u
t
i
o
nt
oE
x
a
m
p
l
e2
.
1
a
ss
e
e
ni
nF
i
g
u
r
e2
.
1
i
sg
e
n
e
r
a
t
e
du
s
i
n
g
MA
T
I
.
A
B[
1.
2
]e
x
:
p
tf
o
rt
h
ei
d
e
n
t
i
f
i
t
i
o
no
ft
h
ei
n
e
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
t
s
.A
l
li
n
e
q
u
a
l
i
t
y
c
o
n
s
t
r
a
i
n
t
sa
r
ed
i
s
t
i
n
g
u
i
s
h
e
db
yh
a
s
hm
a
r
k
s
.Theh
a
s
h
e
ds
i
d
ei
n
d
i
c
a
t
e
st
h
ei
n
f
e
a
s
i
b
l
e
r
e
g
i
o
n
.I
nM
A
'
1.AB5
.
2e
r
ei
sn
of
e
a
t
u
r
et
oi
n
s
e
r
tt
h
e
s
em
a
r
k
st
h
r
o
u
g
hp
r
o
g
r
a
m
m
i
n
g
.
Ana
d
d
o
np
r
o
g
r
a
mc
a
l
l
e
dMatdraww
a
su
s
e
dt
oc
r
e
a
t
et
h
eh
a
s
hi
I
n
e
si
nt
h
ef
i
g
u
r
e
.I
t
c
a
nb
ed
o
w
n
l
o
a
d
e
df
r
o
mt
h
eM
a
t
h
w
o
r
k
ss
i
t
e
.B
e
g
i
n
n
i
n
gw
i
V
e
r
s
i
o
n5
.
3t
h
e
r
ei
sa
p
l
o
t
e
d
i
t
)d
i
t
o
r w
i
l
la
l
l
o
wy
o
ut
oi
n
s
e
r
ta
d
d
i
t
i
o
n
a
lg
r
a
p
h
i
ce
l
e
m
e
n
t
st
oE
f
i
g
u
r
e(
e
. l
es
t
u
d
e
n
tc
a
na
l
s
op
c
i
li
neh
a
s
ht
i
n
e
sa
f
t
e
ro
b
t
a
i
n
i
n
gt
h
ep
r
i
n
t
o
u
tI
ti
s
f
i
g
en
o
t
e
da
tt
h
eg
r
a
p
h
i
c
a
ls
o
l
u
t
i
o
ni
si
n
c
o
m
p
l
e
t
ei
ft
h
ei
n
e
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
t
sa
r
e
t
ob

o
ri
ft
h
ef
e
a
s
i
b
l
er
e
g
i
o
nh
a
sn
o
tb
e
e
ne
s
t
a
b
l
i
s
h
e
di
nsomem
a
n
n
e
r
.
n
o
td
i
s
t
i
n
g
u
i
s
h
e
d

gl(X"X2

.
.
.Xn)o
g
2
(
X
I
.
X
2

.
.
Xn)SO
g
l
l
l
(
X
"
X
2

.
..x
)SO

.
.
.
n
2
x
:
g
x
iX
7
. i=I

(
2
.
3
)

(
2.
4
)

i
g
h
t
h
a
n
d
e
r
oo
nt
h
er
e
c
e
s
s
i
t
yf
o
rz
h
en
o
r
m
a
t
.t
ot
h
ef
d
h
e
r
i
n
gt
w
h
i
l
ea

h
a
p
t
e
r
I
ni
sc

c
o
m
p
r
e
h
e
n
s
i
o
n
.
l
e
a
i
d
a
n
d
1
0
c
o
n
v
e
n
i
e
n
c
e
f
o
r
o
n
e
si
sb
e
i
n
gd
e
l
a
x
e
d
. i
s
i
d
ei
sr
e
x
a
m
p
l
e

f
i
r
s
t
T
h
z
e
r
o
.
o
t
h
e
r
1
h
a
n
v
a
l
u
e
s
n
u
m
e
r
i
c
a
l
h
a
v
e
a
l
s
o
c
a
n
i
d
e
r
i
g
h
t
h
a
n
ds
h
o
s
eg
m
p
1
2
i
c
a
l
u
n
c
t
i
o
n
sw
l
e
m
e
n
t
a
r
yf
n
E
u
s
i
n
ge
i
m
p
l
eo
l
l
u
s
t
r
a
t
i
o
ni
sas
o
ri
c
h
o
s
e
nf
h
eMATLAB
x
a
r
n
i
n
n
t
i
o
no
ft
i
l
lp
e
r
m
i
te
x
a
m
p
l
ew
i
m
p
l
ee
n
o
w
n
.T
h
i
ss
e
l
lk
n
a
t
u
r
ei
sw
e
f
i
n
eE
l
l
o
wu
od
st
tw
i
l
la
l
s
oa
o
l
u
t
i
o
n
.I
h
es
u
r
v
e
sa
n
dt
e
n
c
r
a
t
eec
tw
i
l
lg
c
o
d
ea
p
r
o
b
l
c
m
.
g
r
a
p
h
i
c
a
l
o
p
t
i
m
i
z
a
t
i
o
n
t
o
t
h
e
s
o
l
u
t
i
o
n
o
f
i
s
p
1
a
y
o
rt
h
ed
f
o
r
m
a
tf

1
0

.
1
2
.
1
.
1 Example2

n
e
q
u
a
l
i
t
y
n
dt
w
oi
o
n
s
t
r
a
i
n
t
sa
q
u
a
l
i
t
yc
a
v
et
w
oe
w
i
l
lh
.
1
E
x
a
m
p
l
e2
x
a
m
p
l
e

i
r
s
te
Thef
c
o
n
s
t
r
a
m
t
s
:
M
i
n
i
m
i
z
e

)
2
)
2+(
X
2-2
f
l
.XI'X2)=(XI-3

x1
S
u
b
j
c
c
tt
o
: h
2
)
:2
l
(
x
l
.X
+x2=8

(
2
.
5
)
(
2
.
6
a
)

)
2+(
h2
X
I-1
-4)2=4
(x"XU:(

(
2
.
6
b
)

2S7
2
)
:X
l+x
gl(X"X

(
2
.
7
n
)

g2(x.
XU:XI 0.25~SO

(
2
.
7
b
)

oSXIS10;

0
OSX2S1

(
2
.
8
)

o
t
e
a
r
a
b
o
l
a
.N
n
n
dap
i
r
c
l
e
s
w
oc
t
r
a
i
g
h
tl
i
n
e
s
.t
b
o
v
ed
c
f
m
i
1
i
o
n
h
ea
weh
a
v
et
w
os
I
nt

f
o
r

E
s
c
o
p
e
t
i
sn
o
m
p
l
y
a
r
i
a
b
l
e
si
q
u
a
l
i
t
yc
o
n
s i
n
t
sa
n
dt
w
ov
w
oe
t
h
a
tt
t
h
e
p
r
o
v
i
d
e
d
c
o
n
s
t
r
a
i
n
t
s
u
a
l
i
1
y
d
e
t
c
r
m
i
n
e
d
b
y

w
i
l
l
b
e

ep
r
o
b
l
e
m
a
t
i
o
n
. l
oz
se
x
a
m
p
l
e
e
.Thi
x
n
r
r
u
ei
nt
h
i
se

w
h
i
c
hi
st
n
d
e
p
e
n
d
e
n
t
l
i
In
e
a
r
l
yi
n
t
sa
r
ei
t
w
o
c
o
n
s
r
a
p
h
i
c
a
l
eu
s
e
dt
od
r
a
wg
i
l
lb
ABa
tw
o
d
c
i
n
M
A
'
n
d
e
r
s
t
a
n
d c
e
l
pu
0h
e
d1
w
a
s
1
o
o
k
.
h
i
sb
nt
s
o
l
u
t
i
o
n
si

g
u
r
o2
.
1

4 5 6

x
1v
a
l
u

a
m
p
l
e2
.
1
.
x
G
r
a
p
h
i
c
a
ls
o
l
u
t
i
o
nf
o
rE

1
0

48 G
R
A
P
H
I
C
A
LO
P
T
I
M
I
Z
A
T
I
O
N
I
nt
h
eg
r
a
p
h
i
c
a
ld
i
s
p
l
a
yo
fs
o
l
u
t
i
o
n
so
n
l
yeo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
ni
sdrawnf
o
r
i
r
e
c
t
i
o
nf
o
reminimumc
a
nb
ei
d
e
n
t
i
f
i
e
d
.Eachc
q
u
a
l
i
t
y
s
e
v
e
r
a
lc
o
n
t
o
u
r
ss
ot
h
a
ted
c
o
n
s
t
r
n
ti
sdrawna
sas
i
n
g
l
ec
u
r
v
ef
o
ren
u
m
e
r
i
c
a
lv
a
l
u
eont
h
er
i
g
h
t
h
a
n
ds
i
d
e
.
Theys
h
o
u
l
db
ei
d
e
n
t
i
f
i
e
d
.Eachi
n
e
q
u
a
l
i
t
yc
o
n
st
r
a
In
ti
ss
i
m
i
l
a
r
l
ydrawnf
o
rav
a
l
u
eon
er
i
g
h
t
h
a
n
ds
i
d
e
.Theyn
e
e
dt
ob
ei
d
e
n
t
i
f
i
c
dt
o
o
.Theh
a
s
hmarksa
r
l
r
a
w
n
/
p
e
n
c
i
l
e
d
f
i
g
u .Thee
x
t
e
n
to
ft
h
ed
c
s
i
g
nr
e
g
i
o
ni
se
s
t
a
b
l
i
s
h
e
d
i
nont
h
ef
i
g
u
r
co
rap
r
i
n
t
o
u
to
fe
andt
h
es
o
l
u
t
i
o
ni
d
e
n
t
i
f
i
e
dont
h
ef
i
g
u
r
e
.Ther
e
l
e
v
a
n
tr
e
g
i
o
nc
a
nb
ezoomedf
o
rb
e
t
t
e
r
a
t
af
o
r ep
l
o
ta
r
eo
b
t
a
i
n
e
du
s
i
n
gMATLAB'Sm
a
t
r
i
xo
p
e
r
a
t
i
o
n
sa
f
t
e
r
a
c
c
u
r
a
c
y
. d
e
s
t
a
b
l
i
s
h
i
n
gt
h
ep
l
o
l
t
i
n
gm
e
s
h
.Thi
se
x
p
l
o
i
t
sMA
T
I
.
A
B
'
Sn
a
l
u
r
a
ls
p
e
e
df
o
rm
a
t
r
i
x
o
p
e
r
a
t
i
o
n
s
.

2
.
2 GRAPHICALSOLUTION
MATLABp
o
s
s
sap
o
w
e
r
f
u
lv
i
s
u
a
l
i
z
a
t
i
o
ne
n
g
i
n
et
h
a
tp
e
r
m
i
t
st
h
es
o
l
u
t
i
o
no
fe
t
w
o
v
a
r
i
a
b
l
eo
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
e
mb
yi
n
s
p
e
c
t
i
o
n
.T
h
e
r
ea
r
et
h
r
e
ewayst
ol
a
k
e
a
d
v
a
n
t
a
g
eo
ft
h
eg
r
a
p
h
i
cf
e
a
t
u
r
e
so
fMA
TI.AB
.
ef
i
r
s
ti
sl
h
eu
s
eo
fMA
T
I
.
A
B
'
S
r
i
m
a
r
ywayt
o
h
i
g
h
I
e
v
e
lg
r
a
p
h
i
n
gr
o
u
t
i
n
e
sf
o
rd
a
t
av
i
s
u
a
l
i
z
a
t
i
o
n
.Thi
sw
i
l
lb
e p
s
o
l
v
eg
r
a
p
h
i
c
a
lo
p
l
i
m
i
z
a
t
i
o
np
r
o
b
l
e
m
si
nt
h
i
sb
o
o
k
.T
h
i
swi
l
1a
l
s
ob
eeway10
i
n
c
o
r
p
o
r
a
t
eg
r
a
p
h
i
c
a
le
x
p
l
o
r
a
l
i
o
no
fn
u
m
e
r
i
c
a
lt
e
c
h
n
i
q
u
e
si
neb
o
o
k
.F
o
rmo

p
r
e
c
i
s
ec
o
n
t
r
o
lo
v
e
rt
h
ed
i
s
p
l
a
yo
fd
a
t

a MATLABa
l
l
o
w
su
s
ri
n
t
c
r
a
c
t
i
o
nt
h
r
o
u
g
h
programming u
s
i
n
ga
no
b
j
e
c
t
o
r
i
e
n
t
e
ds
y
s
t
e
mi
d
e
n
t
i
f
i
e
di
n MA
T
I
.
A
Ba
sHandle
et
h
i
r
du
s
eo
f
t
h
eMATLABg
r
a
p
h
i
c
se
n
g
i
n
ei
st
ou
s
et
h
eH
a
n
d
l
eG
r
a
p
h
i
c
s
G
r
a
p
h
i
c
s
.Th
s
y
s
t
e
mt
od
e
v
e
l
o
paG
r
a
p
h
i
c
a
lU
s
e
rI
n
t
e
r
f
a
c
e(
G
U
I
)f
o
rt
h
ep
r
o
g
r
a
mo
rm
f
i
l
e
.Th
i
s
c
r
e
a
l
e
saf
a
c
i
l
i
t
yf
o
rs
o
p
h
i
s
l
i
c
a
t
e
du
s
e
ri
n
t
e
r
a
c
t
i
o
n
.MOSlo
ft
h
ep
l
o
t
t
i
n
gn
e
e
d
sc
a
nb
e
melbyt
h
eh
i
g
h
I
e
v
e
lg
r
a
p
h
i
c
sf
u
n
c
t
i
o
n
sa
v
a
i
l
a
b
l
ei
nM
A
T
L
A
B
.
2
.
2
.
1 M
A
T
L
A
BHigh-LevelGraphlcsFunctlons

:
.
L

m
f
i
l
ef
o
ref
i
r
s
te
x
a
m
p
l
e
.Thef
o
l
l
o
w
i
n
gi
n
t
r
o
d
u
c
e
ssomemoreu
s
e
f
u
lI
n
f
o
r
m
o
n

8h
i
g
h
I
e
v
c
lg
r
a
p
h
i
c
sf
u
n
c
t
i
o
n
s
.
a
b
o
u
tt
h
eu
s
eofMATLAB
Thetwomaing
r
a
p
h
i
c
a
le
l
e
m
e
n
t
s a
r
ct
y
p
i
c
a
l
l
yc
o
n
t
r
o
l
l
e
du
s
i
n
gt
h
eh
i
g
h
I
e
v
e
l
a
n
dt
h
ea
x
e
s
.U
s
i
n
gH
a
n
d
l
eG
r
a
p
h
i
c
syouc
a
nc
o
n
t
r
o
l
g
r
a
p
h
i
c
sf
u
n
c
t
i
o
n
sa
r
et
h
ef
i
g
u
m
o
s
to
ft
h
eo
t
h
e
rg
r
a
p
h
i
c
a
le
l
e
m
e
n
t
s
w
h
i
c
hi
n
c
l
u
d
eelem
n
t
su
s
e
di
nt
h
eGU
I
.These
x
e
so
rf
i
g
u
r
e
.Thef
i
g
u
r
ef
u
n
c
t
i
o
no
r
o
t
h
e
re
l
e
m
e
n
t
sa
r
et
y
p
i
c
a
l
l
yc
h
i
l
d
r
c
no
fea
commandc
r
e
a
t
e
saf
i
g
u
r
ewindoww
i
t
hanumbers
t
a
r
t
i
n
ga
to
n
e

o
rw
i
l
lc
r
e
a
t
a
new
f
i
g
u
r
e window i
n
c
r
e
m
e
n
t
i
n
gt
h
e window c
o
u
n
t by o
n
e
.N
o
r
m
a
l
l
ya
l
lg
r
a
p
h
i
c
s
f
u
n
c
t
i
o
n
sa
r
et
a
r
g
e
t
e
dt
ot
h
ec
r
e
n
tf
i
g
u
r
w
indowwhichi
ss
e
l
e
c
t
e
dbyc
l
i
c
k
i
n
gi
t
w
i
t
ht
h
emouseo
re
x
e
c
u
t
i
n
gt
h
ecommandf
i
g
u
r
e(
n
u
m
b
e
r
)
w
h
e
r
enumberi
st
h
e
numbero
ft
h
ef
i
g
u
r
ewindowa
lw
i
l
lh
a
v
el
h
ef
o
c
u
s
.AI
lcommandsa
r
ei
s
s
u
e
di
n
commandwindow.G
r
a
p
h
i
c
sf
u
n
c
t
i
o
n
so
rcommandsw
i
l
la
u
t
o
m
a
t
i
c
a
l
l
yc
r
e
a
t
ea
windowi
fn
o
n
ee
x
i
s
t
s
.
I
nvokeh
e
l
pcommandnamea
tt
h
eMATLABpromptt
oknow morea
b
o
u
tt
h
e
f
u
n
c
t
i
o
n
so
rc
o
m
r
n
a
n
d
s
.Thewordh
a
n
d
l
ea
p
p
o
nm yp
l
a
o
rmsandi
nmany
a
p
p
l
i
c
a
t
i
o
n
s
p
a
r
t
i
c
u
l
a
r
l
yt
h
o
s
et
h
a
td
e
a
lw
i
t
hg
r
a
p
h
i
c
s
.Theya
r
ew
i
d
e
l
yp
r
e
v
a
l
e
n
ti
n
o
b
j
e
c
t

o
r
i
e
n
t
e
dprogramming p
r
a
c
t
i
c
e
. MA
T
I
.
AB

sv
i
s
u
a
l
i
z
a
t
i
o
ns
y
s
t
e
mi
so
b
j
e
c
t
a
n
d
l
e
"i
nMA
T
I
.
AB
o
r
i
e
n
t
e
d
.Mostg
r
a
p
h
i
c
a
le
l
e
m
e
n
t
sa
r
ec
o
n
s
i
d
e
r
e
da
so
b
j
e
c
t
s
.Theh
i
sas
y
o
f
t
w
a
r
ec
r
e
a
t
e
dnumber c
a
ni
d
e
n
t
'
yt
h
es
p
e
c
i
f
i
cg
r
a
p
h
i
co
b
j
e
c
.
tI
f
t
h
i
sh
a
n
d
l
ei
sa
v
a
i
l
a
b
l
e
e
np
r
o
p
e
r
t
i
e
so
ft
h
eo
b
j
e
c
ls
u
c
ha
sl
i
n
es
i
z
e
m
a
r
k
e
rt
y
p
e
c
o
l
o
r

a
n
ds
oonc
a
nb
ev
i
e
w
c
d
s
e
t

o
rr
e
s
e
ti
fn
e
c
c
s
s
a
r
y
.I
nMATLAB
H
a
n
d
l
eG
r
a
p
h
i
c
s
oc
u
s
t
o
r
n
i
z
eeg
r
a
p
h
i
c
a
le
l
e
m
e
n
t
s
.I
nh
i
g
h
I
e
v
e
lg
r
h
i
c
sf
u
n
c
t
i
o
n
s s
i
sewayt
i
su
s
e
di
nal
i
m
i
t
e
dw
a
y
.I
nt
h
i
sc
h
a
p
t
e
r

wew
i
l
lu
s
ei
ti
nam
i
n
i
m
a
lwayt
oc
h
a
n
g
eE
c
h
a
r
a
c
t
c
r
i
s
t
i
c
so
fsomeo
ft
h
eg
r
a
p
h
i
c
a
le
l
e
m
e
n
t
sont
h
ef
i
g
u
r
e
.
To u
n
d
e
r
s
t
a
n
dt
h
ec
o
n
c
e
p
to
f"
h
a
n
d
l
c
s
" we w
i
l
lr
u
n e f
o
l
l
o
w
i
n
gc
o
d
e
i
n
t
e
r
a
c
t
i
v
e
l
y
.I
td
e
a
l
sw
i
t
hc
r
e
a
t
i
n
gap
l
o
t
whichwasi
n
t
r
o
d
u
c
e
di
nC
h
a
p
t
e
r1
.Th
e
b
o
l
d
f
a
c
ecommentsi
nt
h
ec
o
d
es
e
g
m
e
n
ti
n
d
i
c
a
t
et
h
enewf
e
a
l
u
r
e
st
h
a
ta
r
eb
e
i
n
g
e
m
p
h
a
s
i
z
e
d
.
S
t
a
M
ATLABandi
n
t
e
r
a
c
t
i
v
e
L
yp
e
r
f
o
r
mt
h
ef
o
l
l
o
w
i
n
g
:
0

rh

tr

>>X

con
et-vc
ex
xv
s
eyv
L
aey
e
rat
ceo
rl
cp

r
o

c
e

O:pi/40:2*pi;
X. sin(
x
);
y
'b-');
>> plot(x
>> grid;
>>h plot(x
y
'b-') h is the handle to the plot
> > anewp
lot is overwritten in the same figurewindow
>> an
umerical value is assigned to h
> > We c
an use the variable name h or its value
>> t
o refer to the plot again
>>
'LineWidth'
2); this should make your plot
>> set(h
>
>
t
hickerblue
3
'LineStyle'
':
'
'Color''r')
>> set(h'LineWidth'
>>y

h
e
r
ea
r
et
h
r
e
eu
s
e
f
u
lwindowsd
u
r
i
n
gat
y
p
i
c
a
lMATLABs
c
s
s
i
o
n
.Th
ef
i
r
s
ti
se
'
MATLABCommandwindowt
h
r
o
u
g
hwhichMATLABr
e
c
e
i
v
e
si
n
s
t
r
u
c
t
i
o
nandd
i
s
p
l
a
y
s
a
l
p
h
a
n
u
m
e
r
i
ci
n
f
o
r
m
a
t
i
o
n
.l
es
e
c
o
n
dwindowi
st
h
et
e
x
l
e
d
i
t
o
rwindoww
h
e
r
e
i
Ie
sa
r
ec
o
d
e
d
.Thet
h
i
r
di
st
h
eF
i
g
u
r
ewindoww
h
e
r
et
h
eg
r
a
p
h
i
ce
l
e
m
e
n
t
sa
r
e
m-f
nonef
i
g
u
r
ew
i
n
d
o
w
.Thef
i
g
u
r
ewindowi
s t
a
r
g
e
l
d
i
s
p
l
a
y
e
d
.T
h
e
r
ec
a
nb
emorea
o
ft
h
eh
i
g
h
l
e
v
e
lg
r
a
p
h
i
c
sf
u
n
c
t
i
o
n
s
.
Theg
r
a
p
h
i
c
sf
u
n
c
t
i
o
n
si
nMATLABa
l
l
o
wyout
op
l
o
ti
n2Do
r3
D
.Th
e
ya
l
l
o
w
mesha
n
ds
uaplots
b
a
r

a
r
e
a
p
i
ec
h
a
r
t
s
h
i
s
t
o
g
r
a
m
s
c
o
n
t
o
u
rp
l
o
t
si
n2Da
n
d3D
a
n
i
m
a
t
i
o
n
a
n
dg
r
a
d
i
e
n
tp
l
o
t
s
.S
u
b
p
l
o
t
sc
a
na
l
s
ob
ed
i
s
p
l
a
y
e
du
s
i
n
gt
h
e
s
ef
u
n
c
t
i
o
n
s
.
I
na
d
d
i
t
i
o
nt
h
e
yp
e
r
m
i
to
p
e
r
a
t
i
o
nw
i
t
hi
m
a
g
e
sa
n
d3Dm
o
d
e
l
i
n
g
.Theya
l
l
o
wb
a
s
i
c
c
eo
ft
h
ep
l
o
tt
h
r
o
u
g
hc
o
l
o
r
l
i
n
es
t
y
l
e
andm
a
r
k
e
r
si
s
c
o
n
t
r
o
lo
ft
h
ca
p
p
e
a
r
and p
e
c
tr
a
t
i
oo
ft
h
eg
r
a
p
h
.
'
h
e
yp
e
r
m
i
ta
n
n
o
t
a
t
i
o
no
ft
h
eg
r
a
p
hi
ns
e
v
e
r
a
l
r
a
n
g
e
s
w
a
y
s
.Someo
ft
h
e
s
ef
u
n
c
t
i
o
n
sw
i
l
lb
eu
s
e
di
nt
h
en
e
x
ts
e
c
t
i
o
nwhenwed
e
v
c
l
o
pt
h
e

2
.
2G
R
A
P
H
I
C
A
LS
O
L
U
T
I
O
N 49

50

2
.
2 GAAPHICALSOLUTION

GAAPHICALOPTIMIZATION

>>
>>
>>
>>
>>

The handle is used to refer to the object


whoseproperty is being changed
Usually Property information occurs in pairs of
property-namejproperty-value
property-value can be a text string or number

F
i
l
e
:Ex21.m

>>

>>
>>

>>
'ytick'[
5
2.5
0
2.5
5
])
>> set(gca
>>

youhave reset the ytick marks on the graph

>> set(gca
'FontName'
'
A
rial''FontWeight'.
.
.
>>
'bold'
'FontSize'
14)
>>

Changes the font used for marking the axes


set(gca
'Xcolor'
'blue')
>> c
hanges the x-axis to blue
>> % concludes the demonstration of handles

(
x
13
)**2 + (x22)**2
2x1 + x2 8
(x1-1)^2 + (X2-4)^2 4
x1 + x2 <= 7
x1 - 0.25x2^2 <= 0.0

0 <= x1 <= 10

0 <= x2 <= 10

NOTE: The hash marks for the inequality


constraints must be determined and drawn outside
of this exercise and on the printout

----- ""--

x1=0:0.1:10i
%

the semicolon at the end prevents

the echo
t
hese are also the side constraints
x1 and x2 are vectors filled with numbers starting
at 0 and ending at 10.0 with values at intervals
of 0.1

x2=0:0.1:10i

E
x
2
_
1
.
m
.

Minimize f(x1
x2)
h1(x1
x2)
h2(x1
x2)
gl(X1
x2)
gl(x1
x2)

F
i
g
u
r
e2
.
1i
st
h
eg
r
a
p
h
i
c
a
lr
e
p
r
e
s
e
n
t
a
t
i
o
no
fExample2
.
1
.Ther
a
n
g
ef
o
rt
h
ep
l
o
t
s
m
a
t
c
h
e
st
h
es
i
d
ec
o
n
s
t
r
a
i
n
t
sf
o
rt
h
ep
r
o
b
l
e
m
.Thei
n
t
e
r
s
e
c
t
i
o
no
ft
h
etwoe
q
u
a
l
i
t
y
o
l
u
t
i
o
n
st
ot
h
ep
r
o
b
l
e
ma
p
p
r
o
x
i
m
a
t
e
l
y(
2
.
6

c
o
n
s
t
r
a
i
n
t
si
d
e
n
t
i
f
i
e
stwop
o
s
s
i
b
l s
2
.
8
)a
n
d(
16
)
. The i
n
e
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
tg
2 makcs t
h
ep
o
i
n
t(
2
.
62
.
8
)
1
6
)i
sa
c
cp
t
a
b
l
ew
i
t
hr
e
s
p
e
c
tt
ob
o
t
hc
o
n
s
t
r
a
i
n
t
s
.Thc
u
n
a
c
c
e
p
t
a
b
l
e
.P
o
i
n
t(
s
o
l
u
t
i
o
ni
st
h
e
r
e
f
o
r
ea
t(
1
6
)
.Whilet
h
eo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
nwasn
o
tu
s
e
dt
o
c
o
n
t
o
u
r
so
ft
h
eo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
na
r
edrawni
n
d
i
c
a
t
i
n
g
d
e
t
e
r
m
i
n
et
h
cs
o
l
u
t
i
o
n
d
i
r
e
c
t
i
o
nf
o
reminimumo
ft
h
eo
b
j
e
c
t
iv
f
u
n
c
t
i
o
n
.Thes
o
l
u
t
i
o
nc
a
nb
ei
d
e
n
t
i
f
i
e
d
b
yi
n
s
p
e
c
t
i
o
no
ft
h
ea
s
smbledp
l
o
t
sw
i
t
h
o
u
tr
e
f
e
r
e
n
c
et
ot
h
et
e
r
m
i
n
o
l
o
g
yo
rt
h
e
t
e
c
h
n
i
q
u
c
so
fo
p
t
i
m
i
z
a
t
i
o
n
.R
e
c
a
l
l
h
o
w
e
v
e
r
.t
h
a
tt
h
i
si
so
n
l
yp
o
s
s
i
b
l
ew
i
t
honeo
r
twod
e
s
i
g
nv
a
r
i
a
b
l
e
si
nt
h
ep
r
o
b
l
e
m
.
Th
ec
o
d
ef
o
ri
se
x n
p
l
ei
n
v
o
l
v
e
ss
i
xm
f
i
l
e
s
:as
c
r
i
p
tm
f
i
l
e w
i
l
li
n
c
ude
t
h
eMATLABs
t
a
t
e
m
e
n
t
st
oc
r
e
a
t
eandd
i
s
p
l
a
y ed
a
t
a

a
n
df
i
v
emf
i
1

onef
o
r
c
a
c
hf
u
n
c
t
i
o
ni
n
v
o
l
v
e
di
n Example2
.1
. Web
e
g
i
nw
i
t
ht
h
es
c
r
i
p
tm-file-

the following script should allow the display


of graphical solution

>>

2
.
2
.
2 Example2
.
1-G
r
a
p
h
i
c
a
lS
o
l
u
t
i
o
n

Example 2_1 (Sec 2.1-2.2)

g
raphicalsolutionusingMATLAB (twodesignvariables)

>>

of
i
n
e
t
u
n
et
h
ep
l
o
tyouc
r
e
a
t
eyou
Fromt
h
es
e
s
s
i
o
na
b
o
v
ei
ti
se
v
i
d
e
n
t t
b
j
e
c
t
n
e
e
dt
of
i
r
s
ti
d
e
n
t
i
f
yt
h
eo
b
j
e
c
th
a
n
d
l
e
.Youu
s
et
h
eh
a
n
d
l
et
oa
c
c
e
s
seo
y
.Youc
h
a
n
g
eap
r
o
p
e
r
t
ybys
e
t
t
i
n
gi
t
sv
a
l
u
e
.Wew
i
l
ls
e
emoreg
r
a
p
h
i
c
s
p
r
o
p
f
u
n
c
t
i
o
na
sweo
b
t
a
i
nt
h
eg
r
a
p
h
i
c
a
ls
o
l
u
t
i
o
nt
oExample2
.
1i
n f
o
l
l
o
w
i
n
g
s
e
c
t
l
o
n
.

Chapter 2
: Optimization with MATLAB

%D
r. P
. Venkataraman

>>
>> g
et(gca)

t
his will list the property of the axes
of the curren
p
lot. Note there are a significant
amount of properties you can change to customize
the appearance of the plot

5
1

x2)i
[
X
1 X2] meshgrid(x1
generates matrices X1 and X2 corresponding to
% vectors x1 and x2. It is a mesh of x1 and x2 values
a
t which the functions will be evaluated
f1 obj_exl(X1
X2)i the objective function is
e
valuatedover the entire mesh and stored in f1
M
ATLAB will compute the values for the objective
t
hrougha functionm file called obj_exl.m
ineq1 ineconl(Xl
X2)i the inequality gl is
e
valuatedover the mesh
ineq2 inecon2(Xl
X2)i the inequality g2 is
evaluatedo
ver the mesh
eq1 = eqcon1(Xl
X2)i
t
heequalityh1 isevaluated
o
ver the mesh
eq2= eqco
n2(
X
1
X2)i
t
heequalityh2 is evaluated
over t
he mesh

52

GAAPHICALOPTIMIZATION

[
C
1
han1]
contour(1
x2
ineq1[
7
7
)
'r-')i
a single contour plot of gl (or ineql) is drawn for
t
hevalue of 7 in red color as a continuous line
duplicationof the contour value is necessary
ford
rawing a single contour
han1 i
s the handle to this plot. This handle
c
anbeusedtochange theplotdisplaycharacteristics
Cl c
ontains the value of the contour

han1)i
clabel(C1
labels the contour with the values in vector C1
c
ontour(x1
x2
ineq1[7
7
]
'r-') will draw the contour
w
ithout labeling the value
hold on allowsmultiple plots in the same figure

window
gl')i
gtext(
will place the string '
gl' on the plot at the spot
s
electedby a mouse click. This is a text label
t
his procedure is repeated for remaining
c
onstraints
[
C
2
han2]
contour(x1
x2
ineq2[
0
0
]
'r
)i
clabel(C2
han2)i
gtext('g2')i
[
C
3
han3]
contour(
x
1
x2
eq1[8
8
]
'b-')i
clabel(C3
han3)i
gtext(
h1')i
[
C
4
han4]
contour(x1
x2
eq2[
4
4
]
'b--')i
clabel(C4
han4)i
gtext(
h2')i
[
C
han]
contour(x1
x2
f1
'g')i
contourof '
f
'i
s drawn in green color. The number
of c
ontours are decided by the default value
clabel(
C
han)i
xlabel('x1 values'
'FontName'
'times'
'FontSize'
12
.
'FontWeigh

'bold')i label for x-axes


'FontName'
'times'
'FontSize'
12
ylabel('x2values'
'FontWeight'
'bold')i
grid
hold off

Thef
u
n
c
t
i
o
nm-f
i
I
e
sa
r
e
:

2
.
2 GAAPHICALSOLUTION

5
3

on1.m

function retval inecon1(X1X2)


retval X1 + X2i
X1X2 are matrices
r
etval is the value being returned after the
c
omputation
S
ince X1 and X2 are matrices retva1 is a1so a
m
atrix
By t
his way the entire information on the mesh
i
s generated by a sing1e call to the funcion
i
necon1
2.m

function retva1 inecon2(X1


X2)
retval Xl - 0.2S*X2.^2i
Note t
he use of ^ operator for element by e1ement
o
peration. That is each e1ement of the X2 matrix
i
s squared. Without the dot the imp1ication is a
m
atrixmultiplication -between matrices whose
i
nnerdimensions must agree. Similar operators are
d
efined for element by element muliplication
a
nddivision
Lm

function retval eqcon1(X1


X2)
retval 2.0*Xl + X2i
2.m

function retval eqcon2(X1


X2)
)+ (
X
2 - 4).*(
X
2 -4
)i
retval
(
X
1 -1
).*(X1 1

U
function retval obj_ex1(X1
X2)
X
1 -3
).*(X1 - 3
) +(X2 - 2). (
X
2 2
)i
retval = (
2
.
2
.
3 D
i
s
p
l
a
y
l
n
gtheGraphics

1o
fl
h
cf
i
l
e
sr
e
q
u
i
r
e
df
o
rg
r
a
p
h
i
c
a
ld
i
s
p
l
a
yo
ft
h
ep
r
o
b
l
e
mh
a
v
eb
e
e
nc
r
c
a
l
c
d
.I
nt
h
e
Al
o
m
p
l
e
l
ep
a
t
hf
o
rt
h
cd
i
r
e
c
t
o
r
ya
t
Commandwindow
l
y
p
eaddpatbf
o
l
l
o
w
e
dbyec
y
p
et
h
cnameo
ft
h
cs
c
r
i
p
tf
i
l
ef
o
rt
h
ee
x
a
m
p
l
e
h
o
l
d
st
h
e
s
ef
i
l
e
s
.Att
h
ep
r
o
m
p
tt
w
i
t
h
o
u
tt
h
e.me
X
l
e
n
s
i
o
n
>> Ex
21

54

2
.
2 GRAPHICALSOLUTION

GRAPHICALOPTIMIZAnON

1
0

Th
ef
i
r
s
tc
o
n
t
o
u
rp
l
o
tw
i
l
la
p
p
e
a
r
.Movet
h
emouseo
v
e
rt
oep
l
o
ta
n
dac
r
o
s
sh
a
i
r
a
p
p
e
a
r
s
.C
l
i
c
k
i
n
gont
h
ep
l
o
twi
1
lp
l
a
c
t
h
es
t
r
i
n
g"
g
l
"a
tt
h
ec
r
o
s
sh
a
i
r
.Thes
e
c
o
n
d
c
o
n
t
o
u
rp
l
o
ts
h
o
u
l
da
p
p
e
a
ra
n
dt
h
e
r
ei
sap
a
u
s
et
op
l
a
c
et
e
x
tl
a
b
e.
lA
f
t
e
rt
h
ef
o
u
r
c
o
n
s
a
i
n
t
seo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
ni
sp
l
o
t
t
e
df
o
rs
e
v
e
r
a
lc
o
n
t
o
u
rv
a
l
u
e
s
.F
i
n
a
l
l
yt
h
ep
l
o
t
s
h
o
u
l
dapp
rinF
i
g
u
r
e2
.
1w
i
t
h
o
u
tt
h
eh
a
s
hm
a
r
k
s(
a
n
dw
i y
o
u
rc
h
o
i
c
eo
fB
l
o
c
a
t
i
o
nf
o
rc
o
n
s
t
r
a
i
n
tl
a
b
e
l
s
)
.Youc
a
ni
n
s
e
r
tt
h
eh
a
s
hm
a
r
k
st
h
r
o
u
g
he
d
i
t
i
n
gt
h
ep
l
o
t
d
i
r
e
c
t
l
y(V
e
r
s
i
o
n5.
3onw
d
s
)
. s
o
l
u
t
i
o
nf
o
rt
h
ep
r
o
b
l
e
mi
sa
t
1
(6
)
wheref
o
u
r
p
l
o
t
si
n
t
e
r
s
e
c
.
tThev
a
l
u
eo
ft
h
eo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
ni
s20 es
o
l
u
t
i
o
n
.
I
ft
h
ep
l
o
ti
sa
c
c
e
p
t
a
b
l
e

youc
a
np
r
i
n
tt
h
ei
n
f
o
r
m
a
l
i
o
nont
h
ef
i
g
u
r
eb
yexpo
n
gi
t
(
u
s
i
n
ga
na
p
p
r
o
p
r
i
a
t
ee
x
t
e
n
s
i
o
n
)1
0af
i
l
ea
n
dl
a
t
e
ri
n
c
o
r
p
o
r
a
t
i
n
gi
ti
na
n
o
t
h
e
r
documen
.
tYouc
a
na
l
s
os
e
n
di
tt
oep
r
i
n
t
e
rf
r
o
mI
h
ep
r
i
n
tcommandont
h
eF
i
l
e
menu.T
y
p
i
n
gh
e
l
pp
r
i
n
ti
nt
h
eCommandwindows
h
o
u
l
di
Is
tas
e
to
fcommandsyou
c
a
nu
s
et
os
a
v
eef
i
l
e
.

55

>> print -depsc2 plot_ex_2_1.eps

1
lc
a
t
et
h
el
e
v
e
l2c
o
l
o
rp
o
s
t
s
c
r
i
p
tf
i
l
ec
a
l
l
e
dplocex_2_l
.e
p
si
nt
h
ew
o
r
k
i
n
g
wi
d
i
r
e
c
t
o
r
y
.Ac
o
m
p
l
e
t
ep
a
t
hnames
h
o
u
l
ds
a
v
et
h
ef
i
l
et
oes
p
e
c
i
f
i
e
dd
i
r
e
c
t
o
r
y
.You
c
a
nt
h
e
ne
x
e
c
u
t
es
t
a
n
d
a
r
do
p
e
r
a
t
i
n
gs
y
s
t
e
mcommands1
0dumpt
h
ef
i
l
c1
0e
p
r
i
n
t
c
r
.Youc
a
na
l
s
os
a
v
et
h
ef
i
g
u
r
e(
e
x
s
i
o
n.
f
i
g
)a
n
dl
a
t
e
ropeni
ti
n
a
p
p
l
i
c
a
b
l
MATLABu
s
i
n
gt
h
ef
i
l
co
p
e
ncommandfrommenub
o
r
t
o
o
lb
a
r
.

wi
l
Ib
ec
u
s
t
o
r
n
i
z
c
du
s
i
n
gt
h
e
I
nt
h
i
ss
e
c
t
i
o
nt
h
cb
a
s
i
cf
i
g
u
r
ea
n
dp
l
o
t
sc
r
e
a
t
c
da
b
o
v
H
a
n
d
l
eG
r
a
p
h
i
c
scommandsf
i
r
s
te
x
p
l
o
d
inS
c
t
i
o
n2
.
2
.1
.A
l
lo
f
t
h
ep
l
o
t
sd
e
s
c
r
i
b
i
n
g
ec
o
n
s
t
r
a
i
n
t
sw
i
l
lh
a
v
eal
i
n
e
w
i
d
t
ho
f2p
o
i
n
t
s
.l
eo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
nw
i
l
lh
a
v
ea
l
i
n
e
w
i
d
t
ho
f1p
o
i
nt
.Thet
e
x
tl
a
b
e
l
swi
1
lbei
nt
h
e"
T
i
m
e
s
"f
o
n
t
.I
twi
l
Ib
ei
nb
o
l
d
f
a
c
e
w
i
t
haf
o
n
to
fs
i
z
e1
4a
n
di
nec
o
l
o
ro
ft
h
ec
o
n
s i
n.
tTh
ex dYt
i
c
kmarksw
i
l
lb
e
c
h
a
n
g
e
dt
oe
s
t
a
b
l
i
s
hab
r
o
a
d
e
rg
r
i
d
.Th
ex dyl
a
b
e
1sw
i
l
li
n
c
l
u
d
eas
u
b
s
c
r
i
p
.
tA
t
w
o
r
o
wt
e
x
twi
1
lbei
n
s
e
r
t
e
da
tt
h
ep
o
i
n
ts
e
l
e
c
t
e
dt
h
r
o
u
g
ht
h
em
o
u
s
c
.Thencwf
i
g
u
r
e
i
sshowni
nF
i
g
u
r
c2
.
2
.
F
i
r
s
tpyt
h
ef
i
l
ei
nep
r
e
v
i
o
u
se
x
e
r
c
i
s
ea
n
dr
e
n
a
m
ei
t
.
l
erenamedf
i
l
ew
i
l
lb
e
u
s
e
dmakethen
e
s
s
a
r
y
c
h
g
e
s
t
oec
o
d
e
.Th
e
s
ec
h
a
n
g
e
si
n
c
1udeb
o
t
hnewc
o
d
e
a
swe
I
la
sr
e
p
l
a
c
e
m
e
n
to
fe
x
i
s
t
i
n
gl
i
n
e
so
fc
o
d
e
.I
nef
oI
l
owingo
n
l
yec
h
a
n
g
e
st
o

c
o
d
ea
r
ei
1
Iu
s
t
r
a
t
e
d
.Th
eo
r
i
g
i
n
a
lc
o
d
ei
sshowni
nt
h
ei
t
a
l
i
cs
t
y
l
ea
n
di
d
e
n
t
i
f
i
e
d
w
i
t
ht
h
r
e
ee
l
l
i
p
s
i
sp
o
i
n
t
s(
..
.
)a
t b
e
g
i
n
n
i
n
go
ft
h
el
i
n
e
.Thec
o
d
et
ob
ei
n
s
e
r
t
e
do
r
r
e
p
l
a
c
e
m
e
n
tc
o
d
ei
sshowni
nnorm
s
t
y
l
.I
fnoreplacemcnti
sa
p
p
a
r
e
n
t

t
h
e
nt
h
eo
l
d
c
o
d
ei
su
s
e
d1
0d
e
f
i
n
et
h
el
o
c
a
t
i
o
n
swhe
e
wd
ei
sa
p
p
e
n
d
e
d
.

contour(xl
x2
ineql
{7
7}
'r-'};

set(han1
'LineWidth'
2);
c
ontour to 2 points

sets 1ine width of the gl

F
lg
u
r
e2.2

2
.
2
.
4 CustomlzlngtheFlgure

{Cl
hanl}

a
l
u
e
s
Xv
1

1
0

C
u
s
t
o
m
i
z
e
d
U f
o
r
E
m
p
l
e2
.
1
.

gt t('gl');

k1 gtext(
gl'); k1 is thehandle tothe text element
set(k1
'FontName'
'Times'
'FontWeight'
'bold'
FontSize'
14
'Color'
'red')

..clabel(C2
han2}i

set(han2
'LineWidth'
2
)
k2 = gtext('g2');
set(k2
'FontName'
'Times'
'FontWeight'
'bold'
FontSize'
14
'Color'
'red')

clabel(C3
han3);

set(han3
'LineWidth'
2
)
k3 ~ gtext('h1');
set(k3
'FontName'
'Times'
'FontWeight'
'bold'
'FontSize'
14
'Color'
'blue')
clabel(C4
han4);

set(han4

LineWidth'
2
)
k4 ~ gtext('h2');

56

2
.
3 ADDlTIONALEXAMPLES

GRAPHICALOPTIMIZATION

'FontName'
'Times'
'FontWeigh 'bo1d'
set(k4

FontSize'
14
'Co1or'
'b1ue')
c
:label(
C

han);
set(han
'LineWidth'
1
)
.
. xlabel(

xl values'

FontName''times'
x1abe1(' x 1 va1ues'
'FontName'
'times'

FontSize'
12
.
FontWeight'
'bo1d')i

y1abe1(' x 2 values'
'FontName'
'times'
'
FontSize'
12'FontWeight'
'bo1d')i
set(gca

xtick'[
0 2 4 6 8 10]) set xticks
'ytick'[
0 2.5 5.0 7.5 10]) set yticks
se(gca
k5 gtext({'Chapter 2
: Examp1e 1
'
'pretty graphical display'})
t
he above gtext describes a string array
string array i
s defined using curly braces
'FontName'
'Times'
'FontSize'
12
set(k5

Fontweight'
'bo1d')

!
(
xI'X2)= +bx~ cco) d c
o
s
(
q
x
1
-c
+d
2)-

a=
1

b=
2
c=
0
.
3
d=
0
.
4

p=
3

q=4
F
i
g
u
r
c
s2.
32
.
7a
r
eg
r
a
p
h
i
c
a
ld
i
s
p
l
a
yo
fs
o
l
u
t
i
o
n
s i
ss
e
c
t
i
o
n
.Thef
i
g
u
r
e
si
n
c
t
e
dt
ob
l
a
c
ka
n
dw
h
i
t
ew
i
l
ln
o
tn
v
e
y ei
m
p
a
c
to
fc
o
l
o
ri
m
a
g
e
s
t
h
ebookb
e
i
n
gr
e
s
youw
i
l
ls
e
eont
h
es
c
r
e
e
n
.Thc
r
ea
r
et
w
of
i
l
e
sa
s
s
o
c
i
a
t
c
dw
i
t
ht
h
ep
l
o
t
s
:t
h
es
c
r
i
p
t
m-f
i
I
e(
e
x
2
.
2
.
m
)a
n
dt
h
cf
u
n
c
t
i
o
nm-f
i
I
e(
o
b
L
e
x
2
.
m
)
.
..z.m

Chapter 2
: Optimization with MATLAB
Dr. P.Venkataraman
Examp1e 2.2 Sec.2.3.1

l
o
dC
o
n
l
g

1
1
I
A
b
c

d
e
f
a
u
l
lc
G
l
c
r

Runt
h
ep
r
o
g
r
a
r
nt
os
e
cef
i
g
u
r
ei
nF
i
g 2
.
2
.Oncet
h
ep
l
o
ti
sc
u
s
t
omIzedt
oy
o
u
r
s
a
t
i
s
f
a
c
t
i
o
n
y
o
uc
a
nmakei
t s
t
a
n
d
a
r
df
o
ro
t
h
e
rp
l
o
t
syouw
i
l
lp
r
U inMA
As.
T
h
i
sf
i
l
ec
o
u
l
da
l
s
ob
eat
e
m
p
l
a
t
ef
o
rg
e
n
e
r
a
lc
o
n
t
o
u
rp
l
o
t
t
i
n
g
.N
o
t
ea
ta
1
lf
u
n
c
t
i
o
n
i
I
e
sw
h
i
c
ha
r
ec
o
d
e
do
u
t
s
i
d
et
h
es
c
r
i
p
t
i
n
f
o
r
m
a
t
i
o
ni
so
b
t
a
i
n
e
dt
h
r
o
u
g
hf
u
n
c
t
i
o
nm-f
f
i
l
c
.S
c
t
t
i
n
gupncwp
r
o
b
l
e
m
so
re
x
a
m
p
l
c
so
n
l
yr
e
q
u
i
r
e
sa
d
d
i
n
g
l
c
h
a
n
g
i
n
gnew
f
i
1
.
f
u
n
c
t
i
o
nm-

2
.
3 ADDITIONALEXAMPLES

2
.
3
.
1 Example2
.
2
T
h
i
sc
x
a
m
p
l
ci
I
Iu
s
t
e
ss
c
v
c
r
a
ld
i
f
f
c
r
e
n
twayso
fg
m
p
h
i
c
a
U
yd
i
s
p
l
a
y
i
n
gaf
u
n
c
t
i
o
no
f
r
o
b
l
e
mwasu
s
c
dt
oi
I
I
u
s
l
r
a
t
eg
l
o
b
a
lo
p
t
i
m
i
z
a
t
i
o
ni
nR
e
f
e
r
e
n
c
e3
.
t
w
ov
a
r
i
a
b
l
e
s
.Thep
Th
es
i
n
g
l
co
b
j
c
c
t
i
v
ef
u
n
c
t
i
o
ni
s

(
2
.
9
)

w
i
t
h

(at the end)


c1ear C C1 C2 C3 C4 h h1 h2 h3 h4 k1 k2 k3 k4 k5
g
et rid of variab1es from the workspace

Th
ef
0
1
l
0
w
i
n
ga
d
d
i
t
i
o
n
a
le
x
n
p
l
e
sw
i
l
ls
e
r
v
et
oi
I
l
u
s
t
e
b
o
o
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
e
m
s
w
e
l
la
sa
d
d
i
t
i
o
n
a
lg
m
p
h
i
c
a
lf
c
a
t
u
r
e
so
fM ATLAa
tw
i
l
lb
eu
s
e
f
u
li
nd
v
e
l
o
p
i
n
gg
r
a
p
h
i
c
a
l
s
o
l
u
t
i
o
n
st
oo
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
e
m
s
. l
Cg
r

p
h
i
c
a
lr
o
u
t
i
n
e
si
nMATLABa
r
ep
o
w
e
r
f
u
l d
S
Yt
ou
s
e
.Th
e
yc
a
ng
r

p
h
i
c
a
l
l
yd
i
s
p
l
a
yt
h
ep
r
o
b
l
e
m
s v
e
m
1waysw
i
v
e
r
ys
i
m
p
l
c
mm d
s
.Th
eu
s
e
f
u
ld
i
s
p
l
a
yi
s
h
o
w
e
v
e
r

d
e
t
c
n
n
i
n
e
db
yeu
s
e
r
.Th
ef
i
r
s
te
x
a
m
p
l
ei
n
t
h
i
ss
e
c
t
i
o
nE
x
a
m
p
l
e2
.
2
i
sap
r
o
b
l
e
mi
nu
n
c
o
n
s
i
n
e
do
p
t
t
i
o
n
. s
e
c
o
n
d
c
x
a
m
p
l
ei
sas
t
r
u
c
t
um
1e
n
g
i
n
e
e
r
i
n
gp
r
o
b
l
c
mo
fr
e
a
s
o
n
a
b
l
cc
o
m
p
l
c
x
i
t
y
.l
C e
x
a
r
I
c
s
o
p
t
i
m
i
z
a
t
i
o
ni
nt
h
ca
r
e
ao
fh
e
a
tI
r
a
n
s
f
c
rd
c
s
i
g
n
.
demons

57

0
.
5

g
u

O
Xv
a
l
u

0.
5

2
.
3F
i
l
l
e
dc
o
n
l
o
u
r
sw
i
t
hc
o
l
o
r
b
a
.
rEx
a
m
p
l
e2
.
2
.

58

GRAPHICALOPTIMIZA
T
l
ON

2
.
3 AOOmONALEXAMPLES

5
9

2DCon
w
i
t
hG
r
n
d
i
.
n
lV:10

82-gJ

-0
.
5

g
u
r
e2
.
4

Xv
a
l
u
e
s
1

0
.
5

X
:
!Vu
e
s

C
o
n
t
o
u
rw
l
t
hg
r
a
d
l
e
n
tv
e
c
t
o
r
s
:Exa
m
p
l
e2
.
2
.

graphicalsolutionusingMATLAB (twodesignvariables)
Unconstrained function illustrating global minimum
Example will i
ntroduce 3D plots3D contoursfilled
2D contours with gradient information
--------------

x1=1:0.01:1;

the semi-colon at the end prevents


" the echo
x2=-1:0.01:1i
t
hese are also the side constraints
x1 a
ndx2 are vectors fil1edwith numbers starting
at 1 and ending at 1.0 withvalues at intervals of
0
.01

[
X
l x21 = meshgrid(x1
x2);
generates matrices X1 and X2 corresponding
v
ectors x1 and x2
r
eminder MAT~ is case sensitive

v
a
l
u
e
s

X1

F
l
g
u
r
e
2
.
5 3Dc
o
n
l
o
u
rw
i
l
hl
o
r
b
a
.
rE
x
l
e
2
.
2
.

f1

y
't
.

obj_ex2(X1
X2);

the objective function is


evaluated over the entire mesh
f
illed contour with default colormap
h
elpgraph3d gives you the choices for colormap
han11
contourf(x1
x2
f1
[
C
1
[
0 0.1 0.6 0.8 1.0 1.2 1.5 1.8 2.0 2.4 2.6 2.8 3.01);

s
pecific contour levels indicated above
hanl);
clabel(Cl
colorbar il1ustrates the default color scale
set(gca
'xtick'[
1 -0.5 0.0 0.5 1.01) custom ticks
'ytick'[
1 0.5 0.0 0.5 1.01) custom ticks
set(gca
grid
'FontName'
'times'
xlabel(' X_1 values'
'FontSize'
12);
% label for x-axes
ylabel(' x 2 values'
'FontName'
'times'

60

2
.
3 ADOmONALEXAMPlES

GRAPHICALOPTIMIZATION

6
1

C
o
a

S
u
r
f
a
c
eP
l
o
t
l
o
n
n
n
pj
c
V
d
e
f
n
u
l
t

CO
!
U
'
S
CM
eshP
1
0
1

l
o
r
m
a
p1

3
.
5

3
.
5

2
.
5

2
.
6

1
.
5

1
.
5

1
.
5
a
u
m
w

0
.
5

0
1

0
.
5

0
.
6

Xv
a
l
u

F
l
g
u
r
e2
.
6 3Dm
e
s
hp
l
o
tw
l
t
hc
o
l
o
r
b
a
r
:E
x
a
m
p
l
2
.
2
.

FontSize'
12);
title({'Filled Labelled Contour'
'.
.
.
default color map'}
'FontName'

imes'
'FontSize'
10)

a new figure is used to draw the


basic contour plot superimposed with gradient
information
also information is generatedon a coarser mesh to
keep the figure tidy. grid is removed for clarity

figure a new figure window is drawn


y1 -1:0.1:1.0;
:0.1:1
;
y2 1
Y2]
meshgrid(y1
y2);
[Y1
f2 obj_ex2(Yl
Y2);
han2]
contour(y1
y2
f2
[C2
[
0 0.5 0.75 1.0 1.5 1.75 2.0 2.5 3.0]);

HM

.
0.
5
1

a
v
x

X v
a
l
u
e
s
2

X v
a
l
u
e
s
2

0
.
5

F
l
g
u
r
e2
.
7 3Ds
u
p
l
o
tw
i
t
hd
e
f
a
u
l
tc
o
l
o
r
b
a
:
rE
x
a
m
p
l
e2
.
2
.
clabel(C2
han2)
gradient(f2
0.2);
[GXGY]
% generation of gradient information see help gradient
hold on muliple plos on the same figure
Y2
GX
GY);
quiver(Y1
d
raws the gradient vectors at grid points
see help q
uiver
hold off
'xtick'[-1 -0.5 0.0 0.5 1.0])
set(gca
'ytick'
[-1 -0.5 0.0 0.5 1.0])
set(gca
xlabel(' x_1 values'
'FontName'
'times'

FontSize'
12);
'FontName'
'
imes'
ylabel(' x 2 values'
'FontSize'12);
'with Gradient Vectors'}
title({'2D Contour'

FontName'
'times'
'FontSize'
10)

2
.
3A
O
O
r
r
t
O
N
A
LE
X
A
M
P
L
E
S 6
3

6
2 G
R
A
P
H
I
C
A
LO
P
T
I
M
I
Z
A
T
lO
N
A final example of contour information is a 3D
% contour plot -or a stacked contour plot
figure
colormap(spring)
x2f1
contour3(x1
[
0 0.3 0.6 0.8 1.0 1.5 1.8 2.0 2.4 2.6 2.8 3.0]);
set(gca
'xtick'[
1 0.5 0.0 0.5 1.0])
1 -0.5 0.0 0.5 l.0
]
)
set(gca'ytick'[
c
hange colormap and set colorbar
c
hanging color for display is easy by selecting
a
mong some colormaps providedby MATLAB
colorbar
xlabel(' x 1 values''FontName''times'

FontSize'12);
ylabel(' x 2 values''FontName''times'
.
.
.

FontSize'12);
tile({'Stacked (
3
D
) Contour'
:colormap -spring'}

FontName''times''FontSize'
1
0
)

grid
he next two f
igures w
i
.
l
l display 3D plots
he f
irst is a meshplot of the function

oncemore coarse data is used for clarity


figure
colormap(cool) anothercolormap
y
1
y2
f
2
) using information generated earlier
mesh(
set(gca'xtick'[
1 -0.5 0.0 0.5 l.0
]
)
set(gca'ytick'[
1 0.5 0.0 0.5 1.0])
colorbar
xlabel(' x 1 values''FontName''times'

FontSize'12);
ylabel(' x_2 values''FontName''times'
.
.
.

FontSize'12);
'colormap -cool'}
title({'Coarse Mesh Plot'

FontName''times''FontSize'
1
0
)

grid
h
e final plot in this series

surfaceplot with default colormap


figure
colormap(je}
surf(y1
y2
f
2
) usingold information
colorbar
.
xlabel(' x 1 values''FontName''times'

FontSize'12);
'times'
ylabel(' x_2 values''FontName'

FontSize'12);
set(gca
'xtick'[
1 -0.5 0.0 0.5 l.0
]
)
set(gca
'ytick'
1-1 0
50.0 0.5 l.0
]
)
title({'Coarse Surface Plot'
colormap -jet/
1
0
)
default'}'FontName''times''FontSize'
grid
Q s
.m

function retval obj_ex1(X1


X2)
Optimizationwith MATLAB
D
r. P.Venkataraman
C
hapter2
. Example 2.2

f(x1
x2) a*xl""
2 + b*x2^2 c cos(aa*x1)
d*cos(bb*x2)...
+c+ d
a"
'1
; b 2 c 0.3; d "
' 0.4; aa
3.0*pi; bb 4.0*pi;

n
otematrix operations need a dot operator
retval a*X1.*X1 + b*X2.*X2 -c*cos(aa*X1)
d*cos(bb*X2) + c + d
;

Theb
r
i
e
fc
o
m
m
e
n
t
si
nt
h
ec
o
d
es
h
o
u
l
dp
r
o
v
i
d
ea
ne
x
p
l
a
n
a
t
i
o
no
fw
h
a
ty
o
us
e
eo
n

f
i
g
u
r
e
.F
i
g
u
r
e
sc
a
nbef
u
r
t
h
e
rc
u
s
t
o
r
n
i
z
e
da
ss
e
c
ni
nep
r
e
v
i
o
u
ss
e
c
t
i
o
n
.Froma
n

o
p
t
i
r
n
i
z
a
t
i
o
np
e
r
s
p
e
c
t
i
v
e

F
i
g
u
r
e2.
4p
r
o
v
i
d
e
st
h
cb
e
s
ti
n
f
o
n
n
a
t
i
o
na
b
o
u
tt
h
cn
a
t
u
r
eo
f
o
n
t
o
u
rc
u
r
v
e
si
d
e
n
t
i
f
yt
h
en
e
i
g
h
b
o
r
h
o
o
do
ft
h
e1
0
l
r
n
i
n
i
m
u
m
.
t
h
ep
r
o
b
l
e
m
.The2Dc
l
eg
r
a
d
i
e
n
tv
e
c
t
o
r
si
n
d
i
c
a
t
c d
i
r
e
c
t
i
o
no
fef
u
n
c
t
i
o
n
ss
t
e
e
p
e
s
tr
i
s
ea
1 p
o
i
n
t
s
o sa
n
dv
a
l
l
c
y
sc
a
nb
cd
i
s
t
i
n
g
u
i
s
h
c
d
.Thec
o
n
t
o
u
t
h
e
r
n
s
e
l
v
e
sc
a
nb
ec
o
l
o
r
e
d
w
i
t
h
o
u
tb
e
i
n
gf
i
l
l
c
d
.
Th
cq
u
i
v
e
rp
l
p
1showni
nF
i
g
u
r
e2.
4a
l
s
op
r
o
v
i
d
e
sam
e
c
h
a
n
i
s
m1
0i
n
d
i
c
a
t
ee
f
e
a
s
i
b
l
er
e
g
i
o
nwhend
e
a
l
i
n
gw
i
t
hi
n
e
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
t
ss
i
n
c
ct
h
e
yi
n
d
i
c
a
t
ce
d
i
r
e
c
t
i
o
ni
nw
h
i
c
h c
o
n
s
t
r
a
i
n
tf
u
n
c
t
i
o
nwi
I
Ii
n
c
r
s
e
.I
fs
e
v
e
r
a
lf
u
n
c
t
i
o
n
sa
r
eb
c
i
n
g
wn
e
nt
h
cc
Iu
t
t
e
rp
r
o
d
u
db
yea
r
r
o
w
smaydf
u
s
eec
Ia
r
i
t
y
.Th
eu
s
e
ri
s
e
n
c
o
u
r
a
g
c
d1
0u
s
et
h
e
.
p
o
w
e
r
f
u
lg
r
a
p
h
i
c
a
lf
e
a
t
u
r
e
sofMATLABt
oh
i
sb
e
n
e
f
i
ta
ta
l
lt
i
m
e
s
w
i
t
h
o
u
tl
o
s
i
n
gs
i
g
h
to
feo
b
j
e
c
t
i
v
eo
fh
i
sc
f
f
o
r
t
.MATLABg
r
a
p
h
i
c
sh
a
smanym
o
r
e
f
e
a
t
u
r
e
st
h
a
nw
i
l
lb
ec
o
v
e
r
e
di
ni
sc
h
a
p
t
c
r
.Thee
x
p
o
s
i
nt
h
i
sc
h
a
p
t
e
rs
h
o
u
l
db
c
s
u
f
f
i
c
i
e
n
tf
o
rEd
e
r

c
o
n
f
i
d
e
n
t
l
ye
x
p
l
o
r
emanyo
t
h
e
ru
s
e
f
u
lg
r
a
p
h
i
m
m
a
n
d
s
.
Thc3Dmesha
n
ds
u
r
f
a
c
ep
l
o
t
sh
a
v
el
i
m
i
t
e
du
s
e
f
u
l
n
e
s
s
.Th
e
s
ep
l
o
t
sc
a
nb
eu
s
c
dt
o
r
e
i
n
f
o
r
c
esomeo
ft
h
ef
e
a
t
u
r
e
sf
o
u
n
di
nF
i
g
u
r
e2
.
4
.Thei
n
f
o
n
n
a
t
i
o
ni
nt
h
e
s
ep
l
o
t
smay
lZCSme
a
s
p
e
c
to
ft
h
eg
r
a
p
h
i
I
b
ei
m
p
r
o
v
e
db
yc
h
o
o
s
i
n
gac
a
m
e
r
aa
n
g
l
e emph
d
e
s
c
r
i
p
t
i
o
n
.Thi
se
x
p
l
o
r
a
t
i
o
ni
sl
e
f
tt
oC a
d
e
ra
sa
ne
x
e
r
c
i
s
e
.U
s
i
n
gh
e
l
pv
i
e
wi
nt
h
e
MATLABCommandwindows
h
o
u
l
dg
e
ty
o
us
t
a
r
t
c
di
ni
sd
i
c
t
i
o
n
.

64 GRAPHICALO
P
T
I
M
I
Z
A
T
I
O
N

2
.
3A
D
D
I
T
I
O
N
A
LEXAMPLES

2
.
3
.
2 Example2
.
3
l
en
e
x
te
x
a
m
p
l
ei
sacomplexonefroms
t
r
u
c
t
u
r
a
le
n
g
i
n
e
e
r
i
n
gd
e
s
i
g
n i
sr
e
l
e
v
a
n
t
r
i
n
ga
p
p
l
i
c
a
t
i
o
n
s
.I
ta
p
p
e
a
r
e
da
sap
r
o
b
l
c
mi
n
i
nc
i
v
i
l
l
m
e
c
h
a
n
i
c
a
l
l
a
e
r
o
s
p
a
c
ec
n
g
i
n
R
e
f
e
r
c
n
c
e4
.I
ti
sd
c
v
e
l
o
p
e
di
nd
e
t
a
i
lh
e
r
e
.Thep
r
o
b
l
e
mi
st
or
e
d
e
s
i
g
nt
h
eb
a
s
i
ct
a
l
l
f
l
a
g
p
o
l
ci
nv
i
e
wo
ft
h
ep
h
e
n
o
m
e
n
a
li
n
c
r
e
a
s
ci
nwinds
p
c
e
d
sd
u
r
i
n
ge
x
t
r
e
m
ew
e
a
t
h
e
r
c
o
n
d
i
t
i
o
n
s
.I
nr
e
c
e
n
tc
a
t
a
s
'O
p
h
i
ce
v
e
n
t
s
.t
h
ewinds
p
d
si
nt
o
m
a
d
o
c
shaveb
c
c
n
m
e
a
s
u
r
e
da
to
v
e
r350m
i
l
e
sp
e
rh
o
u
r
.Th
e
s
eh
i
g
hs
p
e
c
d
sapp
r
t
ob
et
h
enormr
a
t
h
e
r
u
n
u
s
u
a
l
c
v
e
n.
t

DeslgnProblem: M
i
n
i
m
i
z
et
h
emasso
fas
t
a
n
d
a
r
dI
O
-m t
u
b
u
l
a
rf
l
a
g
p
o
l
ct
o
l
a
g
p
o
l
ewi

lbemadeofs
t
r
u
c
t
u
r
a
l
w
i
t
h
s
t
a
n
dwindg
u
s
t
so
f350m
i
l
e
sp
e
rh
o
u
r
.Thef
s
t
e
e
.
lUseaf
a
c
t
o
ro
fs
a
f
e
t
yo
f2
.
5f
o
rcs
t
r
u
c
t
u
r
a
Jd
e
s
i
g
n
. l
Cd
e
f
l
e
c
t
i
o
no
ft
h
et
o
p
o
ft
h
ef
l
a
g
p
o
l
es
h
o
u
l
dn
o
te
x
c
e
e
d5cm.Thep
r
o
b
l
e
mi
sd
e
s
c
r
i
b
e
di
nF
i
g
u
r
e2
.
8
.
MathematicalModel: The m
a
t
h
e
m
a
t
i
c
a
l modcl i
sd
e
v
e
l
o
p
e
di
nd
e
t
a
i
lf
o
r
c
o
m
p
l
e
t
e
n
c
s
sandt
op
r
o
v
i
d
ear
c
v
i
e
wo
fu
s
e
f
u
ls
c
t
u
r
a
l(
5
)a
n
da
e
r
o
d
y
n
a
m
i
c
r
e
l
a
t
i
o
n
s[
6
)
.Ther
e
l
a
t
i
o
n
sa
r
ee
x
p
r
c
s
s
e
di
no
r
i
g
i
n
a
ls
y
m
b
o
l
sr
a
t
h
e
rt
h
a
ni
ns
t
a
n
d
a
r
d
f
o
r
m
a
to
fo
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
e
m
st
op
r
o
v
i
d
ea
ni
n
s
i
g
h
ti
n
t
op
r
o
b
l
e
mf
o
r
m
u
l
a
t
i
o
n
.

1owingm
a
t
e
r
i
a
lc
o
n
s
t
a
n
t
s
:
DeslgnParameters:Thcs
t
n
l
c
t
u
c
a
ls
t
c
e
l[
5
]h
a
st
h
ef
ol

65

3
'
Y(
m
a
t
e
r
i
a
Jd
e
n
s
i
t
y
)
:7860kglm
FS(
f
a
c
t
o
ro
fs
a
f
e
t
y
)
:2
.
5
g(
g
r
a
v
i
t
a
t
i
o
n
a
la
c
c
e
l
e
r
a
t
i
o
n
)=9
.
8
1m
l
s
2

F
o
rt
h
ea
c
r
o
d
y
n
a
m
i
cc
a
l
c
u
l
a
t
i
o
n
st
h
ef
o
l
l
o
w
i
n
ga
r
ec
o
n
s
i
d
e
r
e
d
:
3
p(
s
t
a
n
d
a
r
da
i
rd
e
n
s
i
t
y
)
:1
.
2
2
5kglm
Cd( c
o
e
f
f
i
c
i
e
n
to
fc
y
l
i
n
d
e
r
)
:1
.
0
Wp(
f
l
a
gwindl
o
a
da
t8m
)
:5000N
Vw(
w
i
n
ds
p
e
e
d
)
:350mph(
156
.
4
6m
l
s
)

Thegeome
cp
a
r
a
m
e
t
e
r
sa
r

L
p
:thelocationofflagwindload(
8m)
L
:l
e
n
g
o
f
t
h
ep
o
l
(10m)
Ol
I
l
:
lp
e
r
m
i
U
e
dd
e
f
l
e
c
t
i
o
n(
5cm)
b
l
e
s
:Thed
e
s
i
g
nv
a
r
i
a
b
l
e
sshowni
nF
i
g2
.
8a
r
e
DeslgnV
a
l
i

do:o
u
t
s
i
d
ed
i
a
m
e
t
e
r(
X
I
)[N
o
t
e
:x
'
sa
r
en
o
tu
s
e
di
nem
o
d
e
l
]
dj:i
n
s
i
d
ed
i
a
m
e
t
e
r(
.
x
2
)

GeometricRe
u
o
n
s
:Th
ef
o
l
1owingr
e
l
a
t
i
o
n
sw
i
l
lb
eu
s
e
f
u
li
nl
a
t
e
rc
a
Jc
u
l
a
t
i
o
n
s
:
A
:a
r
e
ao
fc
r
o
s
s
c
t
i
o
n=0
.
2
5 *(d~-df)
1
:d
iame
i
c
a
lmomento
fi
n
e
r
t
i
a=
*
(d~ -d
t
)
/
6
4
Q
l
t
:f
i
r
s
tmomcnto
fa
r
e
aa
b
o
v
et
h
en
e
u
t
r
a
l i
sd
i
v
i
d
e
db
yt
h
i
c
k
n
e
s
s
=(d~ +dodi+dr)/6

E(
m
o
d
u
l
u
so
fc
l
a
s
t
i
c
i
t
y
)
:200E 9Pa

0
1
1(
a
l
l
o
w
a
b
l
cnorm
a
Js
):250E+0
6Pa
1
:
1
(
a
l
l
o
w
a
b
l
e
s
h
c
a
r
s
t
r
e
s
s
)
:
1
4
5
E
+
0
6
P
a
0

O
b
j
e
c
t
i
v
eF
u
n
c
t
i
o
n
: Theo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
ni
st
h
ewi
g
h
to
ft
h
e IO-mu
n
i
f
o
r
m
f
l
a
g
p
o
l
e
:

10m

8m

s
8
c
t
l
o
nAA

Weight
:
j
{
x
l
'X2)
:L*
A*
'
Y*
g

(
2
.
1
0
)

Constra
t
F
u
n
c
t
i
o
n
s
:
l
ewind l
o
a
dp
e
ru
n
i
tl
e
n
g
t
h (F
nt
h
ef
l
a
g
p
o
l
ei
s
D) o
c
a
I
c
u
l
a
t
e
da
s
Fo=
0
.
5*
P*
V~

*Cd*do

Thcb
e
n
d
i
n
gmomenta
tt
h
eb
a
s
eo
ft
h
ep
o
l
ed
u
et
oi
su
n
i
f
o
r
mwindl
o
a
dont
h
ee
n
t
i
p
o
l
ei
s
g
u
r
e2
.
8

F
l
a
g
p
o
l
ed
e
s
l
g
n
:E
xa
m
p
l
a2
.
3
.

Mw=0.5*
Fo L*L

f
/

2
.
3A
O
D
I
T
I
O
N
A
LE
X
A
M
P
L
E
S 67

66 G
R
A
P
H
I
C
A
LO
P
T
I
M
I
Z
A
T
I
O
N

j wew
i
l
li
n
c
l
u
d
eag
e
o
m
e
t
r
i
cc
o
n
s
t
r
a
i
n
t
:
Tod
i
s
c
o
u
r
a
g
es
o
l
u
t
i
o
n
swhered
o<d

Th
eb
e
n
d
i
n
gmomentduet
ot
h
ewindl
o
a
dont
h
ef
1agi
s

M~=
W
n
"
'
L
_
F-""
"
P
AJ
D

84(X1X2
)

0.001

(
2.
14
)

Bending(
n
o
n
n
a
l
)s
t
r
e
s
sa
tt
h
eb
a
s
eo
f ep
o
l
ei
s

S
i
d
eC
o
n
s
t
r
a
i
n
t
s
:
l
i
sd
e
f
i
n
e
st
h
ed
e
s
i
g
nr
e
g
i
o
nf
o
rt
h
es
e
a
r
c
h
.

'
b
c
n
d=
0
.
5 (Mw+MF) di

2cmSdoS1
0
0cm; 2cmSdiS100cm

Nonnals
e
s
sduet
ot
h
ew
e
i
g
h
ti
s

w
e
i
g
h
t=
'
Y g L
T
o
t
a
1nonna
1sU
s
s
e
st
ober
e
s
i
s
t
e
df
o
rd
e
s
i
g
ni
sesumo
f nonna
1s
t
r
e
s
s
e
s
computeda
b
o
v
e
.I
n
c
o
r
p
o
r
a
t
i
n
gt
h
ef
a
c
t
o
ro
fs
a
f
e
t
yandt
h
ea
1l
o
w
a
b
l
es
t
r
e
s
sf
r
o
m
m
a
t
e
r
ia
1va
1u
e
s
t
h
ef
i
r
s
ti
n
u
a
1
it
ynsU
i
n
tc
a
nb
es
e
tupa
s
gl(XI'~):

I
F
S
bcnd +

w
c
i
g
h
tS
n
U

(
2
.
1
1
)

Themaximums
h
e
a
rl
o
a
di
nt
h
ec
r
o
s
ss
e
c
t
i
o
ni
s

s=WF+Fo"
'L
Themaximums
h
e
a
rs
t
r
e
s
si
nt
h
ep
o
l
ei
s

=S*QI(I

t
)

Th
es
e
c
o
n
di
n
e
q
ua
1
i
t
yc
o
n
s
t
r
a
i
n
tbedonh
a
n
d
l
i
n
ge rs
t
r
e
s
s
e
si
nt
h
ef
1agpole

(
2
.
1
5
)

MATLABC
ode:The m
f
i
l
e
sf
o
rt
h
i
s example a
r
eg
i
v
e
nb
e
l
o
w
. An i
m
p
o
r
t
a
n
t
o
b
s
e
r
v
a
t
i
o
ni
ni
sproblem

ands
t
r
u
c
t
u
r
a
le
n
g
i
n
e
e
r
i
n
gp
r
o
b
l
e
m
si
np
a
r
t
i
c
u
l
a
r

i
s
o
r
d
e
ro
fm
a
g
n
i
t
u
d
eo
ft
h
eq
u
a
n
t
i
t
i
e
si
nt
h
ec
o
n
s
t
r
a
i
n
i
n
g u
a
t
i
o
n
s
.l
es
t
r
e
s
s
c
o
o
sn
t
sa
r
eo
f eo
r
d
e
ro
f108
+0
6
w
h
i
l
et
h
ed
i
s
p
l
a
c
e
m
e
o
tl
e
n
n
sa
r
eo
feo
r
d
e
r
o
f1
0
8
0
2
.Mostn
u
m
e
r
i
ca
1t
e
c
h
n
i
q
u
e
ss
g
g
l
et
oh
a
n
d
l
et
h
i
sr
a
n
g
e
.T
y
p
i
ca
1l
y
s
u
c
h
p
r
o
b
l
e
m
s0 dt
obenonna
1i
z
e
db
e
f
o
r
eb
e
i
o
gs
o
I
v
e
d
'I
ti
se
s
s
e
n
t
ia
1i
na
p
p
l
y
i
n
g
n
u
m
e
r
i
c
a
lt
e
c
h
n
i
q
u
e
su
s
e
di
no
p
t
i
m
i
z
a
t
i
o
n
.
T
h
i
sexamplei
sp
l
o
t
t
e
di
ntwop 1
5
.I
nt
h
ef
i
r
s
tp
a
r
t
eachi
n
e
q
ua
1
i
t
yC
O
I
i
n
ti
s
1
o
n
e
.Twoc
u
r
v
e
sa
r
eshowlIf
o
re
a
c
hc
o
n
s
n
t
.T
h
i
sa
v
o
i
d
sc
l
u
t
t
e
r
.
'
he
i
n
v
e
s
t
i
g
a
t
e
da
s
i
d
ef
o
rt
h
el
o
c
a
t
i
o
no
ft
h
eh
a
s
hmark00t
h
ec
o
n
s
t
r
a
i
n
ti
sd
e
t
e
n
n
i
n
e
dbyd
r
a
w
i
n
gt
h
e
s
e
C
Vinc
o
l
o
r
.l
eb
l
u
ec
o
l
o
ri
n
d
i
c
a
t
e
st
h
ef
e
a
s
i
b
l
ed
i
r
e
c
t
i
o
n
.A
l
t
e
m
a
t
e
l
y
quiver
p
l
o
t
sc
a
nb
eu
s
e
d
.Thes
e
c
o
n
dp
a
r
ti
sec
o
n
s
o
l
i
d
a
t
e
dc
u
r
v
es
h
o
w
n
.Thec
o
n
s
o
l
i
d
a
t
e
d
c
u
r
v
eh
a
st
ob
edrawn by r
e
m
o
v
i
n
gt
h
e comments on someoft
h
ec
o
d
e d
ti
sn
o
tn
e
e
d
e
d
.S
i
n
c
eeo
p
t
i
m
a
1s
o
l
u
t
i
o
nc
a
n
n
o
tbec
1e
a
r
l
y
commentingt
h
edea
e
s
t
a
b
l
i
s
h
e
d zoomf
e
a
t
u
r
eofMATLABi
su
s
e
dt
on
a
r
r
o
wdownt
h
es
o
l
u
t
i
o
n
.

2
.
:
.
J
.
.
.
.

(
the main script file)

Chapter 2
: Optimizaion with MATLAB
Dr. P.Venkataraman

1
8

g2(X ~):

g:aIJIFS

(
2
.
1
2
)

Example 2.3 Sec.2.3

l
et
h
i
r
dp
r
a
c
t
i
c
a
1c
o
o
s
t
r
a
i
o
ti
sb
a
s
e
d00t
h
ed
ef
1
e
c
t
i
o
oo
ft
h
et
o
po
fep
o
l
e
.
'
h
is
df
l
e
c
t
i
o
oduet
oau
n
i
f
o
n
nwindl
o
a
dont
h
ep
o
l
ei
s

o
w=Fo

0
1
<

L/
(
8 E*

Ref. 2.4

global statement is used to share same information

% between various m-files

global ELAS SIGALL TAUALL GAM FS GRAV


global RHO CD FLAGW SPEED LP L DELT

-----------"-----"

Thet
h
i
r
dc
o
o
s
t
r
a
i
o
tt
r
a
n
s
l
a
t
e
st
o

w+OFSO.u
o

"L _ WF L'
OF (
2 WF'
"L"
'L
p
)
/
(
E'
"

C3(X1X2)

graphicalsolutionu6ingMATLAB (twodesignvariables)
Optimal design of a Flag Pole for high winds

'---------"

Thede
f
1e
c
t
i
o
na
tt
h
et
o
pduet
ot
h
ef
1agwindl
o
a
da
ti
s

(
2
.
1
3
)

Initialize values
ELAS 200e+09i
SIGALL = 250E+06i
TAUALL =145e+06i

Modulus of elasticity -Pa

% allowab
e normal stress -Pa

allowable shear stress - Pa

I l i 7
I
I
I

2
.
3 ADDITIONALE
X
A
M
P
L
E
S 69

68 G
R
A
P
H
I
CAlOPTIMIZATION
GAM= 7860;
FS 2.5;
GRAV
9.81;

densityof material kg/m3


% factor of safety
g
ravitational acceleration

RHO
l.225;
CD l.0
;
FLAGW 5000;
SPEED 156.46;

density of air kg/m3

% drag coefficient

concentrated load on flag -N

I
/
S

"-

LP

8;

10;
DELT 0.05;

location of drag load on flag -m


lengthof pole
allowable deflection - m

--

glval
g2val
g3val
g4val

SIGALL/FS
TAUALL/FS
DELT
0.001

ylabel{'inside diameter''FontName''times'

FontSize'
12)
grid
hold on

right hand side values


for the constraints

----'------

x
0.02:0.01:1; the semi-colon at the end prevents
t
he echo
x2=0.025:0.01:1; these are also the side consraints
t x1 and x2 are vectors filled with numbers
note a way to avoid x1
x2
x2);
[X1 X2]
meshgrid{x1
generates matrices X1 and X2 corresponding to
vectors x1 and x2

figure a new figure window


x2
ineq1[glval
glval]'r-');
contour{x1
hold on
draw another contour at 10 the constraint boundary
contour{x1
x2
ineq1[O.l*glval
O.l*glval]'b-');
set{gca
'xtick'[
0 0.2 0.4 0.6 0.8 l.O])
'ytick'[
0 0.2 0.4 0.6 0.8 l.O])
set{gca
xlabel{'outside diameter'
'FontName''times'
'.
.
.
FontSize'
12);
ylabel{'inside diameter''FontName''times'
'
.
FontSize'
12);
hold off
grid

X2);
f1 obj_ex3(X1
t
he objective function is evaluated over the entire
mesh
Constraints are evaluated
X2);
ineq1 ineq1_ex3{X1
ineq2 ineq2_ex3(X1
X2);
ineq3
ineq3_e
x3{
X1
X2);
X2);
ineq4
ineq4_ex3(X1

han1]
contour{xl
x2
f1[
0
10000
50000
[C1
100000 150000 200000250000300000]'g-');
clabel(C1
han1);
set(gca
'xtick'[
0 0.2 0.4 0.6 0.8 l.O])
'ytick'[
0 0.2 0.4 0.6 0.8 1.0))
set(gca
xlabel('outside diameter''FontName''times'

FontSize'
12);

the following code maybe usefu1 for the consolidated


figure. Not used here __ uncomment below
[
C
2
han2] contour{x1
x2
ineq1[glvalglval]'r-');
clabel{C2
han2);
set{h2
'Linewidth'
1
)
k2 gtext('gl');
set{k2'FontName''Times''FontWeight''bold'
.
FontSize'
14

Color'
red')

figure
x2
ineq2[g2val
g2val]'r-');
contour(x1
hold on
x2
ineq2[
0.1*g2val
0.1*g2val]

b-');
contour(x1
'xtick'[
O 0.2 0.4 0.6 0.8 l.0))
set{gca
set{gca
'ytick'[
0 0.2 0.4 0.6 0.8 l.0))
'FontName''times'
xlabel{'outside diameter'

FontSize'
12);
'FontName'
ylabel{'inside diameter'
'times' 'FontSize'
12);
hold off
grid
[
C
3
han3]

contour(x1
x2
ineq2[g2val
g2val]'r--');

c
label{C3
han3);
s
et(h3
'LineWidth'
1
)

2
.
3A
O
O
I
T
I
O
N
A
LE
X
A
M
P
L
E
S 7
1

70 G
R
A
P
H
I
C
A
LO
P
T
I
M
I
Z
A
T
I
O
N
k3

gtext('g2');

se
(k3'FontName'
'Times'
'FontWeight''bold'

the equality and in~quality constraints are

14
'Color''red')

FontSize'

not written with 0 on the right hand side.

If you do write them that way you would have t

figure
X2
ineq3[g3val
g3val]'r-');
contour(x1
hold on
x2
ineq3[0.1*g3val
0.1*g3val]
'b-');
contour(xl
'xtick'[
0 0.2 0.4 0.6 0.8 1.0])
set(gca
'ytick'[
0 0.2 0.4 0.6 0.8 1.0])
set(gca
xlabel('outside diameter''FontName''times'

FontSize'
12);
'FontName'
'times'
ylabel('inside diameter'

FontSize'
12);
hold off
~rid

[C4
han4] m contour(x1
x2
ineq3[g3val
g3val]'b-');
clabel(C4
han4);
s
et(h4
'LineWidth'
1
)
k4 = g
text('g3')i
set(k4'FontName''Times''FontWeight''bold'
14
'Color'
'blue')

FontSize'

include [
0
0
] in the contour commands

3
.
.
.
.
.m (the first constraint)
1neal ex
function retval
ineq1_ex3(Xl
X2)
g
lobal statement is used to share same information
between various m-files
global ELAS SIGALL TAUALL GAM FS GRAV
global RHO CD FLAGW SPEED LP L DELT

AREA 0.25* pi*(X1.^2 - X2.^2)i matrix


INERTIA = pi*(X1.^4 X2.^4)!64i matrix
FD 0.5*RHO*SPEED*SPEED*CD*X1i
T1Ai
MW 0.25*FD*L*L.*X1.!INER
MF 0.5*FLAGW * LP*Xl.!INERT1Ai

figure
contour(x1
x2
ineq4[g4val
g4val]
'r-')i
hold on
contour(x1
x2
ineq4[0.001*g4val
0.001*g4val]
'b-')i
set(gca
'xtick'[
0 0.2 0.4 0.6 0.8 1.0])
'ytick'[
0 0.2 0.4 0.6 0.8 1.0])
set(gca
xlabel('outside diameter''FontName''times'

FontSize'
12)i
ylabel('inside diameter''FontName''times'

FontSize'
12);
hold off
grid

[C5
han5]

b--')i

set(h5'LineWidth'
1
)
k5 gtext('g4')i
set(k5
'FontName''Times''FontWeight''bold'
14
'Color'
'blue')

FontSize'

GAM GRAV*Li

retval MW + MF + SIGWi SIGW is added to al1 matrix


% elements
1nea2 ex3.m (the second constraint)
function retval
ineq2_ex3(X1
X2)
g
lobal statement is used to share same information
betweenvarious m-files
global ELA8 8IGALL TAUALL GAM FS GRAV
global RHO CD FLAGW SPEED LP L DELT
AREA 0.25* pi*(X1.^2 - X2.^2)i
INERTIA ~ pi*(X1.^4 - X2.^4)!64i
FD 0.5*RHO*SPEED*SPEED*CD*X1;
8
FLAGW + (FD L)i
(X1.*X1 + X1.*X2 + X2.*X2)!6.0i
retval 8.*0./NERTIA;

contour(xl
x2
ineq4[g4val
g4val]

clabel(C5
han5)i

SIGW

1t:L.m (the third constraint)


inea
function retval
ineq3_ex3(X1
X2)
global statement is used to share same info
nation
betweenvarious m-files
global ELAS.SGALLTAUALL GAM FS GRAV
global RHO CD FLAGW SPEED LP L DELT

D
l
T
I
O
N
A
LE
X
A
M
P
L
E
S 73
2
.
3 AD

72 G
R
A
P
H
I
C
A
Lo
p
n
M
I
Z
A
T
I
O
N

AREA 0.25* pi*(Xl.^2 - X2.^2);


INERTIA pi*(Xl.^4 X2.^4)/64;
FD = 0.5*RHO*SPEED*SPEED*CD*Xl;
dw FD*L^4./(8*E
S I
NERTIA);
df
(2.0 FLAGW*L^3 - FLAGW*L*L*LP)./(ELAS*NERTIA);
retval

dw + df;

4 e
x3.m (the fourth constraint)
in
function retval
ineq4_ex3(Xl
X2)
retval Xl X2;

r
o
b
l
c
m
.l
e
c
o
d
c
F
i
g
u
r
e2
.
9d
i
s
p
l
a
y
st
h
eg
r
a
p
h
i
c
a
lo
rc
o
n
s
o
l
i
d
a
t
e
ds
o
l
u
t
i
o
nt
o p
i
sa
v
a
i
l
a
b
l
e l
h
em
.
f
i
l
ea
b
o
v
e(
b
ycommentingt
h
ec
o
d
e
s c
r
e
a
l
ct
h
enewf
i
g
u
r
e
windowsandr
e
m
o
v
i
n
gt
h
ecommentsont
h
ec
o
d
c i
sc
u
r
r
c
n
t
l
yc
o
m
m
e
n
t
c
d
)
.Thc
1e
a
r
.F
i
g
u
2
.
10i
so
b
t
a
i
n
db
yzoomingi
nn
e
a
re
o
p
t
i
m
a
ls
o
l
u
t
i
o
ni
sn
o
tv
e
r
yc
n
e
i
g
h
b
o
r
h
o
o
do
f0
.
6
.Sech
e
l
pz
o
o
mf
o
ri
n
u
c
t
i
o
noni
t
su
s
e
.l
tc
a
nb
ea
c
h
i
c
v
e
dby

F
1
l
l
gP
o
l
cb
l
c
m

0
.
5
80
.
60
.
6
20
.
6
40
.
6
60
.
6
80
.
70
.
7
2
0
.
7
6
o
u
t
s
i
d
ed
i
a
m
C
l
c
r

0
.
8

F
ig
u
r
e2
.
1
0G
r
a
p
h
i
c
a
ls
o
l
u
t
l
o
n(
z
o
o
m
e
d
)
: :
a
m
p
l
e
2
.
3
.
0
.
81

t
y
p
i
n
gzooma
tt
h
ew
o
r
k
s
p
a
c
epromptandu
s
i
n
gt
h
emouset
od
r
a
gar
e
c
t
a
n
g
l
ea
r
o
u
n
d

UES-

aoa

unu

er
e
g
i
o
nt
h
a
tn
d
st
obee
n
l
a
r
g
e
d
.FromF
i
g
u
r
e2
.
1
0cs
o
l
u
t
i
o
ni
sa
r
o
u
n
dt
h
e
l
l
ci
n
s
i
d
ed
i
a
m
e
t
e
ro
f0
.
6
5m.Typingzoomi
n
o
u
t
s
i
d
cd
i
a
m
e
t
e
ro
f0
.
6
8m dt
w
o
r
k
s
p
a
c
ea
g
a
i
nw
i
l
lt
o
g
g
l
e b
a
c
kt
ono
nnaJs
t
a
t
e
.I
nF
i
g
u
r
e2
.
1
0 t
i
c
k
o
rb
c
t
t
c
ri
n
t
e
t
a
t
i
o
n
marksa
r
ep
l
a
c
e
dont
h
ef
i
g
u
r
et
h
r
o
u
g
hecommandwindowf

o
ft
h
es
o
l
u
t
i
o
n
.
Theg
r
a
p
h
i
c
si
nea
b
o
v
ee
x
n
p
l
ew
e
r
ec
r
e
a
t
e
du
s
i
n
gt
h
esames
t
a
t
e
m
e
n
l
s
e
n
c
o
u
n
t
e
r
e
dp
r
e
v
i
o
u
s
l
y
.C
o
l
o
rc
o
n
t
o
u
r
sw
e
r
eu
s
e
dt
oe
s
t
a
b
l
i
s
ht
h
ef
e
a
s
i
b
l
er
e
g
i
o
n
.

h
e
zoomf
e
a
t
u
r
ewasemploy
t
oo
b
t
a
i
nab
e
t
t
c
re
s
t
i
m
a
t
co
ft
h
es
o
l
u
t
i
o
n
.
2
.
3
.
3 Example2
.
4

0
.
2

0
.
2

0
.
4
0
.
6
o
u
t
s
i
d
ed
i
a
m
e
l
e
r

F
l
g
u
r
e2
.
9G
r
a
p
h
l
c
a
ls
o
l
u
l
i
o
n
:E
x
a
m
p
l
e2
.
3
.

0
.
8

Thi
se
x
a
m
p
l
ei
sumea
r
c
ao
fh
e
a
tt
r
a
n
s
f
e
r
.Thep
r
o
b
l
e
mi
st
od
e
s
i
g
nat
r
i
a
n
g
u
l
a
r
m
a
l
l
e
s
tvolumet
h
a
tw
i
l
la
tl
e
a
s
td
e
l
i
v
e
rs
p
e
c
i
f
i
c
df
i
ne
f
f
i
c
i
e
n
c
i
e
s
.The
f
i
no
f s
g
r
a
p
h
i
c
a
li
t
t
u o
ft
h
i
sc
o
d
ei
sv
e
r
ys
i
m
i
l
a
rt
oExample2
.
3
.l
ni
sc
x
a
m
p
l
e
t
h
e
i
n
e
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
t
sa
r
ecomputeda
n
dr
e
t
u
m
e
df
r
o
mas
i
n
g
l
ef
u
n
c
t
i
o
nm
.
f
i
l
er
a
t
h
e
r
a
ns
e
p
a
r
a
t
cf
i
l
e
sc
o
n
s
i
d
e
r
e
di
nt
h
ep
r
e
v
i
o
u
se
x
a
m
p
l
e
.A
n
o
t
h
e
rncwf
e
a
l
u
r
ei
nt
h
i
s
e
x
a
m
p
l
ei
st
oi
n
v
o
k
es
p
e
c
i
a
lm
a
t
h
e
m
a
t
i
c
a
lf
u
n
c
t
i
o
n
st
h
es
e
s
se
If
u
n
c
t
i
o
n
st
h
a
ta

74

GRAPHICALOPT
lMIZATION

2
.
3 ADDITIONALEXAMPLES

75

a
v
a
i
l
a
b
l
ei
nMATLAB.sexamplea
l
s
oi
I
1u
s
t
r
sap
r
o
b
l
e
mwheret
h
eo
p
t
i
m
i
z
a
t
i
o
n
p
r
o
b
l
e
mc
a
nbea
d
e
q
u
a
t
e
l
yd
e
f
i
n
e
db
u
tt
h
es
o
l
u
t
i
o
ni
se
a
s
i
l
yd
e
t
e
r
m
i
n
e
df
r
o
mt
h
e
e
q
u
a
l
i
t
yands
i
d
ec
o
n
s
a
i
n
t
s
.I
no
t
h
e
rwords
t
h
ep
r
o
b
l
e
mc
a
ne
a
s
i
l
yaccommodate
a
d
d
i
t
i
o
n
a
ldemandingc
o
n
s
t
r
a
i
n
t
s
.I
nl
a
r
g
ecomplcxm
a
t
h
e
m
a
t
i
c
a
lm
o
d
e
l
sw
i
t
hmany
n
t
s
d
c
s
i
g
nv
a
r
i
a
b
l
c
si
ti
sn
o
te
a
s
yt
oe
n
s
u
r
et
h
a
ta
tl
e
a
s
to
n
eo
ft
h
ci
n
e
q
u
a
l
i
t
yc
o
n
s
t
r
I
Sa
c
t
i
v
e
.

Des
'
nProblem: M z
e amounto
fm
a
t
e
r
i
a
lu
s
e
di
nedi
o
fas
c
r
i
c
s
o
fi
d
e
n
t
i
c
a
lt
r
i
a
n
g
u
l
a
rf
i
n
st
h
a
tc
o
v
e
rag
i
v
c
na
r
e
aa
n
do
p
e
r
a
t
ea
to
rabovct
h
es
p
e
c
i
f
i
e
d
c
f
f
i
c
i
c
n
c
i

.
MathematlcalModel:T
h
i
st
y
p
eo
fp
r
o
b
l
e
ms
h
o
u
l
db
ea
b
l
et
oaccommodate
r
o
b
l
e
mi
ss
e
tupt
oh
a
v
etwod
e
s
i
g
n
s
e
v
e
r
a
ld
c
s
i
g
nv
a
r
i
a
b
l
e
s
.I
nt
h
i
ss
e
c
t
i
o
nep
v
a
r
i
a
b
l
.1
ef
i
nm
a
t
e
r
i
a
li
sa
l
u
m
i
n
u
m
.

F
J
g
u
r
e2
.
1
1 F
l
nd
e
s
l
g
nf
o
rh
e
a
tt
r
a
n
s
f
e
r
:E
x
a
m
p
l
e2.
4
.

C
o
n
s
t
r
a
i
n
tFunct

ns:Thc h
e
a
tt
r
a
n
s
f
e
re
q
u
a
t
i
o
n
sf
o
rt
h
ef
i
na
r
ca
v
a
i
l
a
b
l
ei
n
ef
i
r
s
tc
o
n
s
i
n
ti
sone
R
e
f
e
r
e
n
c
e7
.F
i
n
sa
r
et
y
p
i
c
a
l
l
ymountedonas
p
e
c
i
f
i
e
d . 1
ona
r
e
a
:

h
(
b
L
)
:A
f 0
.
0
1
5m2

(
2.
17
)

DesignParameters: Foraluminum
N
o
t
e
:i
sc
o
n
s
t
r
a
i
n
te
s
s
e
n
t
i
a
l
l
yf
i
x
e
st
h
ev
a
l
u
eo
fb
.I
neg
r
a
p
ht
h
i
sc
o
n
s
t
r
a
i
n
twi
l

b
eas
t
r
a
i
g
h
tl
i
n
ep
a
r
a
l
l
e
lt
oc
L
a
x
i
s
.
l
fi
swcret
ohappeni
nap
r
o
b
l
c
mw
i
t
hs
c
v
e
r
a
I
d
e
s
i
g
nv
a
r
i
a
b
l
e
si
twouldb
ep
r
u
d
e
n
tt
oe
l
i
m
i
n
a
t
ei
sv
a
r
i
a
b
l
cf
r
o
mt
h
em
a
t
h
e
m
a
t
i
c
a
l
l
e
t
e
r
.
modelbyi
d
e
n
t
i
f
y
i
n
gi
ta
sad
e
s
i
g
np
a
r
ee
f
f
i
c
i
e
n
c
yo
fas
i
n
g
l
ef
i
nc
a
nbee
s
t
a
b
l
i
s
h
e
d

=
=

2[
h 50W/m
c
o
n
v
e
c
t
i
o
nc
o
e
f
f
i
c
i
c
n
t
]
k 1
7
7W/m-K[
t
h
e
r
m
a
lc
o
n
d
u
c
t
i
v
i
t
y
)
N 20[
numbero
ff
i
n
s
]

W=O.1m [
w
i
d
o
f
t
h
ef
i
n
s
)

r (
l/mL)1(
2
m
L
)
l
I
(
2
mL)
o

s
amca
st
h
eb
a
s
el
e
n
g
t
ho
ft
h
et
r
i
a
n
g
u
l
a
rf
i
n
.
F
i
ngapi
st
h

w
h
e
r
e

DesignV
a
r
i
a
b
l
e
s
m =(
2

/
h )
1
1
2

b
:b
a
s
eo
ft
h
et
r
i
a
n
g
u
l
a
rf
i
n
L
:h
e
i
g
h
to
ft
h
ct
r
i
a
n
g
u
l
a
rf
i
n

del
'sa
r
eB
e
s
s
e
le
q
u
a
t
i
o
n
so
ft
h
ef
i
r
s
tk
i
n
d
.Thef
i
r
s
ti
n
c
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
ti
s

g(
b
L):

F
i
g
u
r
c2
.
1
1i
J
lu
s
a
t
egeom

yof f
i
n
s
.

f~

0
.
9
5

(
2.
18
)

Theo
v
e
r
a
l
lc
f
f
i
c
i
e
n
c
yf
o
rt
h
eh
e
a
tt
r
a
n
s
f
e
ri
s

GeometricR
e
l
a
t
l
o
n
s
:Th
f
o
l
l
o
w
i
n
ga
r
esomeo
ft
h
a
r
e
ac
a
l
c
u
l
a
t
i
o
n
st
h
a
ta
r
eu
s
c
d
l
a
t
e
ri
nt
h
ed
c
v
e
l
o
p
m
c
n
to
fccons
a
i
n
t
s
:
Af
=(
2N-1
)
*
bV
A
c=2*W*[
L
2+(
b
!
2)
2
]1
1
2
Ab=
(N-l
)*b*W
A
t N*Ac+Ab

T
Jo=l-N*41)(
Thef
i
n
a
li
n
e
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
ti
s

f
o
o
t
p
r
i
n
to
fef
i
na
n
dg
a
p
f
i
na
r
e
af
o
rh
e
a
tt
r
a
n
s
f
e
r
gapa
r

t
o
t
a
la
r
e
af
o
rh
e
a
tt
r
a
n
s
f
e
r

g2(b
L
)
:T
Jo~ 0
.
9
4

(
2.
19
)

SideC
o
n
s
t
r
a
i
n
t
s
:Thes
i
d
ec
o
n
s
a
i
n
t
s
O
b
j
e
c
t
i
v
eF
u
n
c
t
i
o
n
: Thet
o
t
a
lvolumeo
fma
i
a
lf
o
rt
h
ef
i
ni
s

j
(
b
L
) O.
5*N*W*b*L

0
.
0
0
1SbSO. 5
(
2.
16
)

0
.
0
1SLS0
.
0
3

(
2
.
2
0
)

76 G
R
A
P
H
I
C
A
LO
P
T
I
M
I
Z
A
T
I
O
N

2
.
3A
D
D
I
T
I
O
N
A
LE
X
A
M
P
L
E
S

MATLABC
ode:S
i
m
i
l
a
rt
oep
r
e
v
i
o
u
se
x
a
m
p
l
e
.t
h
eh
a
s
hm
a
r
k
sont
h
ei
n
e
q
u
a
l
i
t
y
a
i
n
tc
a
nb
ee
s
t
a
b
l
i
s
h
e
dbyp
l
o
t
t
i
n
gtwov
a
I
u
e
so
ft
h
ec
o
n
s
t
r
a
i
nt
:t
h
el
i
m
i
tv
a
l
u
e
c
o
n
s
b
e
e
nd
e
t
e
r
m
i
n
e
d
.t
h
e
nt
h
ec
o
n
s
o
l
i
d
a
t
e
df
i
g
u
r
ec
a
n
a
n
dav
a
l
u
eh
i
g
h
e
r
.Onceyhav
Ilt
h
ef
u
n
c
t
i
o
n
si
ns
e
p
a
r
a
t
ew
i
n
d
o
w
s
.
b
ee
s
t
a
b
l
i
s
h
e
d
.T
h
c
o
d
eshownh
e
r
ew
i
l
ldrawa
F
o
rt
h
ec
o
n
s
o
l
i
d
a
t
e
dp
l
o
tt
h
i
sc
o
d
en
e
c
d
st
ob
ee
d
i
t
e
d
.I
ti
sl
e
f
tt
ot
h
es
t
u
d
e
n
t m
e
x
e
r
c
l
s
e
.

eq1

the equality constraint

[ineq1ineq2]
ine~ex4(X1 X2);
inthe above note that the two inequalityconstraints
a
re obtained simultaneously. Keep in mind
that each constraint is a matrix

mTh
emains
c
r
i
p
tf
i
l
e

ready for the plots


[
C
1
han1]
contour(x1
x2
f1[0.000010.00002
.
.
.
0.000040.000060.00008O.OOOl]
'g');
clabel(Cl
han1);
set(gca

xtick'[0.001 0.0015 0.002 0.0025 0.003


0.0035 0.004 0.0045 0.005])
set(gca
'ytick'
[0.010 0.015 0.02 0.025 0.03 0.125]);
'FontName'
'times'
'FontSize'
12);
xlabel('fin length'
'FontName'
'times'
'FontSize'
1
2
)
yl el('finwidth'
grid
% hold on

Optimization with MATLAB


% Dr. P.Venkataraman
% Example 2.4 Sec.2.3

Chapter 2
:

graphicalsolutionusingMATLAB (twodesignvariables)
Minimumfin volume for efficient heat transfer
material Aluminum

-------------------

global statement is used to share same information


betweenvarious m-files
global N H K W AREA

eq1_ex4(X1
X2);

"--------
n
itializevalues

N
W

20
n
umber of fins
0.1
w
idthof fins

50.0 convectioncoefficient W/m*m


K 177.0 thermal conductivity W/m-K
AREA 0.015 available fin foot print area

'FontName'
'times'
'FontSize'
12);
xlabel('fin length'
ylabel('fin width'
'FontName'
'times'
'FontSize'
12)
grid

-------
r
ight hand limits for the functions
h1val AREA;
glval 0.95;
g2val 0.94;

----

x1=0.001:0.0001:0.005;
x2=0.01:0.001:0.03; x1 and x2 are vectors filledwith

numbers
[
X
1 X2]
meshgrid(x1
x2);
g
enerates matrices X1 and X2 corresponding to
v
ectors x1 and x2
X2);
f1 = obj_ex4(X1
the objective function is evaluated over the entire
mesh

Constraints are evaluated

figure open a new figure window


p10tting the equality constraint
x2
eq1[h1val
h1val]
'r-');
contour(x1
'xtick'[0.001 0.0015 0.002 0.0025 0.003
set(gca
0.0035 0.004 0.0045 0.005])
'
y
t
i
c
k
'[0.010 0.015 0
.
.
0
2 0.025 0.03 0.125]);
set(gca

I
'
t
'

p
lotting the first inequality constraint
figure
contour(
x
1
x2
ineq1[
0
.968
0.
968]

r-
);
hold on
contour(x1
x2
ineq1[0.969
0.969]
'b-');
'xtick'[0.001 0.0015 0.002 0.0025 0.003
set(gca
0.0035 0.004 0.0045 0.005])

'ytick'[0.010 0.015 0.02 0.025 0.03 0.125J);


set(gca
'FontName'
'times'
'FontSize'
12);
xlabel('fin length'
ylabel(
'fin width'

FontNar

'times'
FontSize'
12)
hold off
grid
p
lotting the second inequality constraint

2
.
4 AoomONAlM
A
T
l
.
A
DG
A
A
P
H
I
C
S 7
9

78 G
A
A
P
H
I
C
A
LO
P
T
I
M
I
Z
A
T
I
O
N

figure
x2
ineq2[g2val
g2val]'r-');
contour(x1
hold on
contour(x1
x2
ineq2[1.01*g2val
1.01*g2val]
'b-');
'xtick'[0.001 0.0015 0.002 0.0025 0.003
set(gca
0.0035 0.004 0.0045 0.005])
set(gca
'ytick'[0.010 0.015 0.02 0.025 0.03 0.125]);
'FontName'
'times'
'FontSize'
12);
xlabel('fin length'
ylabel('fin width''FontName'
'times''FontSize'
1
2
)
hold off
grid

1
m

0
.
0
4
5

i
O

nb.Lel
eo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
n

0
.
0
2
5

function retval obj_ex4(X1


X2)
volume of the fin
global N H K W AREA
retval 0.5*N*W*X1.*X2

qlJ.mThee
q
u
a
l
i
t
y i
n
t
1
.
5

function retval eql_ex4(Xl


X2)
the e
quality constraint on area
global N H K W AREA
retval
(2.0 N - 1)*W*Xl;
~ T
heinequ~lity

g
u
r
e
2
.
1
2

3
.
5

S
3
x
1
0

G
r
a
p
h
l
c
a
ls
x
a
m
p
l
e2
o
l
u
l
i
o
n
:E
.
4
.

2
.
4 ADDITIONALMATLABGRAPHICS

Thea
b
o
v
es
c
r
i
p
tf
i
l
e
ss
h
o
u
l
dd
i
s
p
l
a
ye
a
c
hf
u
n
c
t
i
o
ni
nas
c
p
a
r
a
t
ef
i
g
u
r
ew
i
n
d
o
w
.
l
u
ec
o
n
t
o
u
r
Thei
n
e
q
u
a
l
i
t
yc
o
n
s
t
n
l
i
n
tf
i
g
u
r
e
ss
h
o
u
l
dc
o
n
t
a
i
ntwoc
o
n
t
o
u
r
sw
i
t
hcb
e
a
s
i
b
l
er
e
g
i
o
n
.F
i
g
2
.
12d
i
s
p
l
a
y
sac
o
m
p
o
s
i
t
ep
l
o
tTh
es
o
l
u
t
i
o
ni
s
i
n
d
i
c
a
t
i
n
gef
d
e
t
c
n
n
i
n
e
dbyt
h
ee
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
ta
n
dt
h
el
o
w
e
rI
i
m
i
ton e
f
i
n
l
e
n
g
d
i
s
c
u
s
s
e
d
p
r
e
v
i
o
u
s
l
y
.

2
.
5

m)
f
i
nb
a
s
e
(

c
o
n
s
t
r
a
i
n
t
s

function [retlret2]
ine~ex4(Xl X2)
returns both the inequality constraints
global N H K W AREA
c 2*sqrt(2.0)*sqrt(/K)*X2./sqrt(
X
l
);
retl
(besseli(1
c
)
.
.
/
((0.5*c).*besseli(O
c)));
A
e 2.0*W*sqrt((X2.*X2 + .25*Xl.*X1));
Ab
(N-1)*W*X1;
e + Ab;
At N*A
Ar Af./At;
ret2
(1.0 -N *Ar. (
1 - ret1));

.
.

1
1
<
'

.
"

'

Th
i
ss
e
c
t
i
o
nprovidab
r
i
e
fe
d
d
i
t
i
o
n
a
lf
e
a
t
u
r
e
so
x
p
o
s
u
r
et
oa
fH
a
n
d
l
eG
r
a
p
h
i
c
s
d
i
n
gd
i
nc
1u
ec
o
v
e
r
a
g
ei
sv
e
r
ym
e
v
e
l
o
p
m
e
n
to
faGUIi
nM
A
T
L
A
B
.Th
o
d
e
s
ta
n
di
s
h
cs
1u
d
e
dh
c
r
ef
o
rt
a
k
eo
fc
o
m
p
l
c
t
e
n
e
s
sa
a
v
ea
c
c
o
m
p
l
i
s
h
c
do
u
rg
i
nc
sweh
o
a
lo
f
h
i
ss
r
e
v
i
o
u
ss
g
r
a
p
h
i
c
a
lo
p
t
i
m
i
z
a
t
i
o
n ep
e
c
t
i
o
n
s
.T
e
c
t
i
o
nc
a
nb
ea
v
o
i
d
e
d
m
p
h
a
s
i
z
e
df
o
rt
o
u
r
s
ea
p
o
s
t
p
o
n
e
do
re
h
ep
r
o
j
e
c
t
si
nec
si
th
a
snoa
d
d
i
t
i
o
n
a
l
r
a
p
h
i
c
a
lo
i
n
f
o
r
m
a
t
i
o
nf
o
rg
p
t
i
m
i
z
a
t
i
o
n
.I
td
o
n
t
a
i
nr
e
a
s
o
n
a
b
l
ci
o
e
sc
n
f
o
n
n
a
t
i
o
ni
n
u
n
d
c
r
s
t
a
n
d
i
n
gMATLABg
r
a
p
h
i
c
sa
n
dp
r
o
g
r
a
m
m
i
n
g
.Ther
e
a
d
e
ri
se
n
c
o
u
r
a
g
e
dt
ot
a
k
e
i
t
hev
e
r
s
i
o
nofMA
n
s
t
a
l
l
e
do
nh
somct
i
m
eg
e
t
t
i
n
gf
a
m
i
l
i
a
rw
TI.ABi
e
r
e
i
sm
a
c
h
i
n
e
.Th
s
u
a
l
l
ys
n
h
a
n
c
e
m
e
n
t
si
a
r
eu
i
g
n
i
f
i
c
a
n
te
e
r
s
i
o
n
so
A
T
L
A
B
.F
o
re
nnewcrv
fM
x
a
m
p
l
e

sb
.
2
)a
l
l
o
w
sy
MATLABV
i
r
s
i
o
n
5
.
3(
sbooki
a
s
e
donV
e
r
s
i
o
n5
o
ut
oc
u
s
t
o
m
i
z
ey
o
u
r
o
u
g
ha
1
o
n
da
n
n
o
t
a
t
i
o
n
st
tr
d
i
t
i
n
gi
t
f w
i
n
d
o
w
.
p
l
o
ta
d
da
oi
d
d
i
t
i
o
n
a
1e
wew
i
l
l
u
s
t
o
m
i
z
ey
o
u
rf
o
d
ea
Youdon
o
th
a
v
et
oc
i
g
u
r
et
l
o
n
e
.I
nt
h
i
ss
e
c
t
i
o
n
h
r
o
u
g
hc
s
i
n
gMATLABH
ed
r
a
p
h
i
c
ss
o weh
a
v
emore
a
1s
ob
e
v
e
l
o
p
i
n
gt
h
eGUIu
a
n
d
l
eG
w
a
n
d
l
eG
oprovidcaGUIc
r
e
a
t
i
o
nt
o
o
l
e
x
p
e
r
i
e
n
MA
TI.ABH
i
r
a
p
h
i
c
s
.MA
TI.ABd
a
g
u
i
d
e
tt
srecommendedt
x
p
l
o
r
e
.
h
eu
s
e
ri
oe
h
ep
a
r
i
a
b
l
e
s
.Anc

l
I
I
nt
r
e
v
i
o
u
ss
e
c
t
i
o
nt
h
e
r
ew
e
r
eo
w
od
e
s
i
g
nv
q
ys
p
a
c
e
d
n
l
yt
o
n
s
t
r
a
i
nt
.A ms
v
e
c
t
o
rwaso
o
re
a
r
i
a
b
l
cb
a
s
e
don s
i
d
cc
h
g
r
i
dw
a
s
b
t
a
i
n
e
df
a
c
hv

80 GRAPHICALO
P
T
I
M
I
Z
A
T
I
O
N

2
.
4A
O
O
I
T
I
O
N
A
LMA
TA
BG
R
A
P
H
I
C
S 8
1

e
ng
e
n
e
r
a
t
e
d
.AlIf
u
n
c
t
i
o
n
sw
e
r
ce
v
a
l
u
a
t
e
dont
h
i
sg
r
i
d
.C
o
n
t
o
u
rp
l
o
1
5weredrawn
f
o
rt
h
ef
u
n
c
t
i
o
n
sr
e
p
r
e
s
e
n
t
i
n
g o
b
j
e
c
t
i
v
ea
n
dc
o
n
s
t
r
a
i
n
t
s
.E
x
c
e
p
tf
o
rt
h
eo
b
j
e
c
t
i
v
e
f
u
n
c
t
i
o
nwhichd
i
s
p
l
a
y
e
ds
e
v
e
r
a
lc
o
n
t
o
u
r
s
t
h
ec
o
n
s
t
r
a
i
n
t
swereo
n
l
ydrawnf
o
ra
s
i
n
g
l
ev
a
l
u
c
.T
h
i
sp
r
o
c
e
d
u
r
es
u
g
g
e
s
t
st
h
ep
o
s
s
i
b
i
l
i
t
yo
fa
u
t
o
m
a
t
i
o
n
.U
s
e
ri
n
p
u
tw
i
l
lb
e
o
rt
h
ev
a
r
i
a
b
l
e
sa
sw
e
l
la
st
h
cf
u
n
c
t
i
o
n
s a
tw
i
l
lb
e
n
e
c
e
s
s
a
r
yf
o
rs
e
l
c
c
t
i
n
g ngcf
p
l
o
t
t
e
d
.
I
nt
h
i
ss
c
c
t
i
o
n
wew
i
l
lc
r
e
a
t
ep
l
o
t
t
i
n
gf
a
c
i
l
i
t
yt
oi
n
c
l
u
d
eaGUIt
h
a
tw
i
l
l

t. o
bjti
ntumh
a
st
h
el
i
n
e
t
e
x
t
andimageobject
.A
l
l0e
c
t
si
n
o
b
j
as
e
to
fp
r
o
p
e
r
t
i
e
sa
ti
su
s
u
a
l
l
yb
a
s
e
doni
1
5f
u
n
c
t
i
o
n
MATLABc
a
nbea
s
s
o
c
i
a
t
e
dwi

o
ru
s
e
f
u
l
n
c
s
s
.D
i
f
f
e
r
e
n
to
b
j
e
c
tt
y
p
e
sh
a
v
ed
i
f
f
e
r
e
n
ts
e
t
so
fp
r
o
p
e
r
t
i
e
s
.Eacho
ft
h
c
s
p
r
o
p
e
e
si
sd
e
s
c
r
i
b
c
dt
h
r
o
u
g
hap
a
i
ro
fr
e
l
a
t
e
di
n
f
o
r
m
a
t
i
o
n(
n
a
m
e
v
a
l
u
e
)
. ef
i
r
s
t
e
l
e
m
e
n
to
ft
h
i
sp
a
i
ri
st
h
enamcf
o
rt
h
ep
r
o
p
e
r
t
yand t
h
es
e
c
o
n
di
st
h
ev
a
l
u
e
c
o
r
r
e
s
p
o
n
d
i
n
gt
o p
r
o
p
e
r
t
y
.Wheneo
e
c
ti
sc
r
e
a
t
e
d
i
ti
n
h
e
r
i
t
sa
l
lo
ft
h
ed
e
f
a
u
l
t
p
r
o
p
e
e
s
.F
o
re
x
a
m
p
l
e
:

fp
l
o
1
5
O
b
t
a
i
n enumbero

sin(2*)'-go'
plot(

LineWid
h'2

MarkerEdgeColor''
k
'

MarkerFaceColor'[0.49 1 0.63]
MarkerSize'
12
)

. Th
er
a
n
g
eandi
n
c
r
c
m
e
n
to
ft
h
ed
e
s
i
g
nv
a
r
i
a
b
l
e
s
x/u
t
i
l
i
t
ya
tt
h
eu
s
e
rw
i
l
lu
s
t
oi
d
e
n
t
i
f
yef
u
n
c
t
i
o
nt
o
Useaf
i
l
es
e
l
e
c
t
i
o
nbo
b
ep
l
o
t
t
e
d

Promptt
h
eu
s
e
rf
o
rt
h
enumbero
fc
o
n
t
o
u
r
f
o
re
a
c
hp
l
o
t

Allowselectionofcontourlevels
Creatctheplot

Thel
a
s
tf
o
u
r1
in
e
sa
b
o
v
cr
e
p
r
e
s
e
n
tf
o
u
rp
r
o
p
e
r
t
i
e
so
fep
l
o
t
t
e
dl
i
n
et
h
a
ti
sb
e
i
n
g
ep
r
o
p
e
r
t
yn
a
r
n
o
neI
c
f
tc
a
nbee
a
s
i
l
yu
n
d
e
r
s
t
o
o
d
.l
ef
o
r
m
a
ts
i
z
e
c
h
a
n
g
e
d
.Th
o
l
o
rv
a
l
u
e
sb
e
t
w
e
e
nz
e
r
o
u
n
i
t
sa
r
ei
np
o
i
n
t
s
.Thec
o
l
o
rv
a
l
u
ei
sarowv
c
c
t
o
ro
f c
a
n
doncr
e
p
r
e
s
e
n
t
i
n
gt
h
er
e
d
g
r
e
e
n
andb
l
u
ev
a
l
u
e
s
r
e
s
p
e
c
t
i
v
e
l
y
.T
h
e
r
ea
r
es
e
v
c
r
a
l
o
t
h
rp
r
o
p
e
e
so
ft
h
ep
l
o
t
t
e
dl
i
n
ewhichw
i
l
lb
es
c
ta
tt
h
c
i
rd
e
f
a
u
l
tv
a
l
u
e
s
.Th
c
is
ta
l
lo
b
j
e
c
1
5andt
h
c
i
rp
r
o
p
e
e
s
.
MATLABo
n
l
i
n
er
c
f
e
r
e
n
c
es
h
o
u
l
dI
I
ng
e
n
e
r
a
l
.o
b
j
e
c
tp
r
o
p
e
r
t
i
e
sc
l
l
nb
ec
h
a
n
g
e
dbyu
s
i
n
gt
h
es
e
rf
u
n
c
t
i
o
nt
oc
h
a
n
g
e
t
h
e
i
rv
a
l
u
e
s
.S
i
m
i
l
a
r
l
y
o
b
j
e
c
tp
r
o
p
e
r
t
i
e
sc
b
er
c
a
dbyt
h
eu
s
eo
ft
h
eg
e
tf
u
n
c
t
i
o
n
.To
i
n
q
u
i
r
ea
b
o
u
tt
h
ep
r
o
p
e
r
t
yo
fa
n0t:

l
ee
x
c
r
c
i
s
ew
i
l
lr
e
q
u
i
r
ctwom
f
i
l
e
s
:o
n
e10s
e
tupeGUIa
n
deo
rt
oa
s
s
i
s
ti
n

t
h
ep
l
o
t
t
i
n
g
.Th
eu
s
e
rw
i
l
lbea
b
l
et
ot
w
e
a
kt
h
cp
l
o
ti
nt
h
eCommandwindowa
f
t
e
ri
t
i
sd
r
a
w
n
.Thef
o
l
l
o
w
i
n
gs
u
b
s
e
c
t
i
o
n
sp

s
e
n
tab
r
i
e
fi
n
t
r
o
d
u
c
t
i
o
nt
oH
a
n
d
l
cG
r
a
p
h
i
c
s
a
n
dGUIc
o
n
t
r
o
l
si
nV
e
r
s
i
o
n5
.
3
.
2
.
4
.
1 HandleGraphics
MA
T
I
.
.
AB H
andleG
r
a
p
h
i
c
sr
e
f
e
rt
oac
o
l
l
e
c
t
i
o
no
fl
o
w
l
e
v
e
lg
r
a
p
h
i
cr
o
u
t
i
n
e
st
h
a
t
i
nthcMA
.A
Bf
i
g
u
r
ewindow.
a
c
t
u
a
l
l
yg
e
n
e
r
a
t
co
rc
a
r
r
yo
u
tg
r
a
p
h
i
c
a
lc
h
a
n
g
e
syous
F
o
rt
h
emostp
a
r
tt
h
e
s
er
o
u
t
i
n
e
sa
r
et
r
a
n
s
p
a
r
e
n
t dt
y
p
i
c
a
l
l
yu
s
e
r
sn
c
e
dn
o
tbea
w
a
r
c
o
ft
h
e
m
.I
nf
a
c

tt
ou
n
d
e
r
s
t
a
n
do
p
t
i
m
i
z
a
t
i
o
nandt
op
r
o
g
r
a
mt
h
et
e
c
h
n
i
q
u
e
si
nMA
T
I
.
.
A
B
i
ti
sn
o
tn
e
c
e
s
s
a
r
yt
oknowa
b
o
u
tH
a
n
d
l
eG
r
a
p
h
i
c
ss
i
n
c
eu
s
i
n
gt
h
eh
i
g
h
e
r
l
e
v
e
lp
l
o
t
commandss
u
c
ha
sp
l
o
tandc
o
n
t
o
u
ri
ss
u
f
f
i
c
i
e
n
tt
og
e
tt
h
ej
o
bd
o
n
e
.
"
T
h
e
s
ccommands
a
c
t
u
a
l
l
yk
i
c
ki
ns
c
v
c
r
a
lo
feHandleG
r
a
p
h
i
c
sr
o
u
t
i
n
e
s10d
i
s
p
l
a
y ep
l
o
tont
h
e
l
yf
e
a
t
u
r
eo
fMA
T
I
.
.
A
Bg
r
a
p
h
i
c
si
s i
ti
si
m
p
l
e
m
e
n
t
e
d
f
i
g
u
r
e
.A notewor
o
b
j
e
c
to
r
i
e
n
t
e
dm
a
n
n
c
r
.Wh
a
tt
h
i
smeansi
st
h
a
tm
o
s
to
ft
h
e e
syous
e
eont
h
e
I
ikea
x
e
sl
a
b
e
l
s
t
c
x

t andl
i
n
e
s
a
r
ca
l
lo
b
j
e
c
1
5
.Thi
smcanst
h
a
tmosto
ft
h
e
f
i
g
u
r
c
g
r
a
p
h
i
c
a
li
t
e
m
sont
h
ef
i
g
u
r
eh
a
v
ep
r
o
p
e
r
t
i
e
s C b
ec
h
a
n
g
e
dr
o
u
g
hprogram
c
o
d
e
.I
nt
h
ep
r
e
v
i
o
u
ss
e
c
t
i
o
n
sweu
s
e
dt
h
i
st
om
a
n
i
p
u
l
a
t
ec
o
l
o
r

f
o
n
ts
i
z
e
l
i
n
w
id

a
n
dt
h
eI
ik
e
.
I
no
r
d
e
rt
oc
h
a
n
g
eep
r
o
p
e
e
so
fo
b
j
e
c
t
s
i
ti
sn
e
c
s
a
r
yt
oi
d
e
n
t
i
f
yo
rr
e
f
e
rt
o
t
h
c
m
.O
b
j
e
c
t
si
nMATLABa
r
ei
d
e
n
t
i
f
i
e
dbyau
n
i
q
u
enumberr
e
f
e
r
r
e
d

a
sah
a
n
d
l
e
.
a
r
t
o
f
H
a
n
d
l
e
sc
a
nb
ea
s
s
i
g
n
e
dt
oa
l
lo
b
j
e
c
1
5whent
h
e
ya
r
cc
r
c
a
t
c
d
.Someo
b
j
e
c
1
5 p
o
t
h
e
ro
b
j
c
c
t
s
.O
b
j
e
c
t
st
h
a
tc
o
n
t
a
i
no
t
h
e
ro
b
j
e
c
1
5a
r
cr
e
f
e
r
r
e
dt
oa
sc
o
n
t
a
i
n
e
ro
rp
a
r
e
n
t
. Th
e
r
ei
sad
e
f
i
n
i
t
ch
i
e
r
a
r
c
h
i
s
c
t
u
r
eamongo
b
j
e
c
t
s
.
'
h
er
o
o
to
fa
l
l
MATLABg
r
a
p
h
i
c01
5i
sef
i
g
u
r
eo
b
jt.Th
ef
i
g
u
r
e0e
c
tf
o
rexamplec
a
nh
o
l
d
s
e
v
c
r
a
lGUIc
o
n
t
r
o
l
s
whicha
r
eknowna
su
i
c
o
n
t
r
o
lo
b
j
e
c
1
5
.I
t3
1
s
oc
o
n
t
a
i
n
st
h
e e
s

get(handle
'Property Name')
ge(h_1
'Color')

I
nt
h
ea
b
o
v
e
h
_
li
seh
a
n
d
l
eo
ft
h
eo
b
j
e
c
twhosec
o
l
o
ri
sd
e
s
i
r
c
d
.S
i
m
i
l
a
r
l
y
se(handle 'Property Name' 'Property Value')
1 0 0])
set(h_l'Color''r') or set(h_1'Color'[

i
nt
h
ea
b
o
v
e v
a
l
u
eo
fC
o
l
o
rf
o
rt
h
eo
b
j
c
c
tr
e
p
r
e
s
c
n
t
byt
h
eh
a
n
d
l
ch
_
li
ss
e
t10
r
c
di
ntwow
a
y
s
.Thes
c
c
o
n
ds
p
e
c
i
f
i

t
h
ec
o
l
o
rr
e
dt
h
r
o
u
g
ht
h
ev
e
c
t
o
rv
a
l
u
c
.I
ti
s
e
x
p
e
c
t
c
dt
h
a
tt
h
eo
b
j
e
c
ti
d
e
n
t
i
f
i
e
dt
h
r
o
u
g
hh
_
1s
h
o
u
l
dh
a
v
eaC
o
l
o
rp
r
o
p
e
r
t
y
.MA
T
t
.
A
B
w
i
l
li
n
f
o
r
myouo
ft
h
ce
r
r
o
ri
nc
a
s
eyoua
r
ea
s
s
i
g
n
i
n
gap
r
o
p
e
r
t
y
/
v
a
l
u
e t
i
sn
o
tv
a
l
i
d
f
o
rt
h
eo
b
j
e
ct
.
2
.
4
.
2 GraphlcalUserI
n
t
e
r
f
a
c
e

Themosts
i
g
n
i
f
i
c
a
n
tf
e
a
t
u
r
co
fs
o
f
t
w
a
r
ed
e
s
i
g
nandd
e
v
e
l
o
p
m
e
n
tt
o
d
a
yi
st
h
e
r
i
e
f
l
yi
I
Iu
s
t
r
a
t
e
di
nep
r
e
v
i
o
u
ss
e
c
t
i
o
n
.
o
b
j
e
c
t
o
r
i
e
n
t
e
dd
e
s
i
g
np
a
r
a
d
i
g
m wasb
Th
ec
o
m
b
i
n
a
t
i
o
no
ft
h
ea
b
o
v
ew
i
t
he
v
e
n
t
d
r
i
v
e
nprogrammingi
sb
o
t
han
a
t
u
r
a
la
n
da
q
u
i
r
e
df
c
a
t
u
r
ci
nc
u
r
r
e
n
ts
o
f
t
w
a
r
cp
r
o
g
r
a
m
s
.I
n
s
t
e
a
do
ft
h
ep
r
o
g
r
a
m
sc
o
n
t
r
o
l
l
e
db
y

82

G
R
A
P
H
I
C
A
LO
P
T
I
M
I
Z
A
T
I
O
N

2
.
4A
D
D
I
T
I
O
N
A
LM
A
T
L
A
BG
R
A
P
H
I
C
S

83

E
J

ap
r
e
d
e
t
e
r
m
i
n
e
dp
r
o
c
e
s
s
i
n
gs
u
e
n
c
e(
b
a
t
c
hp
r
o
c
e
s
s
i
n
g
)
t
o
d
a
y
'
sp
r
o
c
e
s
s
i
n
gi
s
t
e
dmousec
l
i
c
k
sa
n
db
u
t
t
o
n
s
.Th
e
s
eu
s
e
r
govemedbyt
h
eu
s
e
rt
h
r
o
u
g
hsomed
i
r
l
s
e
l
e
c
t
i
o
n
sc
a
u
s
ec
e
r
t
a
i
ne
v
e
n
t
st
h
a
ti
n
s
t
r
u
c
tt
h
es
o
f
t
w
a
r
et
oc
a
r
r
yo
u
tsomea
c
t
i
o
n
.The
o 1f
i
r
eeevenareu
s
u
a
l
l
yr
e
f
e
r
r
e
d1
0 u
s
e
ri
n
t
e
e
lcmc . I
ti
s
i
m
p
o
r
t
a
n
tt
ou
n
d
e
r
s
t
a
n
dt
h
a
tt
h
csamee
v
c
n
t
sc
a
nb
et
r
i
g
g
e
r
e
dbyo
b
t
a
i
n
i
n
gu
s
e
r
r
e
s
p
o
n
s
ct
os
c
e
np
r
o
m
p
t
sa
lt
h
ecommandl
i
n
c(
u
n
a
t
t
r
a
c
t
i
v
c
)
.U
s
i
n
gb
u
t
t
o
n
st
o
a
c
h
i
e
v
et
h
esamer
e
s
u
l
ti
5moreg
l
a
m
o
r
o
u
sa
l
t
h
o
u
g
hi
tu
s
e
smorecompu
t
i
o
n

s
o
u
r
c
e
s
.l
nf
a
c

tf
o
rc
o
m
m
e
r
c
i
a
ls
u
c
c
e
s
si
ti
sn
e
s
s
a
r
yt
oi
n
c
l
u
d
et
h
i
smechanismo
f
u
s
e
ri
n
t
e
r
.
a
c
t
i
o
ni
nt
h
es
o
f
t
w
a
r
ep
r
o
d
u
c.
tAl
s
o
.u
s
e
r
se
x
p
e
c
tt
h
e
s
ee
l
e
m
e
n
t
s1
0b
e
h
a
v
e
i
nas
t
a
n
d
a
r
dw
a
y
.T
h
e
r
ea
r
emanye
l
e
m
e
n
t
s t
h
eu
s
e
rc
a
nu
s
et
oi
n
t
e
r
a
c
tw
i
t
h
p
r
o
g
r
a
m
5
.Manya
p
p
l
i
c
a
t
i
o
n
sa
c
r
o
s
ss
e
v
e
r
a
ld
i
s
c
i
p
l
i
n
e
sh
a
v
cnowd
e
f
i
n
e
daminimum
5
t
a
n
d
a
r
dc
o
l
l
e
c
t
i
o
no
fe
s
ee
l
e
m
e
n
t
s
.MATLAB5
0p
r
o
v
i
d
c
st
h
e
m
.Th
eo
r
g
a
n
i
z
a
t
i
o
n
s
ee
l
e
m
e
n
l
sa
n
dt
h
e
i
rp
r
e
s
e
n
t
a
t
i
o
nt
ot
h
eu
s
e
ri
nag
r
a
p
h
l
c
a
lmanneri
si
d
e
n
t
i
f
i
e
d
o
fe
a
segr
p
h
l
c
a
lu
s
e
ri
n
t
e
r
f
a
.A n
umbero
fGUI-basedt

1
s
.
n
c
t
i
o
n
s
. a
v
a
i
l
a
b
l
e
a
tt
h
eMATLABs
i
t
e
.
T
h
iss
e
c
t
i
o
np
r
o
v
i
d
e
so
n
l
yab
r
i
e
fi
n
t
r
o
d
u
c
t
i
o
nt
oee
l
e
m
e
n
t
su
s
e
di
nt
h
ee
x
a
m
p
l
e
.
Th
s
ee
l
c
m
e
n
t
sa
r
ep
r
i
m
a
r
i
l
yu
s
e
dt
h
r
o
u
g
hMATLABH
a
n
d
l
cG
r
a
p
h
i
c
sp
r
o
g
r
a
m
m
i
n
g
.
C
o
n
s
i
d
e
rt
h
ec
a
s
co
fab
u
t
t
o
n a
tt
h
eu
s
e
rc
a
np
u
s
ht
oc
a
u
s
esomea
c
t
i
o
n
.r
e
f
e
r
r
e
dt
o
a
sap
u
s
hb
u
t
t
o
ni
nt
h
cs
u
b
s
e
q
u
e
n
td
i
s
c
u
s
s
i
o
n
.Att
h
eo
u
t
s
e
t
.t
h
e
r
er
u
t
wop
r
i
n
c
i
p
a
l
lo

r
e
q
u
i
r
e
m
c
n
t
sa
ta
r
ec
x
p
e
c
t
ft
h
ep
u
s
hb
u
t
t
o
n
.F
i
r
s

tt
h
e e
rm
u
s
tb
ea
b
l
et
os
e
e
i
t
.S
e
c
o
n
d
.whcneu
s
e
rc
l
i
c
k
soni
t(
s
i
n
g
l
ev
e
r
s
u
sd
o
u
b
l
ec
l
i
c
k
sc
a
nb
ed
i
s
t
i
n
g
u
i
s
h
e
d
s
omee
x
p
e
c
t
e
d
i
nmanys
o
f
t
w
a
r
ed
e
v
e
l
o
p
m
e
n
tc
o
d
e
s
)t
h
eb
u
t
t
o
nmust spondwi

t
.I
th
a
smanyp
r
o
p
e
t
h
a
t
c
a
n
a
c
t
i
o
n
.I
nMA
T
L
A
Bt
h
ep
u
s
hb
u
t
t
o
ni
sag
r
a
p
h
i
c
a
lo
t
beu
s
e
dt
od
e
s
c
r
i
b
ei
t
sapp
r
a
n
c
eont
h
cs
c
r
e
c
n
.I
nMA'
L
A
B
.a
l
lg
r
a
p
h
i
c
a
lo
b
j
e
c
t
sc
a
n
onlyappe
i
naf
i
g
u
r
cwindow.Henccaf
i
g
u
r
ewindowi
se
s
s
e
n
t
i
a
lt
od
i
s
p
l
a
yt
h
c
p
u
s
hb
u
t
t
o
n
. i
sf
i
g
u
r
ewindowactsac
o
n
t
a
i
n
e
ra
n
dc
a
nc
o
n
t
a
i
nmanyu
s
e
r
i
n
t
e
r
f
a
c
ee
l
c
m
c
n
t
s
.C
o
n
s
i
d
e
r
i
n
gt
h
ep
r
e
s
e
n
t
a
t
i
o
no
ft
h
eb
u
t
t
o
nont
h
cs
c
r
c
c
ni
ti
s
n
e
ssaryt
od
c
c
i
d
ewheret
h
eb
u
t
t
o
ns
h
o
u
l
da
p
p
r
onefigw
w
indow
howb
i
gi
t
s
h
o
u
l
dbe
w
h
a
tl
a
b
e
li
ts
h
o
u
l
dh
a
v
e(
t
y
p
i
c
a
l
l
yi
n
d
i
c
a
t
e
s
k
i
n
do
fa
c
t
i
o
ni
ti
sl
i
k
e
l
y
cause) ds
oo
n
.S
i
n
c
ct
h
ep
u
s
hb
u
t
t
o
ni
sag
r
a
p
h
i
c
a
lo
b
j
e
c

t mucho
ft
h
e
s
e
p
r
o
p
e
e
sc
b
es
e
ti
ft
h
e
r
ci
sah
a
n
d
l
ea
v
a
i
l
a
b
l
ef
o
rI
h
ep
u
s
h
b
u
t
t
o
no
b
j
c
ct
.Oncet
h
e
g
r
a
p
h
l
1a
p
p
e
a
r
a
n
c
ei
st
a
k

'Position'[2.5
0.1
0.7.
2
]
'BackgroundColor'
'red'

String'
EXIT'
'Callback'
'close');

I
nt
h
ea
b
o
v
cI
in
e
so
fc
o
d
c
.Hpushi
st
h
eh
a
n
d
l
et
ot
h
ep
u
s
h
b
u
t
t
o
no
b
j
e
c.
tI
ti
sc
r
e
a
t
e
d
byu
s
i
n
gt
h
ef
u
n
c
t
i
o
nu
i
c
o
n
t
r
o
.
lT
h
i
si
st
h
ef
u
n
c
t
i
o
nt
h
a
tw
i
l
lc
r
e
a
t
et
h
eu
s
e
ri
n
t
c
r
f
a
c
e
e
l
e
m
e
n
t
s
.Th
esamef
u
n
c
t
i
o
ni
su
s
e
dt
op
l
a
c
es
e
v
e
r
a
ld
i
e
r
e
n
tu
s
e
ri
n
t
e
r
f
a
c
ee
l
e
m
e
n
t
s
t
h
r
o
u
g
ht
h
es
e
l
e
c
t
i
o
no
ft
h
es
t
y
l
ep
r
o
p
e
r
t
y
.I
nt
h
eexampleb
e
l
o
w
.t
h
e
s
ei
n
c
l
u
d
el
a
b
e
l
b
o
x
c
sandu
s
e
re
d
i
t
a
b
l
et
e
x
lb
o
x
e
s
.S
e
v
e
r
a
lp
a
r
a
m
e
t
e
r
si
d
e
n
t
i
f
yL
h
et
y
p
o
fl
e
m
c
n
t
U
E
a
ti
sc
r
e
a
t
e
dandt
h
ep
r
o
p
e
r
t
i
e
sU
E
a
tn
e
e
dt
obeset.Eachp
r
o
p
e
n
yi
ss
e
tt
h
r
o
u
g
ha
p
r
o
p
e
r
t
y
n
a
m
e
l
p
r
o
p
e
r
t
y
v
a
l
u
ep
a
i
r
.HfilinIi
n
c
so
fc
o
d
ea
b
o
v
c t
ot
h
e
h
a
n
d
l
eo
ft
h
ef
i
g
u
r
ewindowi
nwhicht
h
i
se
l
e
m
e
n
twi
Ubep
l
a
c
e
d
.Thet
y
p
eo
fe
l
e
m
e
n
t
t
h
r
o
u
g
h S
t
y
l
ep
r
o
p
e
r
t
ys
n
gname
c
r
e
a
t
e
di
sap
u
s
h
b
u
t
t
o
ne
l
e
m
e
n
tc
s
t
a
b
l
i
s
h

p
u
s
h
.
"TheU
n
i
t
su
s
e
dt
os
i
z
eandp
o
s
i
t
i
o
nt
h
ep
u
s
hb
u
t
t
o
ni
si
n
c
h
e
s
.TheP
o
s
i
t
i
o
n
p
r
o
p
e
r
t
yi
sarowv
e
c
t
o
ro
ff
o
v
a
l
u
e
sg
i
v
i
n
gt
h
eLe
f
t
.B
o
t
t
o
m
.W
i
d
t
h
.H
e
i
g
h
to
f
t
h
e
e
l
e
m
e
n
ti
nt
h
eU
n
i
t
ss
c
l
e
c
t
e
de
a
r
l
i
e
r
.Thed
e
f
a
u
l
tl
o
c
a
t
i
o
no
ft
h
eo
r
i
g
i
ni
st
h
eb
o
t
t
o
m
w
indow.TheB
a
c
k
g
r
o
u
n
d
C
o
l
o
rf
o
rt
h
eb
u
t
t
o
ni
ss
c
tt
or
e
d
.Th
c
l
c
f
tcomero
ft
h
ef
i
g
u
l
a
b
e
lt
h
a
twi
l
beont
h
ef
a
c
eo
ft
h
ep
u
s
hb
u
t
t
o
ni
sEXIT.T
h
i
si
st
h
eT
C
X
lp
r
o
p
e
r
t
yo
f
t
h
ep
u
s
h
b
u
t
t
o
no
b
j
e
c.
tTh
cC
a
l
l
Ba
c
kp
r
o
p
e
r
t
yo
ft
h
ep
u
s
hb
u
t
t
o
ni
s es
t
r
i
n
g
c
l
o
s
e
.
"
Thi
ss
t
r
i
n
gw
i
l
lb
et
h
cp
a
r
a
m
e
t
e
ri
nt
h
ee
v
a
lf
u
n
c
t
i
o
na
ti
sc
a
l
l
e
dwhen p
u
s
h
l
c
l
i
c
k
e
v
e
n
ti
st
r
i
g
g
e
d
.MATLAB commandc
l
o
s
ew
i
l
lb
ee
x
e
c
u
t
e
di
nt
h
eCommand
sp s
s
e
d
.T
h
l
sw
i
l
lc
a
u
s
el
h
ef
i
g
u
r
ewindow
w
o
r
k
s
p
a
c
ewhent
h
ep
u
s
hb
u
t
t
o
ni
d
i
s
p
l
a
y
i
n
gt
h
ep
u
s
hb
u
t
t
o
nt
oc
1
o
s
c
.
Onlyafewe
l
e
m
e
n
t
sa
r
eu
s
e
di
ni
ss
c
c
t
i
o
n
.TheCal
1Backs
t
r
i
n
g
sa
r
ea
l
s
oq
u
i
t
e
s
i
g
h
orw
.B
e
f
o
r
ed
c
v
e
l
o
p
i
n
gt
h
cc
o
d
e
.i
ti
su
s
e
f
u
lt
oo
b
s
e
r
v
e h
l
e
r
a
r
c
h
i
c
a
l
o
r
d
e
ri
nw
h
i
c
ht
h
eg
r
a
p
h
i
c
a
lo
b
j
e
c
t
si
nMATLADa
r
eo
r
g
a
n
i
z
e
d(
r
e
p
r
o
d
u
c
e
dfrome
er
o
o
ti
m
p
l
i
e
st
h
es
t
a
r
t
i
n
gr
e
s
o
u
r
c
ef
o
ra
I
l
MATLABd
o
c
u
m
e
n
t
a
t
i
o
n
)i
nF
i
g
u
r
e2
.
1
3
.Th
g
r
a
p
h
i
cd
i
s
p
l
a
y
si
nM
A
T
L
A
B
.I
t
sh
a
n
d
l
eu
s
u
a
l
l
yh
a
sav
a
l
u
eofO
.Ther
o
o
ta
l
s
os
e
r
v
e
s
sr
e
l
a
t
i
o
n
s
h
i
p
a
sac
o
n
t
a
i
n
e
rf
o
ra
l
lo
ft
h
ef
i
g
u
r
ew
i
n
d
o
w
s
.A
n
o
t
h
e
re
x
p
r
c
s
s
i
o
nf
o
ri
i
st
h
a
taf
i
g
u
r
ewindowi
st
h
ec
h
i
l
do
ft
h
er
o
o
.
tFromt
h
eh
i
e
r
a
r
c
h
yi
ti
se
s
s
e
n
t
i
a
lnote
eu
s
e
ri
n
t
e
r
f
a
c
ee
l
e
m
e
n
t
s
.uiconrolo
ruimenuc
o
n
l
ye
x
i
s
tw
ii
na

R
o
+
o
t

1
U
l
c
o
n
t
r
o
l

Hpush

uicontrol(Hfig_1'Style''push'
'Units'
'inches'

Ax
es

Ulmenu

Image L
l
n
e Patch Surface Text

l
i
g
h
t

F
1g
u
r
e2
.
1
3H
l
e
r
a
r
c
h
i
c
a
ls
l
r
u
c
l
u
r
el
o
rMA
TlABg
r
a
p
h
l
c
a
lo
b
j
e
c
l
s
.

8
4 G
R
A
P
H
I
C
A
LO
P
T
I
MlZAnON

2.
4

f
i
g
u
r
ew
i
n
d
o
w
.8
0a
l
5
0t
h
e e
5o
b
j
e
c
.
tI
nt
h
ef
o
l
l
o
w
i
n
gc
o
d
ee
a
c
hi
n
t
e
r
f
a
c
ee
l
e
m
e
n
t
i
sf
i
r
s
ta
s
s
o
c
i
a
t
e
dw
i
ah
a
n
d
l
e5
0ep
r
o
p
e
r
t
yo
ft
h
ee
l
e
m
e
n
tc
a
nb
e f
e
n
t
o
i
fn
e
c
e
s
s
a
r
y
.T
h
i
si
se
s
p
e
c
i
a
l
l
yt
r
u
eo
ft
h
et
e
x
te
n
t
r
yb
o
x
e
st
oi
n
v
o
k
t
h
eC
a
l
l
8
a
c
k
s
n
g
5
.
2
.
4
.
3 GUICode
l
ef
o
l
l
o
w
i
n
gc
o
d
ewi
l
1g
e
n
e
r
a
t
et
h
eGUI
.T
h
eu
i
w
a
i
t q
u
i
r
e
st
h
a
li
sw
i
n
d
o
wb
e

c
l
o
s
e
db
e
f
o
r
es
u
b
s
e
q
u
e
n
tp
r
o
c
e
s
s
i
n
gc
a
nt
a
k
ep
l
a
c
e(
o
ruiresume b
eu
S
e
di
nE

d
e
).
.
G
l
l

.
.
y
l

GUI for Plotting facility


D . P.VenkataramanApplied Optimization Using MATLAB

%C
h.2 Sec. 2.4

GUI is createdwithin a figure window

-------
H
fig_1 is the handle to the
Hfig_1 "
' figure;

figurewindow

set(Hfig_l'Color'[
0
.
3
0.6
0
.
5
] ... set color

NumberTitle''
o
f
f
' nowindow title

Name' 'Set Range for Design variables'


'Units''inches' inchesused to layout
%o
ther controls
'Menubar'
'none'
'Position'[
4
4
4.4
1
.5]);

Positionpropertyis [leftbottomwidthheight] in
Units selected

Ht_1 uicontrol(Hfig_1

Style'

text'
'Units'
'inches'
'position'[
0
.
1
1
.1
2.4.
2
]
'String''Number of Functions to Plo

);

A
D
O
I
T
I
O
N
A
LM
A
T
L
A
DG
R
A
P
H
I
C
S 8
5

BackgroundColor''white'

String''

'Callback'
'Np str2nUm(get(Ht_2
"String"))i');
HLX1L uicontrol(Hfig_l
'Style''text'
'Units'
'inches'
'Position'[
0
.
1
0.8
0.5.
2
]
'String''x1(min)')i
HT_XIL uicontrol(Hfig_l
'Style'
'edit'
'Units'
'inches'
'Position'[
0
.
7
0.8
0.7.
2
]
'BackgroundColor'
'white'
'String'
"
'Callback'
'xlmin str2num(get(HT XIL
'

String'
)
)i')j

HL_XIM uicontrol(Hfig_1
'Style''text'
'Units'
'inches'
'Position'[
1
.5
0.8
0.6.
2
]
'String''incremen')i
HTX1M uicontrol(Hfig_l
'Style''edit'
'Units'
'inches'
'Position'[
2
.
2
0.8
0.7.
2
]
BackgroundColor''white'

String'

'Callback'
'x1inc str2num(get(HT_X1M
'

String'
)
)j')j
HL_X1U uicontrol(Hfig_1

Style'
text'
'Units'
'inches'
'Position'(
3
.
0
0.8
0.5.
2
]
'String''xl(max)')j

--

u
' n
- 1
qd

1 'l
sL''
H'86

sup

f
e
hU
o
--CI

r
Gn'
t
e
--'
n''n
o''o
c
'' 1
1
es
ulti

yi
s
ztno

4'''

HT_X1U uicontrol(Hfig_l
'Style'
'edit'
'Units'
'inches'
'Position'(
3
.
6
0.8
0.7.
2
]

8
6 G
R
A
P
H
I
C
A
l
O
P
T
I
M
I
Z
A
T
I
O
N

2
.
4 A
D
D
I
T
I
O
N
AlM
A
1
1
A
I
iG
R
A
P
H
I
C
S 8
7

BackgroundColor'
'white'
String'

'
C
a
l
l
b
a
c
l

'x1max str2num(ge(HT_X1U

String'
)
);

);
E

_
X2L

uicontrol(Hfig_l

Style'

t
e
x
t
'
'
U
n
i
t
s
'
'inches'
'Position'[
0
.
1
0
.
5
0.5.
2
)

'string''x2(min)');
HT_X2L =uicontrol(Hfig_l

Style'

edit'
'
u
n
i
t
s
''inches'
'Position'[
0
.
7
0.5
0.7.
2
)
'BackgroundColor''white'

String'

'Callback''x2min str2num(get(HT_X2L
'

String'
)
);

);
HL_X2M uicontrol(Hfi9_1

Style'

t
e
x
t
'
'
U
n
i
t
s
'
'inches'
'Position'[1.5
0.5
0.6.
2
)
'String'
'increment');
HT_X2M uicontrol(Hfig_1

Style'
'
e
d
i
t
'
'Units'
'inches'
'position'[
2
.
2
0.5
0.7.
2
)
'BackgroundColor''white'
'String'
'
'Callback'
'x2inc str2num(get(HT_X2M
'
S
tring'
));');

3
.
6
0.5
0.7.
2
]
'Position'[
'BackgroundColor'
'white'

String'
'

Callbacl

'x2max str2num(get(HT_X2U
'

String'
)
);

);
Hlabel uicontrol(Hfig_1

Style'

text'
'
U
n
i
t
s
''inches'
'position'[
0
.
1
0.1
2.2.
2
1
'BackgroundColor''
y
'
'String''After enteringvalues please press
');
EXT
Hpush uicontrol(Hfig_1
'
S
t
y
l
e
'
'
p
u
s
h
'
'
u
n
i
t
s
'
'inches'
'Position'[
2
.
5
0.1
0.7.
2
1
BackgroundColor'
'
r
e
d
'

String'
'
E
X
I
T
'
'Callback''close');
uiwait
t
a
t
e
m
e
nl
Ss
h
o
u
l
db
ecYt
oi
n
t
e
r
p
r
e
.
tM
ucho
ft
h
ec
o
d
ec
b
e
I
nt
h
ea
b
o
v
ec
o
d
ees
e
d
i
t
a
f
t
e
rc
o
p
ya
n
dp
t
eo
p
e
r
a
t
i
o
n
.Thel
a
s
tcommanduiwait i
n
s
t
r
u
c
l
SM
A
T
L
A
B
t
oh
o
l
do
f
fe
x
e
c
u
t
i
o
nu
n
t
i
lt
h
eu
s
e
rc
l
o
s
e
seGUIw
i
n
d
o
w
.F
i
g
u
2
.
14s
h
o
w
st
h
e
t
a
p
p
r
si
nMATLAB.l
et
e
x
tb
o
x
e
sa
r
ci
n
i
t
i
a
l
i
z
.
e
dw
i
t
hn
ov
a
l
u
e
s
i
m
a
g
eo
ft
h
eGUI i
f
o
r
c
i
n
geu
s
e
rt
os
e
tv
a
l
u
e
se
r
e
b
yc
a
u
s
i
n
ga
l
1o
ft
h
eC
a
l
l
B
a
c
ks
t
r
i
n
g
st
ob
e
e
v
a
l
u
a
t
e
d
.I
na
d
d
i
t
i
o
n
t
hd
a
t
ae
n
t
e
r
e
db
yt
h
eu
s
e
ri
sas
i
n
g
.I
th
a
st
ob
ec
o
n
v
e
r
t
e
dt
o
i
sb
a
s
i
c
.I
td
o
e
sn
o
tv
e
r
i
f
ya
tt
h
eu
s
e
rd
i
de
n
t
e
ran
u
m
b
e
r
.
an
u
m
b
e
r
.T
h
edehe
i
sv
e
r
i
f
i
c
a
t
i
o
nc
a
nb
ed
o
n
eb
ye
x
t
e
n
d
i
n
geC
a
l
l
Ba
c
ks
t
r
i
n
go
v
e
rm
u
l
t
i
p
l
e
S
.T
h
i
si
sl
e
f
t a
ne
x
e
r
c
i
s
ei
np
r
o
g
r
a
m
m
i
n
gf
o
rt
h
eu
s
e
r
.
s
t
a
t
e
m
e
nl
T
h
es
c
r
i
p
tm
f
i
l
eu
s
e
dt
os
e
l
e
c
t dc
r
e
a
t
et
h
ep
l
0
l
Si
sa
sf
o
l
l
o
w
s
.
m.m

HL_X2U uicontrol(Hfig_1
'
S
t
y
l
e
''
t
e
x
t
'
'
U
n
i
t
s
''
i
n
c
h
e
s
'
'Position'[
3
.
0
0.5
0.5.
2
]
'string'
'x2(max)');
HT_X2U uicontrol(Hfig_1
'Style''
e
d
i
t
'
'
i
n
c
h
e
s
'
'
U
n
i
t
s
'

Script file to gowithContourplotting


D
r
. P.VenkataramanAppliedOptimizationUsingMATLAB

E
l1li E

F
l
g
u
r
e2
.
1
4I
m
a
g
eo
fG
U
I
.

2
.
4A
O
D
I
T
I
O
N
A
lMA11ABG
R
A
P
H
I
C
S 89

88 G
R
A
P
H
I
C
AlO
P
T
I
M
I
Z
A
T
lO
N

for Mp10t l:Np


text1
['The function which is being p10tedmust
'nbe a MATLAB'
'nfunction M - Fi1e. Given a meshed matrix input'
'nit must return a Matrix'
'nP1ease se1ect function name in the dia10g box .
and hit returnn '1;
% this prints the text to the screen in the command
windowand serves as a prompt
t
he n is a new 1ine command

Ch. 2Sec 2.4

script M fi1e for 20 Contour P10tting of functions

% invo1ved in optimization

A11 functions are expected to be avai1ab1e as


MATLAB function m-fi1es
F
unctionswil1bese1ectedthroughthe input se1ection
box
Function input parameters wi11 be meshed matrices
b
asedon the range of parameters avai1ab1e from GU

)
fprintf(tex1

The Number of p10ts and range for the p10t


are obtained from running the
user interface script GUI2_4 prior
to the ca11 to P1ot20.m

using the uigetfi1e dia10g box


[fi1epathl
uigetfi1e('c:*.m'
'Fi1es of type
300
300);
MATLAB m-fi1e'

check if fi1e is string


strip the .m extension from the fi1e so it can be
c
a11edby the program
if isstr(file)
'.
m
'");
functname strrep(file
e1se
fprintf('nn')
[ You have chosen CANCEL or the fi1e was
text2
'nnot accepab1e'
'nThe program needs a Fi1e to Continue'
'nP1ease ca11 P1ot20 again and choose a fi1e OR '
'npress the up-arrow button to scro11 through
'nprevious commands nn'

Bye 1'];
error(text2);
end

GU
2 4 call the GUI

XnfoYlnfo are introduced for introducing default


values
Xinfo
[xlmin x1max x1inc1;
the va1ues x1minx1maxx1inc are avai1ab1e in
t
he Command workspace due to Ca11Back

if isempty(Xinfo)
Xinfo
[
4 4 0.051
end
fprintf('n')

defau1t

r
l
1
:

2
11

Yinfo
[x2min x2max x2inc1;
if isempty(Yinfo)
Yinfo
4 4 0
.051
end
fprintf('n')
xvar Xinfo(1):Xinfo(3):xinfo(2); % xl vector
yvar Yinfo(1):Yinfo(3):Yinfo(2); % x2 vector
[
X
1X21= meshgrid(xvar
yvar);
matrixmesh

c1ear text1 text2; c1ears the variab1es text1 .


and text2 for reuse
c1ear Fun maxva1 minva1 strcon conva1ue oneva1ue
1abcont 1abcontU
Fu
n feva1(functname
Xl
X2);
maxva1 max(max(Fun));
minva1 min(min(Fun));
fprintf(
'Thecontourranges fromMIN: 12.3fMAX.
1
2.3f '
minva1
maxva1);
fprintf('n');
strcon input('DO you want to set contour va1ues ?

set defau1t number of p10ts to 1

if isempty(Np)

Np 1
;
end
c1 ; c1ear figure

R
E
F
E
R
E
N
C
E
S 9
1

90 G
R
A
P
H
I
C
A
l
o
p
n
M
I
Z
A
T
I
O
N
[no]:''S');
strconU upper(strcon);
'YES') Istrcmp(strco

'Y')
if strcmp(strconU
fprintf(' Input a vector of contour levels );
fprintf('n')
fprintf('between 10.2f and 10.2f '
minval
maxval);
fprintf('n')
convalue input(
' Inputcontourlevel asaVector :
'
)
;
labcont input('Ooyouwant labelledcontours ?
[no]:'
s
');
labcontU upper(labcont);
if strcmp(labcontU
'YES')Istrcmp(labcontU
'Y')
[
C
h
]
contour(xvar
yvar
Fun
convalue);
h);
clabel(C
else
contour(xvar
yvar
Fun
convalue);
end
else
ncon input('Input number of contours [
2
0
] :');
if isempty(ncon)
ncon 20;
labcont input('Oo you want labelled contours ?
[no]:'

s
')
labcontU upper(labcont);
if strcmp(labcontU'YES')Istrcmp(labcontU'Y')
h
]
contour(xvar
yvar
Fun
ncon);
[
C
clabel(C
h);
else
yvar
Fun
ncon);
contour(xvar
end
elseif ncon
1
onevalue input(
'Input thesinglecontourlevel ');
labcont input(
'00youwant labelledcontours ?
[no]:'

s
')
labcontU upper{labcont);
'YES')Istrcmp(labcontU
'Y')
if strcmp(labcontU
[
C
h
] = contour(xvar
yvar
Fun[onevalue
onevalue]);
clabel(
C
h
);
else
yvar
Fun[onevalue
onevalue]);
contour(xvar
end
else

labcont input(
00youwant labelledcontours ? [...
no]:
'
's');
labcontU upper(labcont);
if strcmp(labcontU
'YES')Istrc
(labcontU

Y
')
[
C
h
)
contour(xvar
yvar
Fun
ncon);
h);
clabel(C
else
contour(xvar
yvar
F l

ncon)
end
end
end
if Np > 1
hold on;
end
Hf gcf;
end
figure(Hf)i
grid
hold off
I
nt
h
ea
b
o
v
et
h
ep
l
o
t
t
i
n
gcommandsh
a
v
eb
e
e
nu
s
e
db
e
f
o
r
e
.ThenewMATLAB

clf
isstr
strrep

max
min
upper
strcmp

a
n
d
commandsa
r
eisempty
fprint.Uset
h
eh
e
l
pcommand1
0l
e
a
r
nmorea
b
o
u
tt
h
e
s
e
.Runt
h
es
c
r
i
p
tf
i
l
e d
u
n
d
e
r
s
t
a
n
dt
h
es
u
e
n
c
eo
fa
c
t
i
o
n
sa
sw
e
l
l ep
r
o
m
p
t
sr
e
g
a
r
d
i
n
gt
h
ec
o
n
t
o
u
r
.I
n
i
se
x
a
m
p
l
et
h
ef
u
n
c
t
i
o
n
s c
r
e
a
l
et
h
ep
l
o
tdon
o
tn
e
e
dg
l
o
b
a
lv
a
l
u
e
sf
o
rc
a
l
c
u
l
a
t
i
o
n
s
f
e

'
e
df
r
o
mecommandw
o
r
k
s
p
a
c
e
.I
no
t
h
e
rw
o
r
d
st
h
ef
u
n
c
t
i
o
n
s a
r
et
ob
e
p
l
o
t
t
e
dc
a
nb
ec
a
l
c
u
l
a
t
e
di
n
d
e
p
e
n
d
e
n
t
l
y
.

REFERENCES
1
.MATLAB The Lan
googe ofT
e
c
l
l
1
li
c
a
lC
o
m
p
l
u
i
n
g
:U
s
i
l
l
g MATLAB
V
e
r
s
i
o
n5

Ma
Wo
s
I
n
c
.1
9
9
8
.
2
.MATLABT
h
eLa
nguageo
f
T
e
c
h
n
i
c
a
lCO
u
l
i
n
g
:U
s
i
n
gMATLABG
r
a
p
h
i
c
s
V
e
r
s
i
o
n5

M
a
t
h
W
o
r
k
sI
n
c
.1
9
9
6
.

1
.0
.
1
0
h
n
8
0
n
M.E
.
a
n
dS
l
e
i
n

M.L
.
G
e
n
e
r
a
l
i
z
e
dS
i
m
u
l
a
t
e
dA
n
n
e
a
l
i
n
g
3
.B
o
h
a
c
h
e
v
s
k
y
T
e
c
l
m
o
m
e
t
r
i
c
Vo
l
.2
8

N
o
.3
1
9
8
6
.
f
o
rF
u
n
c
i
L
o
nO
p
t
i
m
i
z
a
t
i
o
n

J
.S
.
l
n
t
r
o
d
u
c
l
i
o
n10O
p
l
i
l
l
l
u
mD
e
s
i
g
n
McGraw-H

I
i NewYor

k1
9
8
9
.
4
.A
r
o
r
a
5
.Be

'

rF
.
P
.
a
n
d
J

0
s

E
.
R

.
J
r
Y
o
r

k1
9
9
2
.

R
.W.

a
n
dM
c
D
o
n
a
l
d

A
.T
.
l
n
t
r
o
d
u
c
l
i
o
n10F
l
u
i
dM
e
c
h
a
n
i
c
s
4 e
d
.
W
i
l
e
y
New
6
.F
o
x
Y
o
r
k
1
9
9
2
.
7
.A
r
p
a
c
i
V
.S
.
C
o
n
d
u
c
t
i
o
l
lHeatT
r
a
l
l
S
f
e
r

A
d
d
i
s
o
n
W
e
s
l
e
y

R
e
a
d
i
n
g
MA

1
9
6
6
.

92

GRAPHICALOPTIMIZATION

8
.H
a
n
s
e
l
m
a
n
D
.
a
n
dL
i
t
t
l
e
t
i
e
l
d

B
.
M
a
s
t
e
r
i
n
gM
a
t
l
a
b5

AC
o
m
p
r
e
h
e
n
s
i
v
eT
u
t
o
r
i
a
la
n
d
R
e
f
e
r
e
n
c
e
. M
A
'
'
L
ABC
u
r
r
i
c
u
l
u
mS
e
r
i
e
s
P
r
e
n
t
i
c
e
Ha
Jl
E
n
g
l
e
w
o
o
dC
l
i
f
f
s
.NY.
1
9
9
6
.
M
ATLAB.o
n
l
i
n
er
e
f
e
r
e
n
c
ei
np
d
ff
o
r
m
a
t
.Math
W
o
r
k
sI
n
c
.
.1
9
9
8
.
9
.B
u
i
l
d
i
n
gGUIsw
i

PROBLEMS
(Manyo
ft
h
eg
r
a
p
h
i
c
a
1enhancementsa
r
enowp
o
s
s
i
b
l
et
h
r
o
u
g
hp
l
o
t
e
d
i
tf
e
a
t
u
r
e
si
n
l
a
r
r
e
l
e
a
s
e
so
fMA
1LAB.T
hef
o

l
owingp
r
o
b
l
e
m
sa
r
ej
u
s
ts
u
g
g
e
s
t
i
o
n
s
.Th
es
u
g
g
e
s
t
e
d
p
r
o
b
l
e
m
su
s
ea
d
d
i
t
i
o
n
a
lg
r
a
p
h
i
c
a
le
l
e
m
e
n
t
s
.anda
1s
or
u
i
r
ed
e
l
e
t
i
n
gh
a
n
d
l
e
s
.P
l
e
a
s

c
o
n
s
u
l
tMA
T
I
.
.
A
Bd
o
c
u
m
e
n
t
a
t
i
o
n
.
)
2
.
1 P
l
o
tt
h
eba
1l
i
s
a
j
e
c
t
o
r
yoms
i
m
p
l
et
w
o
d
i
m
e
n
s
i
o
na
1m h
a
n
i
c
s
.
2
.
2 P
r
o
d
u
c
a
na
n
i
m
a
t
e
dd
i
s
p
l
a
yo
ft
1
1
j
e
c
t
o
r
y
.

LINEARPROGRAMMING

2
.
3 C
r
e
a
t
eat
w
o
d
i
m
e
n
s
i
o
n
a
lp
r
o
g
r
a
mt
h
a
tw
i
l
ld
i
s
p
l
a
yarandomf
i
r
i
n
gl
o
c
a
t
i
o
n
a
n
d arandom t
a
r
g
e
tl
o
c
a
t
i
o
n
.A
llow t
h
eu
s
e
rt
oc
h
o
o
s
ei
n
i
t
ia
1v
e
l
o
c
i
t
y
e
c
t
o
r
yandc
a
1
c
u
l
a
t
eh
i
se
r
r
o
r
.
m
a
g
n
i
t
u
d
ea
n
da
n
g
l
e
.P
l
o
th
i
st
r

d
i
m
e
n
s
i
o
n
a
lprograma
tw
i
l
ld
i
s
p
l
a
yarandomf
i
r
i
n
gl
o
c
a
t
i
o
n
2
.
4 C
r
e
a
t
eat
h
r
e
a
n
d arandom t
a
r
g
e
tl
o
c
a
t
i
o
n
. Allow eu
s
e
rt
oc
h
o
o
s
ei
n
i
t
ia
1v
e
l
o
c
i
t
y
m
a
g
n
i
t
u
d
eandd
i
r
e
c
t
i
o
na
n
g
l
e
s
.P
l
o
th
i
st
r
a
j
e
c
t
o
r
ya
n
dc
a
l
c
u
l
a
t
eh
i
se
r
r
o
r
.

10wo
v
e
ral
1a
tp
l
a
t
e
.Drawl
i
n
e
s
2
.
5 Drawt
h
eb
o
u
n
d
a
r
yl
a
y
e
rp
r
o
f
i
l
eo
fl
a
m
i
n
a
rl
i
n
d
i
c
a
t
i
n
gv
e
l
o
c
i
t
yp
r
o
f
i
l
ea
tt
e
np
o
i
n
t
si
nt
h
eb
o
u
n
d
a
r
yl
a
y
e
r
.
2
.
6 C
r
e
a
t
eap
r
o
g
r
a
ma
tw
i
l
lp
l
o
t v
e
l
o
c
i
t
yp
r
o
f
i
l
ef
o
ru
s
e
r
s
p
e
c
i
f
i
ci
n
p
u
t
s
.
Al
I
owt
h
eu
s
e
rt
oe
x
p
r
e
s
sh
i
si
n
p
u
t
st
h
r
o
u
g
had
ia
1
0gb
o
x
.

Them
a
j
o
rp
a
r
to
ft
h
i
sbookd
e
a
l
sw
i
t
hm
n
t
h
e
m
n
t
i
c
a
1modelsn
tn
r
ec
h
n
r
n
c
t
e
r
i
z
e
dby
n
o
n
l
i
n
e
a
re
q
u
a
t
i
o
n
s
.I
ikemostof examplesu
s
e
df
o
ri
I
Iu
s
t
r
a
t
i
o
nu
sf
a
r
.The
p
r
e
s
e
n
c
eo
fas
i
n
g
l
en
o
n
l
i
n
e
a
re
q
u
a
t
i
o
ni
nt
h
emodeli
ss
u
f
f
i
c
i
e
n
tt
oi
d
e
n
t
i
f
yt
h
e
p
r
o
b
l
e
mf
r
o
mt
h
ec
l
a
s
so
fn
o
n
l
i
n
e
a
rprogramming(NLP)p
r
o
b
l
e
m
s
.M
a
t
h
e
m
a
t
i
c
a
l
e
rt
e
r
mt
h
a
ti
su
s
e
dt
od
e
s
c
r
i
b
es
u
c
hm
o
d
e
l
sa
n
de
i
rs
o
l
u
t
i
o
n
programmingi
sano
c
h
n
i
q
u
e
s
.Mosto
fe
n
g
i
n
e
r
i
n
gd
e
s
i
g
nf
a
l
l
si
nt
h
i
sc
a
t
e
g
o
r
y
.Havingr
e
p
r
e
s
e
n
t
e
ds
u
c
h
1
1
yi
nt
h
ep
r
e
v
i
o
u
sc
h
a
p
t
e
r
.i
tapp
r
s t
h
eg
r
a
d
i
e
n
ta
n
dt
h

p
r
o
b
l
e
m
sg
r
a
p
h
i
ca
C v
a
r
eo
fn
c
t
i
o
n
si
nt
h
em
a
t
h
e
m
a
t
i
c
a
lmodelhadas
i
g
n
i
f
i
c
a
n
ti
m
p
a
c
ton
i
d
e
n
t
i
f
y
i
n
ges
o
l
u
t
i
o
n
eventhoughnoe
f
f
o
r
twasmadet
op
o
i
n
to
u
ti
sf
e
a
t
u
r
e
i
f
i
c
a
l
l
y
.I
nf
a
c

tg
r
a
p
h
i
c
a
ls
o
l
u
t
i
o
n waso
b
t
a
i
n
e
db
yi
n
s
p
e
c
t
i
o
nr
a
t
h
e
rt
h
a
n
s
p
d
e
t
e
r
m
i
n
e
dbym
a
t
h
e
m
a
t
i
c
a
lr
e
l
a
t
i
o
n
so
re
n
t
i
t
i
e
s
.
T
h
e
r
ei
sa
ne
q
u
a
l
l
yimpo
t
c
Ia
s
so
fp
r
o
b
l
e
m
swhos
m
a
t
h
e
m
a
t
i
c
a
lmodeli
smade
i
n
e
a
r
s
u
c
ha
sExample1
.
3

modeledi
n
upex
c
Iu
s
i
v
e
l
yo
ff
u
n
c
t
i
o
n
st
h
a
ta
r
eo
n
l
yI
Eq
u o
n
s1
.21-1
.
2
4
.Thesameexamplew m
o
d
i
f
i
e
di
nEq
u
a
t
i
o
n
51
.
2
5
1
.
2
85
0 t
ti
sa
p
p
e
n
tom g
r
a
p
h
i
c
a
1d
e
s
c
r
i
p
t
i
o
n
ag
r
a
p
h
i
c
a
ls
o
l
u
t
i
o
nc
o
u
l
db
ed
i
5
c
u
s
S
e
d
.I
.6a
t i
snoc
u
r
v
a
t
u
r
ee
v
i
d
e
n
ti
nef
i
g
u
r
e
.The
s
ep
r
o
b
l
e
m
sa
r
e
i
nF
i
g
u
r
e1
m dl
i
n
e
a
rprogramming (
L
P
)p
r
o
b
l
e
m
s
.Theya
r
en
a
t
u
r
a
li
nes
u
b
j
e
c
to
f
whichcoversav
a
s
tv
a
r
i
e
t
yo
fm
o
d
e
l
su
s
e
df
o
rs
e
v
e
r
a
lk
i
n
d
so
f
o
p
e
r
a
t
i
o
n
sr
e
s
e
a
r
c
h
n
p
l
e1
.
3d
o
e
sr
e
p
r
e
s
e
n
tad
e
c
i
s
i
o
n
m
a
k
i
n
gp
r
o
b
l
e
m
.T
y
p
i
c
a
l
l
y

d
e
c
i
s
i
o
nm
a
k
i
n
g
.Ex
LPc
a
nbeac
o
u
r
s
ebyi
t
s
e
l
f(
o
rs
e
v
e
r
a
lc
o
u
r
s
e
sf
o
r tm
a
t
t
e
r
)
.As
i
g
n
i
f
i
c
a
n
tp
o
r
t
i
o
n
o
fs
u
c
hac
o
w
ouldb
et
od
e
v
e
l
o
pm
a
t
h
e
m
a
t
i
c
a
1modelsfromd
i
f
f
e
r
e
n
ta
r
e
a
so
f
a
st
h
e
r
ei
su
s
ua
1
1
yonen
u
m
e
r
i
c
a
lt
e
c
h
n
i
q
u
ea
ti
scommonlyu
s
e
dt
o
a
p
p
l
i
c
a
t
i
o
n
s
s
o
l
v
eLPp
r
o
b
l
e
m
s
.I
ti
sc
a
l
1edt
h
es
i
m
p
l
e
xmethodandi
sb
a
s
e
dont
h
ea
l
g
o
r
i
t
h
mb
y
g[
1
]
. The method i
n
v
o
l
v
e
sm
o
s
t
l
ye
l
e
m
e
n
t
a
r
y row o
p
e
r
a
t
i
o
n
st
y
p
i
c
a
l
1
y
D i
e
n
c
o
u
n
t
e
r
e
di
nG
a
u
s
s
e
l
i
m
i
n
a
t
i
o
nt
y
p
emo
d
st
h
a
ta
r
ep
a
r
to
ft
h
en
u
m
e
r
i
c
a
l
93

94 L
lN
EARPROGRAMMING

3
.
1

p
p
o
s
i
t
et
y
p
e
andt
r
a
n
s
f
o
r
m
i
n
gt
h
ec
o
n
s
t
r
a
i
n
1
5t
o e
q
u
a
l
i
t
yt
y
p
e
.
f
u
n
c
t
i
o
nt
o o
Thev
i
a
b
l
e
sXJandX2
r
e
p
r
e
s
e
n
t
i
n
gt
h
enumbero
fComponentP
l
a
c
e
m
e
n
tM
a
c
h
i
n
e
s
n
dBr
e
s
p
e
c
t
i
v
e
l
yc
a
nbee
x
p
e
c
t
e
dt
ob
es
e
m
i
p
o
s
i
t
i
v
eP
-0) i
se
i
t
h
e
r
o
fmodelAa
somem
a
c
h
i
n
e
sw
i
l
lbeo
r
d
e
r
e
do
rnom
a
c
h
i
n
e
sw
i
l
lbeo
r
d
e
r
e
d
.Note t
h
ed
i
s

e
t
e
mreo
ft
h
ep
r
o
b
l
e
mi
sb
e
i
n
gi
g
n
o
r
a
s3
.
7
5m
a
c
h
i
n
e
sofTypeAi
sn
o
ts
o
m
e
t
h
i
n
g
t
h
a
twouldbeo
r
d
e
r
e
d
.Th
es
o
l
u
t
i
o
nw
i
l
lb
er
o
u
n
d
e
dt
oas
u
i
t
a
b
l
ei
n
t
e
g
e
r
.
1
te
r
ea
r
e
e
x
t
e
n
s
i
v
ed
i
s
c
u
s
s
i
o
n
si
nd
i
s
c
a
n
di
n
t
e
g
e
rprogrammingl
i
t
e
r
a
t
os
u
g
g
e
s
t
i
smayn
o
lbeeoptimumv
a
l
u
e
.Th
ej
u
s
t
i
f
i
c
a
t
i
o
nf
o
rr
o
u
n
d
i
n
go
f
fi
s i
ti
s
c
o
n
v
e
n
i
e
n
t1
0dos
oa
tt
h
i
sj
u
n
c
t
u
r
e
.
I
nt
h
eo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
nj
(
x
) c
o
e
f
f
i
c
i
e
n
1
5990a
n
d900a
r
ec
a
l
l
e
dt
h
ec
o
s
t
c
i
e
n
1
5
.Ea
c
hc
o
e
f
f
i
c
i
e
n
t i
sa
s
s
o
c
i
a
t
e
dw
i ad
e
s
i
g
nv
a
r
i
a
b
l
e
:
p
r
e
s
e
n
1
5
c
o
e
.
i
n
c
r
e
a
s
e(
d
e
c
a
s
e
)i
nc
o
s
tp
e
ru
n
i
tc
h
a
n
g
ei
nt
h
er
e
l
a
t
e
dv
a
r
i
a
b
l
e
.l
es
i
m
p
l
e
s
twayt
o
l
s
f
o
r
m oe
c
t
i
v
ef
u
n
c
t
i
o
nt
oer
e
q
u
i
r
e
df
o
r
m
a
ti
st
om
u
l
t
i
p
l
ya
l
lo
ft
h
et
e
r
m
s
by1
.T
h
e
r
e
f
o
r
eenewo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
ni
s

3
.
1 PROBLEMDEFINITION

Maximize j
(
x
)
:990XI+900X2+5250

(
3
.
1
)

S
u
b
j
e
c
tt
o
: g
l
(
x
)
:0.
4XI+0
.
6X2S8
.
5

(
3
.
2
)

g2(X):3XJ- X2S2
5

(
3
.
3
)

g
)
(
x
)
:3XJ+6X2S70

(
3.
4
)

M
i
n
i
mize

l
ep
r
o
b
l
e
m
.d
e
f
i
n
e
di
nE
q
t
i
o
n
s(
3
.1
)
(
3
.
4
)i
san
a
t
u
r
a
lr
e
p
r
e
s
e
n
t
a
t
i
o
no
f

g
J
(
x
)
:0.
4xJ+0
.
6~ S8
.
5

m
a
t
h
e
m
a
t
i
c
a
1modeli
ne
n
g
i
n
e
e
r
i
n
gd
e
s
i
g
no
r yo
t
h
e
rd
i
s
c
i
p
l
i
n
e
.Bya
l
l
o
w
i
n
g
u
a
l
i
t
y(
=
)c
o
n
s
t
J
0
1
5 dgr
t
e
ra
no
req
1
(

)c
o
n
s
t
r
a
i
n
1
5t
h
ep
r
o
b
l
e
mc
o
u
l
d
accommodatea
n
ym
a
t
h
e
m
a
t
i
c
a
lmodel a
ti
sc
h
a
r
a
c
t
e
r
i
z
e
dbyl
i
n
e
a
rf
u
n
c
t
i
o
n
s
.T
h
i
s
a
nbec
o
n
s
i
d
e
r
e
dt
h
ei
n
e
q
u
a
l
i
t
yf
o
r
mo
ft
h
em
a
t
h
e
m
a
t
i
c
a
lmodel
r
e
p
r
e
s
e
n
t
a
l
i
o
nc
a
l
t
h
o
u
g
hi
ti
sn
o
te
x
p
s
s
e
di
nes
t
a
n
d
a
r
d
f
o
r
m
a
tf
o
rt
h
eLPp
r
o
b
l
e
m
.

Thes
t
a
n
d
a
r
df
o
r
m
a
toft
h
eLPp
r
o
b
l
e
mi
n
c
l
u
d
e
so
n
l
ye
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
1
5
.I
ti
ss
c
tup
a
s am
i
n
i
m
i
z
a
t
i
o
np
r
o
b
l
e
m (somca
u
t
h
o
r
sp
r
c
f
e
r am
a
x
i
m
i
z
a
t
i
o
np
r
o
b
l
e
m
)
.l
n
a
d
d
i
t
i
o
n
.a
l
lo
ft
h
ev
a
r
i
a
b
l
e
si
nemodel ee
x
p
e
c
t
e
dt
ob
es
e
m
i
p
o
s
i
t
i
v
e
(0
)o
r
h
ec
o
n
s
i
n
tl
i
m
i
1
5eq
u
a
n
t
i
t
yont
h
er
i
g
h
t
h
a
n
ds
i
d
eo
f
n
o
n
n
e
g
a
t
i
v
e[
4
]
.F
i
n
a
l
l
yt

j
(
x
)
:-990XJ- 900X2- 5250

(
3
.
5
)

I
nE
q
u
a
t
i
o
n(
3
.
5
) c
o
n
s
t
a
n
tt
e
r
mont
h
er
i
g
h
t(
5
2
5
0
)c
a
nb
ea
b
s
o
e
di
n
t
ot
h
e
o
p
t
a
l
l
e
f
l
h
a
n
ds
i
d
ebyd
e
f
i
n
i
n
ganewo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
n
L
(
x
)w
i
t
h
o
u
ta
f
f
e
c
t
i
n
ge
1u
e
sf
o
rt
h
ed
e
s
i
g
nv
a
r
i
a
b
l
e
s
.l
tc
a
nbeo
b
s
e
r
v
e
da
ti
n
c
r
e
a
s
i
n
gt
h
ev
a
l
u
e
so
ft
h
e
va
v
a
r
i
a
b
l
e
sw
i
l
lmaket
h
eo
b
j
e
c
t
i
v
emoren
e
g
a
t
i
v
e
whichi
sgooda
swea
r
et
r
y
i
n
gt
o
(
x
)a
slowa
sp
o
s
s
i
b
l
e
.I
fap
a
r
t
i
c
u
l
a
rc
o
s
tc
o
e
f
f
i
c
i
e
n
th
a
dap
o
s
i
t
i
v
es
i
g
n
t
h
e
n
makej
i
n
c
r
e
a
s
i
n
gt
h
eamounto
ft
h
ec
o
r
r
e
s
p
o
n
d
i
n
gv
a
r
i
a
b
l
ewouldl
e
a
dt
oa
ni
n
c
r
e
a
s
eo
ft
h
e
whichi
sn
o
ld
e
s
i
r
a
b
l
ei
f ef
u
n
c
t
i
o
nn
e
e
d
st
obem
i
n
i
m
i
z
e
d
.T
h
is
o
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
n
i
d
e
ai
se
x
p
l
o
i
t
e
di
n eS
i
m
p
l
e
xmethodp
r
e
s
e
n
t
e
dl
a
t
e
r
.
Theq
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
t
sh
a
v
et
obet
r
a
n
s
f
o
r
m
e
dt
oan u
a
l
i
t
yco
a
i
n
t
.The
s
i
m
p
l
e
s
twayachievet
h
i
sc
h
a
n
g
i
s
t
o
i
n
d
u
c
ea
na
d
d
i
t
i
o
n
a
ls
e
m
i
p
o
s
i
t
i
v
ev
a
r
i
a
b
l
e
f
o
re
a
c
hi
n
e
q
ua
1
i
t
yc
o
n
s
t
r
a
i
nt
.T
h
a
lw
i
l
lc
r
e
a
t
eE p
o
n
d
i
n
ge
q
u
a
l
i
t
yc
o
n
sn
t
i
nas
t
r
a
i
g
h
t
f
o
r
w
a
r
dw
a
y
.C
o
n
s
i
d
e
ref
i
r
s
tc
o
n
s
n
t
:

x
J~O; X2~0

ndardFormat
3
.
1
.
1 S

95

ec
o
n
s
t
r
a
i
n

ta
r
er
e
q
u
i
r
e
dt
ob
ep
o
s
i
t
i
v
e(
>0
)
.I
nep
r
o
b
l
e
md
e
f
i
n
e
di
nE
q
u
a
t
i
o
n
s
(
3.
1)
(
3.
4
)
t
h
ec
h
a
n
g
e
sa
tmustbec
a
r
r
i
e
do
u
ti
n
c
l
u
d
ec
o
n
v
e
r
t
i
n
gt
h
eo
b
j
e
c
t
i
v
e

t
e
c
h
n
i
q
u
e
su
s
e
di
nl
i
n
e
a
ra
l
g
e
b
r
a
.I
n
i
sc
h
a
p
t
e
r

o
n
l
yal
i
m
i
l
e
d

b
u
tu
s
e
f
u
ld
i
s
c
u
s
s
i
o
n
i
s
p
s
e
n
t
e
d
.F
i
r
s
t
i
ti
si
m
p
o
r
t
a
n
tt
ou
n
d
e
r
s
t
a
n
dLPp
r
o
b
l
e
m
s m
o
d
e
l
i
n
gi
s
s
u
e
sa
n
d
t
h
e
i
rs
o
l
u
t
i
o
n
a
se
ya
r
ed
i
f
f
e
r
e
n
tfrommosto
feo
t
h
e
rp
r
o
b
l
e
m
si
nt
h
i
sb
o
o
k
.
S ond
m Yo
f c
u
r
r
e
n
tn
u
m
e
r
i
c
a
lt
e
c
h
n
i
q
u
e
sf
o
rNLPp
r
o
b
l
e
m
so
b
t
a
i
nt
h
e
i
r
.The
s
el
i
n
e
a
r
i
z
e
d
s
o
l
u
t
i
o
nbyl
i
n
e
a
r
i
z
i
n
gt
h
es
o
l
u
t
i
o
na
tt
h
ec
u
r
r
e
n
td
e
s
i
g
np
o
i
nt
u
a
t
i
o
n
s nb
es
o
l
v
e
dbyemeth
So
ft
h
i
sc
h
a
p
t
e
r
.l
e a
d
e
ri
sd
i
r
t
e
dt
o y
o
ft
h
eb
o
o
k
sonLPf
o
ramored
e
t
a
i
l
e
dd
e
s
c
r
i
p
t
i
o
no
ft
h
em
o
d
e
l
sa
n
dt
h
et
e
c
h
n
i
q
u
e
s
s
e
n
t
a
t
i
o
ni
ni
sbooki
ss
i
m
p
l
ea
n
db
r
i
e
f
.R
e
f
e
r
e
n
c
e
s1
3a
r
eu
s
e
f
u
lf
o
rt
h
i
s
a
st
h
ep
o
s
e
.
pu

Example1
.
3w
i
l
lbeu
s
e
dt
od
e
f
i
n
et
h
et
e
r
m
i
n
o
l
o
g
ya
s
s
o
c
i
a
t
e
dw
i
L
Pproblem. l
e
m
o
d
i
f
i
e
dv
e
r
s
i
o
Do
ft
h
ep
r
o
b
l
e
mi
su
s
e
d1
0d
e
v
e
l
o
pef
o
r
m
a
to
ft
h
eLPp
r
o
b
l
e
ms
o
a
twec
a
nf
o
l
l
o
wt
h
r
o
u
g
hw
i
t
heg
r
a
p
h
i
c
a
ls
o
l
u
t
i
o
n
.Them
a
t
h
e
m
a
t
i
c
a
lmodelh
a
s
twod
e
s
i
g
nv
a
r
i
a
b
l
e
sxJ
X2 t
h
a
lr
e
p
r
e
s
e
n
tenumbero
fComponentP
l
a
c
e
m
e
n
t
.r
e
s
p
e
c
t
i
v
e
l
y
.Theo
b
j
e
c
t
i
v
ei
s1
0maximizet
h
enumbero
f
M
a
c
h
i
n
e
so
ft
y
p
eAandB
b
o
a
r
d
st
obem
a
n
u
f
a
c
t
u
r
e
d
.C
o
n
s
t
r
a
i
n
tg
lr
e
p
r
t
sea
c
q
u
i
s
i
t
i
o
ndo
J
]a
r
sa
v
a
i
l
a
b
l
e
.
C
o
n
s
t
r
a
i
n
tg2 r
e
p
r
e
s
e
n
1
5ef
l
o
o
rs
p
a
c
ec
o
n
s
t
r
a
i
n.
tC
o
n
s
t
r
a
i
n
tg3 r
e
p
r
e
s
e
n
1
5e
numbcro
fo
p
e
r
a
t
o
r
sa
v
a
i
l
a
b
l
c
.

PR08LEMD
E
F
I
N
I
T
I
O
N

(
3
.
2
)

i
sc
a
nbet
r
a
n
s
f
o
r
m
e
dt
o

4x
0
.
6~ +x
.
5
g
J
(
x
)
:0.
J+
3 8
}

(
3
.
6
)

Th
enewv
a
r
i
a
b
l
eX3 i
sr
e
f
e
r
r
e
dt
oa
st
h
s
l
a
c
kv
a
r
i
a
b
l
e
.Th
ed
e
f
i
n
i
t
i
o
ni
sq
u
i
l
e
a
p
p
r
o
p
r
i
a
t
eb
e
c
a
u
s
ei
tt
a
k
e
supt
h
es
l
a
c
ko
ft
h
eo
r
i
g
i
n
a
lc
o
n
s
a
i
n
tt
e
r
m
swheni
ti
sl
e
s
s
8
.
5
.
1
t
es
l
a
c
kwi
1
lbep
o
s
i
t
i
v
ea
sef
r
r
s
ttwot
e
r
m
sw
i
l
lbel
e
s
s o
r u
a
lt
o
r
i
g
i
n
a
ld
e
f
i
n
i
t
i
o
ni
nE
q
u
a
t
i
o
n(
3
.
2
)
.I
no
t
h
e
r
8
.
5(
f
o
rc
o
n
s
t
r
a
i
n
ts
a
t
i
s
f
a
c
t
i
o
n
)f
r
o
meo
i
ti
st
h
ed
i
f
f
e
r
e
n
c
eb
e
t
w
e
c
n c
o
n
s
t
r
a
i
n
tl
i
m
i
t(
8
.
5
)a
n
d f
i
r
s
ttwot
e
r
m
so
f
words
t
h
ec
o
n
s
t
r
a
i
n
.
tI
ft
b
ev
a
l
u
e
so
fXIa
n
dX2c
a
u
s
et
h
ec
o
n
s
t
r
a
i
n
tt
obea
ti
1
5l
i
m
i
tv
a
l
u
e
.
e
nX3w
i
l
lb
ez
e
r
o
.T
h
i
si
st
h
er
e
a
s
o
ni
ti
sd
e
t
i
n
e
dt
obes
e
m
i
p
o
s
i
t
i
v
e
.Byd
e
f
i
n
i
t
i
o
n
.

96

3
.
1 PR08LEMDEANmON 97

UNEARPROGRAMMING

t
h
e
r
e
f
o
r
e

X
3i
ss
i
m
i
l
a
rt
ot
h
eo
r
i
g
i
na
Jv
a
r
i
a
b
l
e
si
nt
h
eLPp
r
o
b
l
e
m
.T
r
a
n
s
f
o
r
m
i
n
gt
h
e

t
h
ec
o
m
p
l
e
t
ee
q
u
a
t
i
o
n
si
ns
t
a
n
d
a
r
df
o
r
m
a
t
r
e
m
a
i
n
i
n
gc
o
n
s
t
r
a
i
n
t
si
nas
i
m
i
l
a
rm
a
n
n
e
r
a
r
e

4x
+0
.
6X
2+X
J=8
.
5
8
'
(
X
)
:0.
8
2
(
X
)
:3X
-X2+X4=2
5
+6X
2+X
s=7
0
8
3
(
X
)
:3x

(
3
.
5
)

o
:
S
u
b
j tt

(
3
.
6
)

~~0
;

x3~ 0
; x4~ 0
; Xs~ 0

M
i
n
i
m
i
z
e j
(
x
)
:-990x 9
0
0X
2-5
2
5
0

x~ 0
;

w
h
e
r
ea
l
l=0
.
4
.a12=0
.
6
.al3=1
.B
e
c
a
u
s
et
h
e
r
ea
r
ef
i
v
ed
e
s
i
g
nv
a
r
i
a
b
l
e
si
ne
d
d
i
t
i
o
n
a
lt
e
r
m
sw
i f
o
l
l
o
w
i
n
gc
o
e
f
f
i
c
i
e
n
t
sa
'
4(
=0
)a
n
d
p
r
o
b
l
e
md
e
f
i
n
i
t
i
o
na
al
5
(
=0
)
.a
r
ei
n
c
l
u
d
e
di
nee
x
p
r
e
s
s
i
o
nf
o
rg(
x
)w
i
t
h
o
u
tc
h
a
n
g
i
n
gi
t
sm
e
a
n
i
n
g
:

(
3
.
7
)
(
3
.
8
)

g(
x
)
:alX +a
'
2x
'
3x
'
4x
4+asx
.
5
s 8
2+a
3+a
l
ec
o
n
s
t
r
a
i
n
tg
2
(
X
)c
a
nb
es
i
m
i
1
a
r
l
ye
x
p
s
s
e
da
s

g
2
(
X
)
:

z
Gx
+au +a2
X
s=2
5
4x
4+s
3x
3+a2
Th
ec
o
e
c
i
e
n
t
si
nt
h
ea
b
o
v
ee
q
u
a
t
i
o
nc
a
nb
e s
i
l
ye
s
t
a
b
l
i
s
h
e
db
yc
o
m
p
a
r
i
s
o
nw
i
t
h

E
q
u
a
t
i
o
n(
3
.
7
)
.Le
t
t
i
n
ger
i
g
h
t
h
a
n
ds
i
d
ev
a
Ju
e
so
ft
h
ec
o
n
s
t
r
a
i
n
t
sb
er
e
p
r
e
s
e
n
t
e
db
y
b"b2 bm e s
t
r
df
o
r
m
a
ti
nE
q
u
a
t
i
o
n
s(
3
.
5
)
(
3
.
9
)c
b
e s
l
a
t
e
dt
o

(
3
.
9
)
M
i
n
i
m
i
z
e

j
(
x
)
:c
T
x

(
3
.
1
0
)

S
u
b
j 0
:

g
(
x
)
:Ax=b

(
3
.
1
1
)

Equ
a
t
i
o
n
s(
3
.
5
)(
3
.
9
)e
x
p
r
l

t
h
eLPp
r
o
b
l
e
mi
ns
t
a
n
d
a
r
df
o
r
m
a
t
.Ani
n
c
o
n
s
i
s
t
e
n
c
y

MIN

p
t

r -EE E -E E -EEEa--

l
J
l
l
l

c
c
.c

-B
Elli--J
12m

U U . L U

r'EBEEE'EEEEEEaEEE

hu

EEEEEEEE'EaEEJ

g(
x
)
:a
" x +a
'
2x
'
3x
.
5
2+a
3=8

.
h

aM
aa

-W

C
o
m
p
a
r
i
n
gE
q
u
a
t
i
o
n
s(
3
.
1
0
)a
n
d(
3
.
1
1
)w
i
t
hE
q
u
a
t
i
o
n
s(
3
.
5
) (
3
.
8
)ev
e
c
t
o
r
sa
n
d
m
a
t
r
i
xa
r
e

(
3
.
5
)

t
t
h
ec
o
s
t
Th
f
o
l
l
o
w
i
n
gd
e
f
i
n
i
t
i
o
n
sa
r
cu
s
e
dt
or
e
w
r
i
t

t
I
l
ep
r
o
b
l
e
mu
s
i
n
gm
a
l
r
i
c
e
s
.Le
c
o
e
f
f
i
c
i
e
n
t
sb
er
e
p
r

e
n
t
e
db
yc
C2
.
c
"
.Thec
o
e
f
f
i
c
i
e
n
t
si
nec
o
n
s
t
r
a
i
n
t
sa
r
e
1
a
i
j(
u
s
i
n
gd
o
u
b
l
es
u
b
s
c
r
i
p
t
s
)
.l
ef
i
r
s
ts
u
b
s
c
r
i
p
ti
d
e
n
t
i
r
e
p
r
e
s
e
n
t
e
db
yesym
c
o
n
s
t
r
a
i
n
tn
u
m
b
e
rwh
i
1
et
h
es
e
c
o
n
dc
o
n
s
t
r
a
i
n
ti
st
h
es
a
m
ea
st
h
ev
a
r
i
a
b
l
ei
ti
s
m
u
l
t
i
p
l
y
i
n
g
.I
ti
sa
Js
ou
s
e
f
u
1t
oi
d
e
n
t
i
f
yr
s
ts
u
b
s
c
r
i
p
tw
i
t
ht
h
er
o
wn
u
m
b
e
ra
n
d
t
h
es
e
c
o
n
ds
u
b
s
c
r
i
p
tw
i
t
h
ec
o
l
u
m
nn
u
m
b
e
rt
or
r
e
s
p
o
n
dw
i
o
p
e
m
t
i
o
n
s tu
s
e
i
I
n
e
a
ra
l
g
e
b
m
.F
o
re
x
a
m
p
l
e
t
h
e a
i
n
t8(
x
)wi
l
1b
er
e
p
s
e
n
t
e
d

1h

(
x
)+5
2
5
0=-990x 9
0
0~
L
(
x
)
:j

(
3
.
1
2
)

xr
e
p
r
e
s
e
n
t
st
h
ec
o
l
u
m
nv
e
c
t
o
ro
fd
e
s
i
g
nv
a
r
i
a
b
l
e
s
i
n
c
l
u
d
i
n
ges
l
a
c
kv
a
r
i
a
b
l
e
s
l
x
"
X
2

.
.
.x]Tmentst
h
ea
n
s
p
o
s
i
t
i
o
ns
y
m
b
o
l
)
.c p
r
e
s
e
n
t
st
h
ec
o
l
u
m
nv
e
c
t
o
r
o
fc
o
s
tc
o
e
f
f
i
c
i
e
n
t
s
.[
c
"C2.
.
C
J
T
.br
e
p
r
e
s
e
n
t
sec
o
l
u
m
nv
e
c
t
o
ro
ft
h
ec
o
n
s
t
r
a
i
n
t
i
I
m
i
t
s
.[
b
"b2
b
m
]
T
.Ar
e
p
r
e
s
e
n
t
st
h
emxnm
a
t
r
i
xo
fc
o
n
s
t
r
a
i
n
tc
o
e
f
f
i
c
i
e
n
t
s
.I
n
i
sb
o
o
ka
l
lv
e
c
t
o
r
s c
o
l
u
m
nv
e
c
t
o
r
su
n
l
e
s
so
I
t
l
e
r
w
i
s
ei
d
e
n
t
i
f
i
e
d
.
12m
a
aaj
rEEEEEEEEEEEEEEEEL

M
i
n
i
m
i
z
e

i
n
t
s
:
"~ 0
S
i
d
en a

r
e
s
p t
t
oef
o
r
m
a
to
ft
h
ep
r
o
b
l
e
m
sp
r
e
v
i
o
u
s
l
yd
e
f
i
n
e
di
nt
h
e
c
a
nb
eo
b
s
e
r
v
e
dw
i
b
o
o
k
.Namely

8
(
"
)w
a
sr
e
s
e
r
v
e
d

d
e
f
i
n
ei
n
e
q
u
a
J
i
t
yc
o
n
si
n
t
sw
h
i
l
eh(

)d
e
f
i
n
e
d
u
a
Ji
t
yma
i
n
t
s
.H
e
r
e

8
(
X
)i
sa
J
s
ob
e
i
n
gu
s
e
df
o
re
q
ua
J
i
t
yn
s
a
i
n
t
s
.T
h
i
sc
h
a
n
g
e
sc
h
a
p
t
e
ro
n
l
y
.I
ns
u
b
s
e
q
u
e
n
tc
h
a
p
t
e
r
st
h
ep
r
i
o
rm
e
a
n
i
n
gf
o
r
i
sa
c
c
o
n
u
n
o
d
a
t
e
di
ni
s
y
m
b
o
l
sa
r
en
o
tr
e
q
u
i
r
e
d ec
o
n
s
t
r
a
i
n
tu
n
d
e
r
t
h
es
y
m
b
o
l
si
sr
e
s
t
o
r
e
d
.A
c
t
u
a
l
l
y
s
i
o
n na
1
s
ob
e r
r
e
d1
0b
ye
q
u
a
t
i
o
nn
u
m
b
e
r
.Thef
o
r
m
u
1
a
t
i
o
ni
nE
q
u
a
t
i
o
n
s
d
i
s
c
5)
(
3
.
9
)c
a
nb
ee
x
p
s
s
e
ds
u
c
c
i
n
c
t
1
yb
yr
e
s
o
r
t
i
n
gt
oam
a
t
r
i
xn
o
t
a
t
i
o
n
.

l
i
s
(
3.
c
o
n
v
e
r
s
i
o
ni
sq
u
i
t
es
t
r
a
i
g
h
t
f
o
r
w
a
r
d
.B
e
f
o
r
ee
m
b
a
r
k
i
n
go
ni
st
r
a
n
s
f
o
r
m
a
t
i
o
n
t
h
e
d
a
g
a
i
n
.Thec
o
n
s
t
a
n
tt
e
r
m
r
e
p
r
e
s
e
n
t
a
t
i
o
no
ft
h
eo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
nn
e
e
d
st
ob
ec
h
a
n
g
'
i
sb
r
o
u
g
h
tt
ot
h
er
i
g
h
l
h
a
n
ds
i
d
es
o o
b
j t
i
v
ef
u
n
c
t
i
o
n nb
ee
x
p
r
e
s
s
w
i

t
e
r
m
so
n
l
yi
n
v
o
l
v
i
n
ged
e
s
i
g
nv
a
r
i
a
b
l
e
so
nt
h
er
i
g
h
t
.T
h
i
si
ss
h
o
w
na
s
L
(
x
)b
e
l
o
w
.
Oncea
g
a
i
nf
o
rc
o
n
v
e
n
i
e
n
c
et
h
es
a
m
es
y
m
b
o
lj
(
x
)i
su
s
e
di
np
l
a
c
eo
ft
h
i
snewf
u
n
c
t
i
o
n
i
ns
u
b
s
e
q
u
e
n
td
i
s
c
u
s
s
i
o
n
s
.T
h
e
s
ee
x
t
r
as
y
m
b
o
l
sh
a
v
eb
e
e
nt
h
r
o
w
ni
nt
ok
e
e
pt
h
e
.T
h
i
si
sas
m
a
l
lp
r
i
c
ef
o
rt
h
ea
b
i
l
i
t
y1
0d
e
v
e
l
o
per
n
a
t
h
e
m
a
t
icaJ
d
e
r
i
v
a
t
i
o
nd
i
s
t
i
n
ct
m
o
d
e
li
nan
a
t
u
m
lwaya
n
d c
o
n
s
t
r
u
c
ti
tt
oc
o
n
f
o
r
mt
ot
h
es
t
a
n
d
a
r
dr
e
p
s
e
n
t
a
t
i
o
ni
n
m
o
s
tr
e
f
e
r
e
n
c
e
s
.T
h
e
r
e
f
o
r
e
:

c=[
9
9
09
0
000O
]
T
b=[
8
.
52
57
0
]
T

A
z
h
?
i:
!

Thef
o
U
o
w
i
n
ga
d
d
i
t
i
o
na
Jd
e
f
i
n
i
t
i
o
n
sh
a
v
eb
e
e
nu
s
e
di
nt
h
es
t
a
n
d
a
r
df
o
r
m
a
t
.ni
sE
n
u
m
b
e
ro
fd
e
s
i
g
nv
a
r
i
a
b
l
e
s dmisen
u
m
b
e
ro
fc
o
n
s
t
r
a
i
n
t
s
.I
nt
l
t
ee
x
a
m
p
l
e

n=5
.
a
n
dm=3
.N
o
l
eI
t
l
a
tt
h
es
l
a
c
kv
a
r
i
a
b
l
e
sa
r
ei
n
c
l
u
d
e
di
nt
h
ec
o
u
n
tf
o
rn

98

LINEAAPAOGRAMMING

3
.
1 PAOBLEMOEFINJTION

NegatlveValuesofDesignV
a
r
l
a
b
l
e
s
:T
heLPs
t
a
n
d
a
r
dm
a
t
h
e
m
a
t
i
c
a
1m
o
d
e
l
a
l
l
o
w
so
n
l
yf
o
rn
o
n
n
e
g
a
t
i
v
ed
e
s
i
g
nv
a
r
i
a
b
l
e
s
.T
h
i
sd
o
e
sn
o
tl
i
m
i
tea
p
p
l
i
c
a
t
i
o
nofLP
i
na
n
yw
a
y
.The
r
ei
sas
t
a
n
d
a
r
dp
r
o
c
e
d
u
r
et
oh
a
n
d
l
en
e
g
a
t
i
v
ev
a
l
u
e
sf
o
rv
a
r
i
a
b
l
e
s
.I
fa
d
e
s
i
g
nv
a
r
i
a
b
l
e

s
a
yx
.
.i
sc
x
p
e
c
t
e
dt
ob
eu
n
r
e
s
l
r
Ic
t
e
di
ns
i
g
n
e
ni
ti
sd
e
f
i
n
e
da
se
d
i
f
f
e
r
e
n
c
o
ft
w
on
o
n
n
e
g
a
t
i
v
ev
a
r
i
a
b
l
c
s
1 .
.
1
x
.
=
x
;-^
x
I-^J
J

(
3
.
1
3
)

w
h
e
r
ex~ a
n
dx~ a
r
ea
c
c
e
p
t
a
b
l
eLPv
a
r
i
a
b
l
e
s
.I
ft
h
el
a
t
t
e
ri
sh
i
g
h
e
rt
h
a
nt
h
ef
o
r
m
e
r

x
J
q
u
a
t
i
o
n(
3
.
1
3
)i
su
s
e
dt
or
e
p
l
a
c
ex i
nt
h
em
a
t
h
c
a
l
c
a
nh
a
v
en
e
g
a
t
i
v
ev
a
l
u
e
s
.E
.
lN
o
t
et
h
a
ti
sp
r
o
c
e
d
u
r
ca
l
s
oi
n
c
r
e
a
s
e
st
h
cn
u
m
b
e
ro
fd
e
s
i
g
nv
a
r
i
a
b
l

i
nt
h
e
mode
mode
.
l

TypeofC
o
n
s
t
r
a
i
n
t
s
: Th
el
e
s
s a
no
re
q
u
a
lt
o(
:
:
:
:
)c
o
n
s
t
r
a
i
n
tw
a
sh
a
n
d
l
e
d
l
.
l
es
t
a
n
d
a
r
dn
u
m
e
r
i
c
a
l
n
a
t
u
r
a
l
l
yi
nd
e
v
e
l
o
p
i
n
gt
h
eLPs
t
a
n
d
a
r
dm
a
t
h
e
m
a
t
i
c
a
lmode
t
e
c
h
n
i
q
u
ef
o
ro
b
t
a
i
n
i
n
gt
h
es
o
l
u
t
i
o
nr
e
q
u
i
r
e
st
h
a
tes
l
a
c
kv
a
r
i
a
b
l
e
sb
ep
o
s
iv
e
t
o
s
t
a
r
t
l
s
ob
ep
o
s
i
t
i
v
et
h
r
o
u
g
h
o
u
.
tI
ft
h
e
r
ei
sac
o
n
s
t
r
a
i
n
t i
sr
e
q
u
i
r
e
dt
ob
e
a
n
dba
m
a
i
n
t
a
i
n
e
da
b
o
v
eas
p
e
c
i
f
i
e
dv
a
l
u
e

f
o
re
x
a
m
p
l
e

x
)
:a
" x +a
'
2x
2+al
=
bJ
g(
3x
3-x4+Xs
'
h
ev
a
r
i
a
b
l
eXsi
sc
a
l
l
e
d

a
r
t
i
f
i
c
i
a
lv
a
r
i
a
b
l
e
.A
r
t
i
f
i
c
i
a
lv
a
r
i
a
b
l
e
s a
l
s
oi
n
t
r
o
d
u
c
e
d
f
o
r c
he
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
t i
sn
a
t
u
r
a
l
l
yp
r
e
s
e
n
ti
nt
h
emode
l
.T
heu
s
eo
ft
h
e
s
e
v
a
r
i
a
b
l
e
swi
l
1b
ei
l
l
u
s
t
r
a
t
e
dl
a
t
e
rt
h
r
o
u
g
ha
nx
a
m
p
l
e
.

3
.
1
.
2 ModellngIssues
a
t
h
e
m
a
t
i
c
a
lm
o
d
e
ld
e
f
i
n
di
nE
q
u
a
t
i
o
n
s
TheLPp
r
o
g
r
a
mi
sc
h
a
r
a
c
t
e
r
i
z
e
db
yem
(
3.
10
)(
3
.
1
2
)
.Duet
ot
h
el
i
n
e
a
rn
a
t
u
r
eo
fa
1
1
o
ft
h
ef
u
n
c
t
i
o
n
sinv~lved t
h
ed
i
s
c
u
s
s
i
o
n
o
ft
h
es
o
l
u
t
i
o
nn
e
e
do
n
l
yc
o
n
s
i
d
e
r c
o
e
f
f
i
c
i
e
n
t
so
ft
h
ef
u
n
c
t
i
o
n
s
t
h
ev
a
r
i
a
b
l

LU

Wh
i
l
ei
ti
sp
o
s
s
i
b
l
et
oc
h
g
ees
i
g
nonX4b
ym
u
l
t
i
p
l
y
i
n
gw
i
t
h-1r
o
u
g
h
o
u
t

i
s
c
a
u
s
e
ser
i
g
h
t
h
a
n
dc
o
n
s
t
r
a
i
n
tl
i
m
i
tt
ob
en
e
g
a
t
i
v
e

w
h
i
c
hs
h
o
u
l
db
ea
v
o
i
d
e
d
.I
ns
u
c
h
c
a
s
e
s
a
n
o
t
h
e
rv
a
r
i
a
b
l
e

x
s
i
si
n
c
l
u
d
e
di
nt
h
ee
x
p
s
s
i
o
n
.I
nt
h
i
sc
a
s
e
t
w
oa
d
d
i
t
i
o
n
a
l
v
a
r
i
a
b
l
e
sa
r
enowp
a
r
to
ft
h
em
o
d
e
l
:

r
EE -L

g(
x
)
:al
'
3x3-x4bJ
l
2x
2+a
lX +a

A
-

t
h
e
ni
ta
p
p
e
s
n
a
t
u
r
a
lt
oi
n
t
r
o
d
u
c
ean
e
g
a
t
i
v
es
l
a
c
kv
a
r
i
a
b
l
e-X4t
oc
h
a
n
g
eEqu
o
n
(
3
.
1
4
)t
oa
ne
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
t
:

(
3
.
1
4
)

F
i
g
u
r
e3
.
1i
l
l
u
s
t
r
a
t
e
st
h
eg
r
a
p
h
i
c
a
ls
o
l
u
t
i
o
nt
ot
h
ep
r
o
b
l
e
m
.Thes
o
l
u
t
i
o
ni
sa
tx =
0
.
5
a
n
dX2=1
.5
.T
h
i
si
si
n
d
e
p
e
n
d
e
n
to
ft
h
eo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
n(
w
h
a
t
e
v
e
ri
tmayb
e
)
.l
n
m
a
t
r
i
xf
o
r
m
.Aa
n
dba
r
e

-lE

-EAt-

g
l
: xl+X
z=2
g
2
: -x1+
X
z=1

t
h
e
m
s
e
l
v
e
sc
a
nb
ei
g
n
o
r
e
d
.T
h
e
s
ec
o
e
f
f
i
c
i
e
n
t
sw
h
i
c
ha
r
ep
r
e
s
e
n
ta
sv
e
c
t
o
r
so
ra
ma
x

c
a
nb
em
a
n
i
p
u
l
a
t
e
dt
oy
i
e
l
ds
o
l
u
t
i
o
n
su
s
i
n
gc
o
n
c
e
p
t
sf
r
o
ml
i
n
e
a
ra
l
g
e
b
r
a
.
Al
o
o
ka
tt
h
eg
r
a
p
h
i
c
a
ls
o
l
u
t
i
o
nt
ot
h
eLPp
r
o
b
l
e
ms
u
g
g
e
s
t
st
h
a
tt
h
es
o
l
u
t
i
o
nt
ot
h
e
p
r
o
b
l
e
m

i
fi
te
x
i
s
t
s

m
u
s
tl
i
eo
nt
h
ec
o
n
s
t
r
a
i
n
tb
o
u
n
d
a
r
yb
e
c
a
u
s
et
h
ec
u
r
v
a
t
u
r
eo
ft
h
e
e
c
t
i
v
ef
u
n
c
t
i
o
ni
sz
e
r
oe
v
e
r
y
w
h
e
r
e(
i
ti
sl
i
n
e
a
r
)
.F
u
r
t
h
e
r
m
o
r
e

t
h
es
o
l
u
t
i
o
nt
o LP
o
p
r
o
b
l
e
mi
sp
r
i
m
a
r
i
l
yi
n
f
l
u
e
n
c
e
db
yE
q
u
a
t
i
o
n(
3.
11
)
.B
e
i
n
gai
In
e
a
re
q
u
a
t
i
o
n
.t
o
p
i
c
s
f
r
o
ml
i
n
e
a
ra
l
g
e
b
r
aa
r
en
e
c
e
s
s
a
r
yt
ou
n
d
e
r
s
t
ds
omeo
fei
m
p
l
i
c
a
t
i
o
n
so
fe
m
a
t
h
e
m
a
t
i
c
a
lm
o
d
e
l
.
I
fn=m

a
ti
s

t
h
en
u
m
b
e
ro
fu
n
k
n
o
w
n
si
st
h
es
a
m
ea
st
h
en
u
m
b
e
ro
fv
a
r
i
a
b
l
e
s
.
t
h
e
r
ec
a
nb
enos
c
o
p
ef
o
ro
p
t
i
m
i
z
a
t
i
o
na
sE
q
u
a
t
i
o
n(
3
.
1
1
)w
i
l
ld
e
t
e
r
m
i
n
e s
o
l
u
t
i
o
n
.
p
r
o
v
i
d
e
do
n
ee
x
i
s
t
s
'
.I
nt
h
ef
o
l
1o
w
i
n
g
.manyo
f c
o
n
c
e
p
t
sa
r
ec
x
p
l
a
i
n
e
dt
h
r
o
u
g
h
o
rb
yc
o
n
s
i
d
e
r
a
t
i
o
no
fs
i
m
p
l
ee
x
a
m
p
l
e
s
.T
h
e
ym
o
s
t
l
y
s
i
m
p
l
eg
r
a
p
h
i
c
a
li
l
l
u
s
t
r
a
t
i
o
n
x
"X2).F
o
re
x
a
m
p
l
e

c
o
n
s
i
d
r
t
w
o
c
o
n
s
a
i
n
t
s
d
e
a
lw
i
t
he
x
a
m
p
l
e
so
ft
w
ov
a
r
i
a
b
l
e
s(

g(x):a
" x +aJ
z+a'3x3~ b
2x

99

l
ee
x
i
s
t
e
n
c
o
fas
o
l
u
t
i
o
nd
e
p
e
n
d
so
nt
h
er
o
w
so
fA(
o
rc
o
l
u
m
n
so
fA-wew
i
l
ld
e
a
l

i
n
e
a
r
l
yi
n
d
e
p
e
n
d
e
n
t

t
h
e
nt
h
e
r
e
w
i
t
hr
o
w
so
n
l
yf
o
r
c
o
n
v
e
n
i
e
n
c
e
)
.I
f
t
h
er
o
w
so
fAa
r
el
i
sau
n
i
q
u
es
o
l
u
t
i
o
nt
oes
y
s
t
e
mo
fe
q
u
a
t
i
o
n
s
.
h
ea
b
o
v
ee
x
a
m
p
l
e
)
T
h
D
e
f
i
n
e
a
n
o
t
h
e
rp
r
o
b
l
e
ma
sf
o
l
l
o
w
s(
o
n
l
yc
h
a
n
g
i
n
gb omt
enewp
r
o
b
l
e
mi
s

g:
x
+ =2
g
2
: 2xl+2x
2=4

T
h
enewf
u
n
c
t
i
o
ng
2r
e
p
r
e
s
e
n
t
st
w
i
c
eev
a
l
u
eo
fg
lT
h
i
sm
e
a
n
sg
2c
a
nb
eo
b
t
a
i
n
e
d
m8
J
'Th
e
r
e
f
o
r
e
8
2i
ss
a
i
dt
od
e
p
e
n
don8
1T
e
c
h
n
i
c
a
l
l
y

etwoq
u
a
t
i
o
n
s e
l
i
n
e
a
r

d
e
p
e
n
d
e
n
t
ea
c
d
e
p
e
n
d
e
n
c
eo
ft
h
ef
u
n
c
t
i
o
n
si
sn
o
tac
o
n
c
e
ma
smuch
a
st
h
ef
a
c
te
y
a
r
n
o
tl
i
n
e
a
r
l
yi
n
d
e
p
e
n
d
e
n
.
tB
o
t
h8
1a
n
d8
2e
s
t
a
b
l
i
s
ht
h
es
a
m
el
i
n
e
ti
st
h
es
o
l
u
t
i
o
nt
ot
h
ep
r
o
b
l
e
m
?T
h
e
r
ea
r
ei
n
f
l
n
i
t
es
o
l
u
t
i
o
n
s
a
sl
o
n
g
g
r
a
p
h
i
c
a
l
l
y
.Wha
a
st
h
ep
a
i
ro
f(XlX2) v
a
l
u
e
ss
a
t
i
s
f
i
e
s8
1I
nF
i
g
u
r
e3

1
.a
n
yp
o
i
n
to
nt
h
el
i
n
e
r
e
p
r
e
s
e
n
t
i
n
g8
1i
sas
o
l
u
t
i
o
n
.N
o
to
n
l
y(
0
.
5
1
.5
)
b
u
t(
2

0
)
(
0

2
)a
r
ea
l
s
os
o
l
u
t
i
o
n
s
.
p
t
i
m
i
z
a
t
i
o
np
e
r
s
p
e
c
t
i
v
ei
si
sq
u
i
t
eg
o
o
df
o
ru
n
l
i
k
et
h
ep
r
e
v
i
o
u
sc
a
s
eo
f
From o
u
q
u
es
o
l
u
t
i
o
n

t
h
eo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
nc
a
nb
eu
s
e
dt
of
i
n
dt
h
eb
e
s
ts
o
l
u
t
i
o
n
.1
tw
o
u
l
d

h
o
w
e
v
e
r

i
ft
h
ec
h
o
i
c
e
sw
e
r
ef
i
n
i
t
er
a
t
h
e
rt
h
a
ni
n
f
i
n
i
t
e
.
b
eb
e
t
t
e
r

1
0
0 L
I
N
E
A
RP
R
O
G
R
A
M
M
I
N
G

HM

RM

n
H

auy

RM

e
e

RH

u
u

Cas
ena 2
;ma 3
:N
os
o
l
u
t
l
o
n
;I
n
n
s
l
s
t
e
n
!5
e
!

lo

F
V

g.f-

FM

3
.
1P
R
O
B
L
E
MD
E
F
I
N1
0
N 1
0
1

vb//
KAA

m
d////

2
1

)
(
'
"1

)(~ 1

-3

-2

X
1

F
l
g
u
r
e3
.
1U
n
i
q
u
es
o
l
u
t
i
o
n
U
n
e
a
ri
n
d
e
p
e
n
d
e
n
c
e
.
W
h
i
l
es
e
v
er
a
Jc
x
a
m
p
l
e
sc
a
nbec
o
n
s
t
r
u
c
t
e
dt
oi
l
l
u
s
t
r
a
t
el
i
n
e
a
rd
c
p
e
n
d
e
n
c
ei
t
a
p
p
e
a
r
sr
e
a
s
o
n
a
b
l
et
oa
s
s
u
m
et
h
a
t
f
o
r
a 1
1 e
r
c
i
s
e o
p
t
i
m
i
z
a
t
i
o
n

t
h
enumbero
f
c
o
n
s
t
r
a
i
n
t
sc
a
n
n
o
te
q
u
a
lt
h
enumbero
f
u
l
I
k
n
o
w
n
s(m n
)
.I
nt
h
ep
r
e
v
i
o
u
si
l
l
u
s
n

n=2 dm=1(
t
h
e
r
ewaso
n
l
yo
n
ee
f
f
e
c
t
i
v
ec
o
n
s
a
i
nt
)
.I
ti
su
s
e
f
u
lt
or
e
c
o
g
n
i
z
et
h
a
t
i
sc
o
r
r
e
s
p
o
n
d
st
ot
h
ec
a
s
en>m
.
I
fm>n
t
h
e
r
ea
r
emoree
q
u
a
t
i
o
n
st
h
a
nt
h
en
u
m
b
e
ro
fv
a
r
i
a
b
l
e
s
. l
i
si
m
p
l
i
e
s
t
h
es
y
s
t
e
mo
fe
q
u
a
t
i
o
n
sr
e
p
r
e
s
e
n
t
e
db
yE
q
u
a
t
i
o
n(
3
.
1
1
)i
s i
n
c
o
n
s
i
s
t
e
n
ts
e
to
rh
a
sa
n
dn=2
r
e
d
u
n
d
a
n
ts
e
to
fc
q
u
a
t
i
o
n
s
.C
o
n
s
i
d
e
rt
h
ef
o
l
l
o
w
i
n
gi
l
l
u
s
t
r
a
t
i
o
nf
o
rm=3a
w
h
i
c
hu
s
e
st
h
es
a
m
e8
a
n
dg
2a
si
nt
h
cf
i
r
s
ti
l
l
u
s
t
r
a
t
i
o
nw
h
i
l
ea
d
d
i
n
ganew8
3
:

-3

g
3
:

+2x
2=
1

F
i
g
u
r
c3
.
2i
l
l
u
s
t
r
a
t
e
s t
h
es
e
to
fe
q
u
a
t
i
o
n
si
si
n
c
o
n
s
i
s
t
e
n
ts
i
n
c
eas
o
l
u
t
i
o
ndonot
t
.I
fo
n
ew
e
r
et
oe
x
i
s

tt
h
e
nt
h
et
h
r
e
el
i
n
e
sm
u
s
tp
a
s
st
h
r
o
u
g
hes
o
l
u
t
i
o
n
.S
i
n
c
e
e
x
i
s
e
y
a
r
a
l
ls
g
h
tl
i
n
e
se
r
ec
a
nbeo
n
l
yo
n
eu
n
i
q
u
ei
n
t
c
r
s
e
c
t
i
n
gp
o
i
n
tI
nF
i
g
u
r
c
3
.
2o
n
l
yt
w
oo
ft
h
t
h
r
e
el
i
n
e
si
n
t
e
r
e
c
ta
td
i
f
f
e
r
e
n
tp
o
i
n
t
s
.

X
1

F
l
g
u
r
3
.
2N
os
o
l
u
l
i
o
n
I
n
c
o
n
s
l
s
t
e
n
ts
e
t0
1e
q
u
a
t
i
o
n
s
.
R
c
d
e
f
i
n
eg
3a
s
g
3
:
X
I+2 =
3
.
5
a
s
h
e
dl
i
n
ei
nF
i
g
u
r
e3
.
3
.Nowau
n
i
q
u
s
o
l
u
t
i
o
nt
oe
p
r
o
b
l
e
m
T
h
i
snew8
3i
sed
a
t(
0

5
.1
.5
)i
se
s
t
a
b
l
i
s
h
e
d
.l
i
si
s ot
h
es
o
l
u
t
i
o
ne
s
t
a
b
l
i
s
h
e
db
yc
o
n
s
i
d
e
r
a
t
i
o
no
f8

dg
2a
l
o
n
e
.Thi
si
m
p
l
i
e
s g
3i
sr
e
d
u
n
d
a
n
t
.I
f8
1i
sm
u
l
t
i
p
l
i
c
db
y1
.5
8
2i
s
3d
e
f
i
n
e
da
b
o
v
e
.Tha
ti
s
m
u
l
t
i
p
l
i
e
db
y0.
5a
n
db
o
t
ha
d
d
e
d
.t
h
cr
e
s
u
l
ti
sg
g
3=1
.5g
l+0
.
5g
2

: X.+X
2=2

g
2
: -x1+
X
2=1

c
a
nb
eo
b
t
a
i
n
e
db
yl
i
n
e
a
r
l
yc
o
m
b
i
n
i
n
gg
a
n
dg
2
. i
s
.a
d
d
i
n
gc
o
n
s
t
a
n
tm
u
l
t
i
p
l
e
s
o
f n
c
t
i
o
n
s
.T
h
i
si
sa
n
o
t
h
e
re
x
a
m
p
l
eo
fl
i
n
e
a
rd
e
p
e
n
d
e
n
c
e
.T
h
i
sl
i
n
e
a
r
d
e
p
e
n
d
e
n
c
ea
n
dr
e
d
u
n
d
a
l
l
C
Yi
sa
l
s
oi
I
Iu
s
t
r
a
t
e
db
yt
h
es
e
t
g

=2

X+"
2

g
2
:x =1
g
3
: 2x1
.
s
a
m
ea
sg

2
x
l
)
2=4(

3
.
1 PROBLEMDEFINITION

102 L
lNEARPROGRAMMING

103

F
o
rn=
=ma
ndw
i
t
hg
l
:x
l+x
2=
=2a
n
dg
2
:X
1+X2=1

em
a
t
r
i
xA i
s

a
n
di
t
sd
e
t
e
n
n
i
n
a
n
ti
s

E-

-EEEEEE J

11

C
a
s
ena 2
;m
.=
3
:U
n
l
q
U
8s
o
l
u
l
l
o
n
U
n
o
a
rd
e
p
e
n
d
o
n
c
9
4

1
)
*
(
1
)

(
1
)
*
(
1)=2
d
e
t
(
A
)=I
A
I=(

.
r1

- "3.5

Fromt
h
e
o
r
e
m
si
nI
in
e
a
ra
l
g
e
b
r
a[
5
]
i
fd
e
t
(
A
)i
sn
o
tz
e
r
o

i
nw
h
i
c
hc
a
s
et
h
em
a
t
r
i
xA
t
e
r
m
e
dn
o
n
s
i
n
g
u
l
a
r

au
n
i
q
u
es
o
l
u
t
i
o
ne
x
i
s
t
st
ot
h
es
e
to
fe
q
u
a
t
i
o
n
s
.Th
i
swasu
cf
o
r
ee
xamplei
I
Iu
s
t
e
da
b
o
v
ea
s o
n
l
ys
o
l
u
t
i
o
nw l
o
c
a
t
e
da
t(
0
.
51
.5
)shown
.1
.
g
r
a
p
h
i
c
a
l
l
yi
nF
i
g
u
r
e3
C
o
n
s
i
d
e
rt
h
ee
x
a
m
p
l
ew
i
t
hg
l
:XI+X
2=2dg
2
:2
x1+2
x
n i
sc
a
s
e
2=4.I
d
e
t
(
A
)=I
A
I=(
1)
*
(
2
)(1)2
)=
=0
l
fd
e
t
(
A
)i
sz
e
r
ot
h
a
ti
sm
a
t
r
i
xA i
ss
i
n
g
u
l
a
r
r
ea
r
ee
i
t
h
e
rn
os
o
l
U
l
i
o
n
so
ri
n
f
i
n
i
t
e
se
x
a
m
p
l
ee
r
e were i
n
f
i
n
i
t
es
o
l
u
t
i
o
n
s
.D
c
t
c
r
m
i
n
a
n
t
so
f
s
o
l
l
l
l
i
o
n
s
. For i

-3
-3

X
1

.
3U
n
l
q
u
es
o
l
u
t
l
o
n
I
I
n
e
a
rd
e
p
e
n
d
e
n
c
e
.
F
l
g
u
r
e3

Th
i
ss
u
g
g
e
s
t
s ec
o
n
c
e
p
to
fr
e
d
u
n
d
a
n
c
yc
a
n be a
s
s
o
c
i
a
t
e
dw
i
t
hl
i
n
e
a
r
d
e
p
e
n
d
e
n
c
e
.Th
ed
i
s
c
u
s
s
i
o
no
fl
i
n
e
a
rd
e
p
e
n
d
e
n
c
eandi
n
d
e
p
e
n
d
e
n
c
ewase
s
t
a
b
l
i
s
h
e
d
ge
q
u
a
t
i
o
n
sa
b
o
v
e
.Thesamed
i
s
c
u
s
s
i
o
nc
a
nt
a
k
ep
l
a
c
cbyr
e
a
s
o
n
i
n
gont
h
e
nt
h
i
sc
a
s
et
h
ec
o
e
f
f
i
c
i
e
n
tma
xA s
h
o
u
l
dl
e
a
du
st
oe
c
o
e
f
f
i
c
i
c
n
t
st
h
e
m
s
e
l
v
c
s
.I
Iu
s
i
o
nr
c
g
a
r
d
i
n
gel
i
n
e
a
ri
n
d
e
p
e
n
d
e
n
c
eo
fas
e
to
fI
i
n
e
a
re
q
u
a
t
i
o
n
s
. e
sameconc
e
l
e
r
m
i
n
a
n
li
sn c
s
s
a
r
yt
od
e
v
e
l
o
pt
h
ec
r
i
t
e
r
i
af
o
rl
i
n
e
a
rd
e
p
e
n
d
e
n
c
e
.
c
o
n
c
e
p
to
fad

t
h
em
a
t
r
i
c
e
s
EEE E

35

(
3
.
1
6
)

1 n

Il

--

EEEEEEJ

35
--a

FEE-EEaE

hu

--EE E. J

Il

a
a

fl L
=
A

EEEEBEE-J

Ea.
a

E a

rMlML

)
=

a
=

a
21 2

g
2
: -X +0
.
5x3 1
.5

Th
enewm
a
t
r
i
xA.i
sc
a
l
l
e
dea
u
g
m
e
n
t
e
dmatrix-thecolumnso
fba
r
ea
d
d
e
dt
oA
.
A
c
c
o
r
d
i
n
g1
0r
e
m
so
fl
i
n
e
a
ra
l
g
e
b
r
a(
p
r
e
s
e
n
t
e
dh
e
r
ew
i
t
h
o
u
tp
r
o
of
)
:

ed
e
t
e
r
m
i
n
a
n
ti
se
x
p
r
e
s'
i
e
dande
v
a
l
u
a
t
e
da
s

a
a
1

g
t
: x1+X
2+X
3 3

D
e
t
e
r
m
l
n
a
n
t
:Thed
e
t
e
r
m
i
n
a
n
ti
sa
s
s
o
c
i
a
t
e
dw
i
t
has
q
u
a
r
cm
a
t
r
i
x
.F
o
rag
e
n
e
r
a
l2x
2m
a
t
r
i
xA
where

h
i
g
h
e
r
r
d
e
rs
q
u '
em
a
t
r
i
c
e
sa
r
ee
v
a
l
u
a
t
e
dbys
e
t
t
i
n
gu
pl
o
w
e
r
o
r
d
e
rd
e
t
e
r
m
i
n
a
n
t
s
u
n
t
i
leyare ducedt
oa2x2d
e
t
e
r
m
i
n
a
n
t
whichi
se
v
a
l
u
a
da
sa
b
o
v
e
.An
y
t
e
x
t
b
o
o
kone
n
g
i
n
e
e
r
i
n
gm
a
t
h
e
m
a
t
i
c
so
rl
i
n
e
a
ra
1gebrashouldi
l
l
u
s
t
r
a
t
e i
st
e
c
h
n
i
q
u
e
.
r
o
d
u
c
e
dh
e
r
e
.
I
ti
sn
o
t p
I
fn"
"m
d
i
s
c
u
s
s
i
o
no
ft
h
ee
x
i
s
t
e
n
c
co
fs
o
l
u
t
i
o
n
sr
e
q
u
i sa
d
d
i
t
i
o
n
a
lc
o
n
c
e
p
t
sI
ik
e
r
a
n
ko
fam
a
t
r
i
xa
n
da
u
g
m
e
n
l
e
dm
a
t
r
i
x
.S
i
n
c
e c
a
s
en>m i
so
fi
n
t
e
r
e
s
ti
n
o
p
t
i
m
i
z
a
t
i
o
no
n
l
yt
h
a
tc
a
s
ei
s employed i
ns
u
b
s
u
e
n
ti
l
Iu
s
t
r
a
t
i
o
n
.A u
s
e
f
u
l
d
i
s
c
u
s
s
i
o
nw
i
l
ln
e
e
da
t1
e
a
s
tt
h
r
e
ev
a
r
i
a
b
l
e
sa
n
dtwoe
q
u
a
t
i
o
n
s
.T
h
r
e
ev
a
r
i
a
b
l
e
sw
i
l
l
d
e
n
yt
h
eu
s
eo
fg
r
a
p
h
i
c
st
od
e
v
e
l
o
pt
h
ef
ol
Iowingc
o
n
c
e
p
t
s
.Usingt
h
es
e
t

(
3.
15
)

I
ft
h
ea
u
g
m
e
n
t
e
dm
a
t
r
i
x(A
)a
n
dt
h
em
a
t
r
i
xo
fc
o
e
f
f
i
c
i
e
n
t
s(
A
)h
a
v
et
h
esame

e
n e
r
M m
anys
o
l
u
t
i
o
n
s
.
r
a
n
kr<n

1
0
4

L
lNEARPROGRAMMING

3
.
1 PROBLEMDEFINITION 105

.I
f
ea
u
g
m
e
n
t
e
dma
x(A

)
dem
a
t
r
i
xo
fc
o
e
f
f
i
c
i
e
n
t
s(
A
)d
on
o
th
a
v
et
h
e

as
o
l
u
t
i
o
nd
o
e
sn
o
te
x
i
st
.
s
a
m
er
a
n
k
I
ft
h
ea
u
g
m
e
n
t
e
dm
a
t
r
i
x(A
)a
n
dt
h
em
a
t
r
i
xo
fc
o
e
f
f
i
c
i
e
n
t
s(
A
)h
a
v
et
h
es
a
m
e
r
a
n
kr=1

w
h
e
r
em=n
t
h
e
nt
h
e
r
ei
sau
n
i
q
u
s
o
l
u
t
i
o
n
.

a
l
1 a

a
ll
"
l

~2

33

I~.

1
-~ I

-~:l l

(
3
.
1
7
)

2A
L

a
a

z
a

'Ea'EA

tA

a
a
A

11

'

-EA--A2

''

Ill

=(
1)
(
0
)+(
1
)
(
2
)+(
1
)
(
3
)=-1

A
.i
sn
o
n
s
i
n
g
u
l
a
ra
n
di
t
sr
a
n
ki
s3wh
I
i
et
h
er
a
n
ko
fAi
s2
.T
h
es
y
s
t
e
mo
fe
q
u
a
t
i
o
n
s
o
r
c
ma
n
dc
a
nb
es
e
e
ng
r
a
p
h
i
c
a
l
l
yi
nF
i
g
u
r
e
w
i
l
lh
a
v
en
os
o
l
u
t
i
o
n
sf
r
o
mt
h
ea
b
o
v
et
h
3
.
2
.F
o
rt
h
es
a
k
eo
fc
o
m
p
l
e
t
e
n
e
s
su
s
i
n
gt
h
ep
r
e
v
i
o
u
s
l
yd
e
f
i
n
e
dd
e
p
e
n
d
e
n
ts
e
to
f
e
q
u
a
t
i
o
n
s
:
g
l
: x1+x
2=2
g
2
: -x1+x
1
2=
g
3
:2
x1+2
x
a=4

g
2
: -x1+ =1

w
h
i
c
hy
i
e
l
d
st
h
ef
o
l
l
o
w
i
n
gma
c
e
s
:

g
3
: x1+2
x
2=1

[
l}
l

Them
a
t
r
i
c
e
sf
o
rt
h
i
ss
y
s
t
e
ma
r
e

i
]
l

1
!
]

.N
o
t
e u
s
i
n
gt
h
ef
i
r
s
ta
n
dl
a
s
tr
o
wo
fAf
o
rt
h
ec
a
l
c
u
l
a
t
i
o
no
ft
h
e
Ther
a
n
ko
fAi
s2
i
n
g
u
l
a
r2x2subma
x
.Thed
e
t
e
r
m
i
n
a
n
to
fA i
s
d
e
t
e
r
m
i
n
a
n
ty
i
e
l
d
sas

nu

00

'A

2A

'

.E

'ia T

'
'
z

.
z
'
Z

a ' A a

'1 1 L

A--a--E

'

Ther
a
n
ko
fAc
a
n
n
o
tb
eg
r
e
a
t
e
r 2
s
i
n
c
en=2a
n
di
sl
e
s
st
h
m=3
.Fromp
r
i
o
r
e
t
e
r
m
i
n
a
n
to
ft
h
ef
i
r
s
tt
w
or
o
w
so
fAi
s2
.T
h
i
si
sg
r
t
e
rt
h
a
nz
e
r
o
c
a
l
c
u
l
a
t
i
o
n
sed
r
e
f
o
r
et
h
er
a
n
ko
fAi
s2
.F
o
rA.t
h
ed
e
t
e
r
m
i
n
a
n
to
ft
h
e3x3m
a
t
r
i
xh
a
st
ob
e
a
n
dt
h
e
x
a
m
i
n
e
d
.Thed
e
t
e
r
m
i
n
a
n
to
fag
e
n
e
r
a
lnxns
q
u
a
r
em
a
t
r
i
xAC
a
nb
ec
a
l
c
u
l
a
t
e
di
n
'
ac
t
o
r
so
ft
h
em
a
t
r
i
x
.Thec
o
f
a
c
t
o
ri
n
v
o
l
v
e
sede
n
n
i
n
a
n
to
fam
i
n
o
r
t
e
r
m
so
ft
h
ec
m
a
t
r
i
x
w
h
i
c
hi
sa
n(
n-1
)x(
n-1)m
a
t
r
i
xo
b
n
e
db
yd
e
l
e
t
i
n
ga
na
p
p
r
o
p
r
i
a
t
es
i
n
g
l
e
rowa
n
das
i
n
g
l
ec
o
l
u
m
nf
r
o
mt
h
eo
r
i
g
i
n
a
lnXnma
i
xA
m
u
l
t
i
p
l
i
e
db
y1o
r-1
d
e
p
e
n
d
i
n
go
ne
t
o
o
ft
h
er
o
wa
n
dc
o
l
u
m
nv
a
l
u
e
s
.R
e
f
e
r
e
n
c
e
s5a
n
d6c
p
r
o
v
i
d
e
i
n
f
o
r
m
a
t
i
o
n
.H
e
r
e d
e
t
e
r
m
i
n
a
n
tf
o
rt
h
e3X3m
a
t
r
i
xi
sd
e
f
i
n
e
df
o
ri
I
lu
s
t
r
a
t
i
o
n
mor

b=[!} A'=H

A=

b=[I} A'=H

l
e
r
e
f
o
r
e

g
l
: x1+x2=2

}
l

~1

A=H

0
;lln ailn(+l)a111273+( l)a1212J2+(+l)a13

A=1

Rank01aM
a
t
r
i
x
:W
h
i
l
et
h
e
r
ea
r
ef
o
r
m
a
ld
e
f
i
n
i
t
i
o
n
sf
o
rd
e
f
i
n
i
n
gt
h
er
a
n
ko
fe
x

au
s
e
f
u
lwayt
od
e
t
e
r
m
i
n
e r
a
n
k
i
s

l
o
o
k e
d
e
t
e
n
n
t.
'
h
e
r
a
n
k
o
f
a
ma
m
a
t
r
i
xAi
st
h
eo
r
d
e
ro
fel
a
r
g
e
s
tn
o
n
s
i
n
l
a
rs
q
u
a
r
es
u
b
m
a
t
r
i
xo
fAt
h
a
ti
se
n
z
e
r
o
.
l
a
r
g
e
s
ts
u
b
m
a
t
r
i
xw
i
t
had
e
t
e
r
m
i
n
a
n
to
t
h
e
ra
I
nt
h
ee
x
a
m
p
l
ea
b
o
v
e

t
h
el
a
r
g
e
s
ts
q
u
a
r
es
u
b
m
a
t
r
i
xi
sa2x2m
a
t
r
i
x(
s
i
n
c
em=2
a
n
dm <n
)
.T
a
k
i
n
gt
h
esubma
i
xw
h
i
c
hi
n
c
l
u
d
e
st
h
ef
i
r
s
tt
w
oc
o
l
u
m
n
so
fAe
t
h
e
r
e
f
o
r
en
o
n
s
i
n
g
u
l
a
r
.
d
e
t
e
r
m
i
n
a
n
tw
a
sp
r
e
v
i
o
u
se
s
t
a
b
l
i
s
h
e
dt
oh
a
v
eav
a
l
u
eo
f2
T
h
u
s
t
h
er
a
n
ko
fAi
s2(
r=2
)
.l
es
a
m
ec
o
l
u
m
n
sa
p
p
e
a
ri
nA m
a
k
i
n
gi
t
sr
a
n
ka
l
s
o
2
. l
e
r
e
f
o
r
ei
n
f
i
n
i
t
e
l
ym ys
o
l
u
t
i
o
n
se
x
i
s
t(
r<n
)
.Onew
ayde
r
m
i
n
eE
s
o
l
u
t
i
o
n
si
st
oa
s
s
i
g
n(
n-r
)v
a
r
i
a
b
l
e
sa
r
b
v
a
l
u
e
sa
n
du
s
et
h
e
mt
od
e
t
e
r
m
i
n
e
v
a
l
u
e
sf
o
rt
h
er
e
m
a
i
n
i
n
grv
a
r
i
a
b
l
e
s
.Thev
a
l
u
en-ri
sa
l
s
oo
f
t
e
ni
d
n
t
i
f
i
e
da
s
d
e
gr
e
eo
f
f
r
e
e
d
o
m(DOF)f
o
rt
h
es
y
s
t
e
mo
fe
q
u
a
t
i
o
n
s
.l
nt
h
ep
a
r
t
i
c
u
l
a
re
x
a
m
p
l
ea
b
o
v
e

eDOFc
a
nb
ei
d
e
n
t
i
f
i
e
dw
i
t
hav
a
l
u
eo
f1(
i
.
e
.
3-2
)
.F
o
ri
n
s
t
a
n
c
e

x3 c
a
nb
e
a
s
s
i
g
n
d
av
a
l
u
eo
f1i
nw
h
i
c
hc
a
s
eXI=0
.
5 dX2=1
.5
.Ont
h
o
t
h
e
rh
a
n
c
J
.i
fX3=2
t
h
e
nXI=0
.
2
5a
n
dX2=0
.
7
5
.Thea
b
o
v
ec
a
s
e
si
l
l
u
s
t
r
a
t
et
h
a
tf
o
raDOFo
f1

o
n
c
ex3i
s
e
d
.Xl dX2c
a
nb
eo
b
t
a
i
n
e
dt
h
r
o
u
g
hsomef
u
r
t
h
e
rp
r
o
c
e
s
s
i
n
g
.
s
e
l
e
c
t
I
1
u
s
t
r
a
t
i
n
ger
a
n
ki
sei
n
c
o
n
s
i
s
t
e
n
ts
y
s
t
e
mo
fe
q
u
a
t
i
o
n
s
A
n
o
t
h
e
re
x
a
m
p
l
ef
o
ri
i
n
o
d
u
c
e
de
l
i
e
rw
h
e
r
en=2 dm=3

r
e
p
r
o
d
u
c
e
df
o
rc
o
n
v
e
n
i
e
n
c
e

i
neb
o
o
k
. l
e(+1)d(
1
)b
e
l
o
wa
r
o
b
t
a
i
n
e
db
yu
s
i
n
g
(1
)
;

w
h
e
r
i
a
n
d
j
r
e
p
r
e
s
e
n
terowa
n
dc
o
l
u
m
nv
a
l
u
e
so
ft
h
ec
o
e
f
f
i
c
i
e
n
tm
u
l
t
i
p
l
y
i
n
gt
h
ev
a
l
u
e

w
h
i
c
hm
a
k
e
sA.s
i
n
g
u
l
a
r
.However

t
h
er
a
n
ko
fA
i
sa
tl
e
a
s
t2d
u
et
o p
r
e
s
e
n
c
e
o
f
l
e
s
u
b
m
a
i
xw
h
i
c
hd
e
t
e
r
m
i
n
e
dt
h
er
a
n
ko
fA
.Fromt
h
ea
b
o
v
et
h
e
o
r
e
m

e
r
k
w
h
i
c
hi
s2a
n
dt
h
i
si
sl
e
s
s m=3
.I
tc
a
nb
ec
o
n
c
l
u
d
e
d
o
fAi
se
q
u
a
lt
ot
h
er
a
r
t
ko
fA.
t
e
r
ea
r
ei
n
f
i
l
l
i
t
es
o
l
o
n
sf
o
r s
y
s
t
e
mo
fe
q
u
a
t
i
o
n
s
.

106

L
lNEARPROGRAMMING

l
ON
3
.
2 GRAPHICALSOLUT

43

eq

d
L
'

EE--EEEBEESEEJ

nunu'E

FEE--EE E -BEE-L

EBB-a'EBEE--J

43

nu' nu
rlll111L

1111111
'znunu

(
3
.
1
8
)

Thea
d
v
a
n
t
a
g
eo
fu
n
i
tv
e
c
t
o
r
si
se
yc
a
nc
s
t
a
b
l
i
s
h yp
o
i
n
ti
nt
h
eCar
s
i
a
n
s
p
a
c
e
t
h
r
o
u
g
ha
na
p
p
r
o
p
t
et
r
i
a
do
fnumberss
i
g
f
y
i
n
gc
o
o
r
d
i
n
a
t
ev
a
l
u
e
so
r10 t
i
o
n
.
a
n
yp
o
i
n
ti
neCartians
p
a
c
ec
a
nbed
e
t
e
n
n
i
n
e
db
yanys
e
to
ft
h
l

A
c

a
l
l
y
l
i
n
e
a
r
l
yi
n
d
e
p
e
n
d
e
n
tv
e
c
t
o
r
s
.Th
e
s
ev
e
c
t
o
r
sa
r
et
e
r
m
e
dt
h
eb
a
s
i
sv
e
c
t
o
r
s
.I
nt
h
i
sc
a
s
e
.
t
h
es
i
m
p
l
ec
o
n
n
e
c
t
i
o
nb
e
t
w
e
e
nt
h
ep
o
i
n
ta
n
dt
h
ee
l
e
m
e
n
t
a
r
ya
d
d
i
t
i
o
ni
l
l
u
s
t
r
a
t
e
d
w
i
l
ln
o
tb
ea
v
a
i
l
a
b
l

.
F
u
e
r
m
o
r
eought
h
em
e
t
h
o
d
so
fi
I
n
e
a
ra
l
g
e
b
r
a

t
h
e
s
e
a
b
o
v
e
t
h
r
e
ev
c
c
t
o
r
sc
a
nb
er
e
d
u
c
c
dt
ot
h
eu
n
i
tv
c
c
t
o
r
st
h
r
o
u
g
he
l
e
m
e
n
t
a
r
yr
o
w
/
c
o
l
u
m
n
o
p
c
r
a
t
i
o
n
s
a
l
s
oc
a
l
l
e
dG
a
u
s
s
J
o
r
d
a
ne
l
i
m
i
n
a
t
i
o
no
rr
e
d
u
c
t
i
o
n
.I
nnd
i
m
e
n
s
i
o
n
a
l
E
u
c
l
i
d
e
a
ns
p
a
c
es
p
a
n
n
e
dbyt
h
end
e
s
i
g
nv
a
r
i
a
b
l
e
s
ep
o
i
n
tp
"(
p
.
.P
2
.
.
.

p
"
)
w
i
l
l
a
l
u
e
se
n
c
l
o
s
e
di
np
a
r
n
e
s
e
s
.Thec
o
r
r
e
s
p
o
n
d
i
n
g
b
er
e
p
r
e
s
e
n
t
e
dbyac
o
l
l
e
c
t
i
o
no
fnv
u
n
i
tv
e
c
t
o
r
sa
r
edf
i
n
e
da
s

o
o

(
3
.
1
9
)

e
"
=
I
.
l
ev
e
c
t
o
rfrom o
r
i
g
i
n(
0
)t
oP"i
s P s
e
n
t
e
da
s

X
3

U
n
/
tV
e
c
t
o
r
s
:R
e
a
l
w
o
r
l
do
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
e
m
sf
r
e
q
u
e
n
t
l
yi
n
v
o
l
v
e al
a
r
g
e
fv
a
r
i
a
b
l
e
s
numbero
fd
e
s
i
g
nv
a
r
i
a
b
l
e
s(
nv
a
r
i
a
b
l
e
s
)
.I
nr
e
a
lLPp
r
o
b
l
e
m
s
. enumbero
z
a
t
i
o
np
r
o
b
l
e
m
.i
ti
s
c
a
nbes
t
a
g
g
e
r
i
n
g
.F
o
ras
u
c
c
e
s
s
f
u
ls
o
l
u
t
i
o
nt
oeop
i
m
p
l
i
c
i
t
l
yassumedt
h
a
tt
h
es
e
to
fd
e
s
i
g
nv
a
r
i
a
b
l
e
si
sl
i
n
e
a
r
l
yi
n
d
e
p
e
n
d
e
n
.
tTh
a
ti
s
.
t
h
e
r
ei
snod
i
r
e
c
td
e
p
e
n
d
e
n
c
eamonga
n
ytwov
a
r
i
a
b
l
e
so
ft
h
es
ct
.I
ni
sc
a
s
e
.t
h
e
a
b
s
t
r
a
c
t~pace spawnedbyt
h
ed
e
s
i
g
nv
a
r
i
a
b
l
e
si
sc
a
l
1
eden
d
i
m
e
f
l
s
i
o
f
l
a
lE
u
c
l
i
d
e
a
n
eworda
b
s
t
r
a
c
ti
su
s
c
dh
e
r
ct
os
i
g
n
i
f
yt
h
a
ti
ti
sac
o
n
s
t
r
u
c
ti
nt
h
ei
m
a
g
i
n
a
t
i
o
n
s
p
a
c
e
.Th
a
so
n
l
yt
h
r
e
ev
a
r
i
a
b
l
e
sc
a
nb
eaccommodatedi
n f
a
m
i
l
i
a
r3Dp
h
y
s
i
c
a
ls
p
a
c
e
. l
e
ss
p
a
c
e
.Th
eE
u
c
l
i
d
e
a
ns
p
a
i
sad
i
r
e
c
te
x
t
r
a
p
o
l
a
t
i
o
no
f
d
e
s
i
g
nv
e
c
t
o
ri
sap
o
i
n
ti
n i
c
a
l3Ds
p
a
c
e ts
u
r
r
o
u
n
d
su
s
.T
h
i
ss
p
a
c
ei
st
e
r
m
e
d
t
h
ep
h
y
s
i
c
a
lgeome
C
a
r
t
e
s
i
a
ns
p
a
c
e
.Fora
ni
1
1u
s
t
r
a
t
i
o
n
c
o
n
s
i
d
e
rt
h
ep
o
i
n
t(
2
3
2
)i
nC
a
r
t
c
s
i
a
ns
p
a
c
eo
ri
n
t
h
er
e
c
t
a
n
g
u
l
a
rc
o
o
r
d
i
n
a
t
es
y
s
t
e
m
.I
nF
i
g
u
r
e3

4
.t
h
ep
o
i
n
ti
smarkedP
.The adof
numbers w
i
t
h
i
np
a
r
e
n
t
h
e
s
e
sd
e
n
o
t
e
st
h
ev
a
l
u
e
sa
l
o
n
gt
h
ex
"X2andX3 es
s
p
e
c
t
i
v
e
l
y
.
Th
eC
a
r
t
e
s
i
a
ns
p
a
c
ei
si
d
e
n
t
i
f
i byas
e
to
fm
u
t
u
a
l
l
yp
e
e
n
d
i
c
u
l
l
i
n
e
so
ra
x
e
s
markedx"x2'andx3
Thep
o
i
n
twherct
h
e
yi
n
t
e
r
s
e
c
ti
st
h
eo
r
i
g
i
n(
0
)o
ft
h
ec
o
o
r
d
i
n
a
t
e
s
y
s
l
e
m
.Thep
o
i
n
tPc
a
na
l
s
ob
ea
s
s
o
c
i
a
t
e
dW
i
t
l
lt
h
ev
e
c
t
o
rOP.drawnf
r
o
mt
l
l
eo
r
i
g
i
n
otoepointP.Veryoftenapointandavector-twoverydifferententities-are
t
h
ev
e
c
t
o
rOPc
a
n
u
s
e
ds
y
n
o
n
y
m
o
u
s
l
y
.Frome
l
e
m
e
n
t
a
r
yv
e
c
t
o
ra
d
d
i
t
i
o
np
r
i
n
c
i
p
l
e
s
becons
c
t
e
dbyc
o
n
s
e
c
u
t
i
v
e
l
yj
o
i
n
i
n
gt
h
r
e
ev
e
c
t
o
r
sa
l
o
n
gt
h
et
h
r
c
o
o
r
d
i
n
a
t
e e
s
.
x
i
sa
d
d
i
n
gav
e
c
t
o
r
I
nt
h
ef
i
g
u
r
et
h
i
si
sshownbyl
a
y
i
n
go
u
tav
e
c
t
o
ra
l
o
n
gt
h
ex a
a
l
o
n
geX2a
x
i
st
oi
tandf
i
n
a
l
l
ya
d
d
i
n
gt
h
ct
h
i
r
dv
e
c
t
o
ra
l
o
n
gt
h
cX3 a
x
i
s
.The
a
d
d
i
t
i
o
n
a
ls
m
a
l
la
r
r
o
w
si
nt
h
e
f
i
g
u
r
ea
r
eeL
m
i
tv
e
c
t
o
r
s
.l
e
yh
a
v
eam
a
g
n
i
t
u
d
eo
f
oneandad
i
r
e
c
t
i
o
na
l
o
n
ge
a
c
ho
ft
h
e e
s
.Th
e
s
eu
n
i
tv
e
c
t
o
r
sa
r
el
a
b
e
l
e
de
"e
2
and
e3i
nt
h
ef
i
g
u
r
e
.M
a
t
h
e
m
a
t
i
c
a
l
l
y

p
o
i
n
tP(
2
3
2
)andt
h
ev
e
c
t
o
ra
d
d
i
t
i
o
nc
a
nb
e
rcpresented(columnv
e
c
t
o
r
s
)

r'EEEaEaE E -EL

EEEEEEEEEEEEJ

rhHh

Au
s
e
f
u
lo
b
s
e
r
v
a
t
i
o
no
fm
a
t
r
i
xA thep
c
e
d
i
n
ge
x
a
m
p
l
ei
s twoo
ft
h
er
o
w
s
c
e
ss
i
n
g
u
l
a
r
.Th
esame
a
r
eac
o
n
s
t
a
n
tm
u
l
t
i
p
l
eo
fe
a
c
ho
t
h
e
r
.smakesema
ta
r
em
u
l
t
i
p
l
e
so
fe
a
c
ho
t
h
e
r
.
h
o
l
d
sf
o
rtwocolumnsa

107

OP=
pe +P
2
e
2+.
.
.
+p"ell
T
h
e
s
enewd
e
f
i
n
i
t
i
o
n
sw
i
l
lb
es
u
f
f
i
c
i
e
n
tt
ou
n
d
e
r
s
t
a
n
dt
h
et
e
c
h
n
i
q
u
e
su
s
e
dt
os
o
l
v
eLP
p
r
o
b
l
e
m
s
.

3
.
2 GRAPHICALSOLunON
l
eg
r
a
p
h
i
c
a
ld
i
s
c
u
s
s
i
o
ni
sa
g
a
i
nl
i
m
i
t
e
d10twod
e
s
i
g
nv
a
r
i
a
b
l
e
si
n o
r
i
g
i
n
a
l
P
.
p
r
o
b
l
e
md
e
s
c
r
i
p
t
i
o
n
.I
ti
su
s
e
dh
e
r
et
od
e
v
e
l
o
pc
e
r
t
a
i
ng
e
o
m
e
t
r
i
ci
d
e
a
sr
e
l
a
t
e
d10L

t
h
i
n
t
r
o
d
u
c
t
i
o
no
fs
l
a
c
kv
a
r
i
a
b
l
e
sw
i
l
le
a
s
i
l
ye
x
c
e
e
dt
h
i
sl
i
m
i
to
ftwov
a
r
i
a
b
l
e
s
InLP
a
t
i
o
n
.s
ot
h
eg
r
a
p
h
i
c
a
ls
o
l
u
t
i
o
ni
si
l
l
u
s
t
e
dw
i
t
h s
p
e
c
tt
oe
f
o
rg
r
a
p
h
i
c
a
li
l
l
o
r
i
g
i
n
a
lv
a
r
i
a
b
l
e
si
nt
h
ed
e
s
i
g
np
r
o
b
l
e
m
.MATLABi
so
n
c
ea
g
a
i
nu
s
e
dt
o awe
g
r
a
p
h
i
c
a
ls
o
l
u
t
i
o
n
.Th
ep
r
o
b
l
e
mi
sn
o
tr
e
d
u
c
e
d10t
h
es
t
a
n
d
a
r
df
o
r
m
a
tf
o
rt
h
eg
r
a
p
h
i
c
a
l
s
o
l
uo
n
.I
ti
si
m
p
o
r
t
a
n
tt
odo50f
o
ran
u
m
e
r
i
c
a
ls
o
l
u
t
i
o
n
.S
i
n
c
eI
h
ep
r
i
m
a
r
yg
r
a
p
h
i
c
a
l
e
l
e
m
e
n
ti
ses
t
r
a
i
g
h
tl
i
1
Le
afunctionf
o
rd
r
a
w
i
n
g1
in
e
si
sd
e
v
e
l
o
p
e
dt
h
r
o
u
g
ht
h

f
u
n
c
t
i
o
nm
f
i
l
edrawLine.mb
e
l
o
w
. i
sf
u
n
c
t
i
o
ni
sa
l
s
ou
s
e
dt
op
r
o
v
i
d
emore
e
x
p
o
s
t
oMATLABp
r
o
g
r
a
m
m
i
n
g
.

3
X2

F
l
g
u
r
e3
.
4R
e
c
t
a
n
g
u
l
a
rc
o
o
r
d
l
n
a
t
es
y
s
t
e
m
.

L
l
ne.m: i
saf
u
n
c
t
i
o
nmf
i
l
ef
o
rd
r
a
w
i
n
gs
t
r
g
h
tl
i
n
e
s
.l
e1
i
n
e10b
e
draw
drawni
sr
e
p
r
e
s
e
n
t
e
da
s

108

3
.
2 GAAPHICALSOLUTION

L
lNEAAPAOGAAMMING

+by=c

S
i
n
c
ex
ya
r
eusedsymbolsf
o
rv
a
r
i
a
b
l
e
s
x
lr
e
p
r
e
s
e
n
t
st
h
el
o
w
e
rr
a
n
g
eonx
w
h
i
l
e
L
im
i
t
i
n
gI
in
e
s
/
c
o
n
s
t
r
a
i
n
t
sc
a
nb
cd
r
a
w
nb
ys
e
t
t
i
n
ge
i
t
h
e
r
ao
rbt
oz
e
r
o
.Color(
r
e
d
)i
su
s
c
dt
od
r
a
wap
'
a
l
l
e
ll
i
n
ea
t10%i
n
c
r
e
a
s
e
/
d
e
c
r
e
a
s
ci
n
t
h
eo
r
i
g
i
n
a
lv
a
l
u
eo
fc(
d
e
p
e
n
d
i
n
gont
h
ed
i
r
e
c
t
i
o
no
ft
h
eo
r
i
g
i
n
a
li
n
e
q
u
a
l
i
t
y
)i
n
s
t
e
a
d
o
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
ni
si
nb
l
u
e
o
fh
a
s
hm
a
r
k
s
.L
i
m
i
tc
o
n
s
t
r
a
i
n
t
sa
r
ei
nm
a
g
e
n
t
aa
n
dt
h
1
1
.
A
B c
ommandh
e
r
ei
sel
i
n
ecommand. Youa
r
e
d
a
s
h
e
dl
i
n
e
. ThenewMA
e
n
c
o
a
g
e
dt
os
kh
e
l
pont
h
ecommand.Th
ec
o
d
e
:

else cmult
end
end

-1;

X2r
e
p
r
e
s
e
n
t
st
h
eu
p
p
e
rr
a
n
g
c
.

drawLine.m
D
rawinglinearconstraintsforLPprogrammingproblems
D
r. P.Venkataraman
O
ptimizationUsing Matlab
C
hapter 3 linear Programming

Lines are represented as: ax + by c (c >= 0


)
x1x2 indicate the range of x for the line
typ indicates type of line being drawn 1 (<=)
9 (>=)
n (none)

The function will draw line(s) in the figure window


t
he green solid line represents the actual value
% of the constraint
t
he red dashed line is 10 larger or smaller
(
in lieu of hash marks)
t
he limit constraints are identified inmagentacolor
t
he objective function is in blue dashed lines

function drawLine(x1
x2
a
b
c
typ)
recognize the types and set color
if (typ
'n')
str1 'b';
str2
'
b
'
cmult 1;
else
str1 '
"'
g
'
str2 '
" 'r';
end

values for drawing hash marks


dependingon the direction of inequality
if (typ = 'n')
if (typ == 'l
')
cmult +1;

s
et up a factor for drawing the hash constraint
if (abs(c) >= 10)
cfac 0.025;
) & (abs(c) < 1
0
)
elseif (abs(c) > 5
cfac 0.05;
else
cfac 0.1;
end

if (
c

0)

cdum
else

m
cd
end

cmult*O.l;

if b

(
1 + cmult* cfac)*c;

0 then determine end points of line x line

if (
b=0)

y1
(
c -a*x1)/b;
y1n
(cdum -a* x1)/b;
y2
(
c - a* x2)/b;
y2n (cdum a*x2)/b;
else
i
dentify limit constrains by magenta color
str1 '
m
'
str2 'm';
y1 x1; set y1 same length as input x1
y2 x2; set y2 same length as input x2
x1 c/a; adjust x1 to actual value
x2 c/a; adjust x2 to actual value
y1n "
'0
; set y 0
;
y2n 0
; set y 0
end
if (
a

0)

str1
'
m
'; set color for limit line
str2
'm'; set color for limit line
end;
drawa
xis with solid black color
hh

x2][
0
0]);
line([
x
1

109

110

3
.
2G
R
A
P
H
I
C
A
LS
O
L
U
T
I
O
N 1
1
1

L
lN
E
A
RP
R
O
G
R
A
M
M
I
N
G

'LineWidth'
1
'Color'
'k');
set(hh

8
Or'

hv 1ine([
0
0
][xl
x2]);
set(hv
'Linewidth'
1
'Color'
'
k
'
)
;

60

start drawing the lines

hl

y2]);
line([xl x2] [yl

if (
t
p

'n')

set(hl
'LineWidth'
2
'LineStyle'"

'
'Color'
strl);
else
set(hl
'LineWidth'
1
'LineStyle'
'

'
'Color'
strl);
end
if (b -= O)&{a -= 0
)
text(x2
y2
num2str(c));
end
if( b =
=0)I
(a =
=0)I( yp

.
.---

'n')

~~

grid
returnend
grid;
y2n]);
h2 line([xl x2] [yln
'LineWidth'
0.S
'LineStyle'
'
:
''Color'
str2);
set(h2
grid
hold on

..

..

.
.

.
.
.
..
.

.
.

-4
0

1
0
1
5
2
0
N
u
m
b
o
r0
1m
a
c
h
l
n
a
s0
1I
y
p
eA

.J5
2
5
0

2
5

3
0

x
a
m
p
l
e3
.
1
.
F
l
g
u
r
e3
.
5G
r
a
p
h
l
c
a
ls
O
l
u
l
i
o
nE

3
.
2
.
1 Example3
.
1

3
.
2
.
2 C
h
a
r
a
c
t
e
r
l
s
t
l
c
so
ftheSolutlon

Thep
r
o
b
l
e
mf
r
o
mp
a
g
e94i
sr
e
p
r
o
d
u
c
e
d00 moref
o
rc
o
o
v
e
o
i
e
o
c
e
.

l
m
beg
e
t
r
ye
v
i
d
n
ti
nt
h
eg
r
a
p
h
i
c
a
ls
o
l
u
t
i
o
no
fExample3
.
1i
su
s
e
dt
oe
x
p
l
a
i
nsome
o
ft
h
ec
o
n
c
e
p
t
sa
s
s
o
c
i
a
t
e
dw
i
t
hLPa
n
di
l
Sn
u
m
e
r
i
c
a
1t
e
c
h
n
i
q
u
e
.F
i
g
u
r
e3
.
6i
sE
g
r
a
p
h
i
ca
Jr
c
p
s
e
n
l
a
t
i
o
no
f c
o
n
s
t
r
a
i
n
t
si
n
v
o
l
v
e
di
nExample3
.
1(
s
a
m
ea
sF
ig
u
r
e
c
t
i
o
n
)
.I
no
r
d
e
rt
o l
a
t
eegcometryLPc
o
n
c
e
pl
S
t
h
c
3
.
5w
i
t
h
o
u
tt
h
eo
b
j
e
c
t
i
v
cn
s
t
a
n
r
df
o
r
r
n
a
to
fLPi
sn
S
y
.Thes
t
a
n
d
a
r
df
o
r
r
n
a
te
s
t
a
b
l
i
s
h
e
db
e
f
o
r
ei
s

{
X
)
:990X
J+9
00X
2+5250
Maximize j

1
)
(
3.

.
(
X
)
:0
.
4
x
.+0
.6
x
.
5
S
u
b
j
t
t
o
: g
2S8

(
3
.
2
)

g
2
(
X
)
:3
x
J-X
2S2
5

(
3
.
3
)

):3
x
.+6x
2S70

(
3
.
4
)

;.

i~ 0
x
.~ 0

T
h
g
r
a
p
h
i
c
a
1s
o
l
u
t
i
o
oi
sshowni
nF
i
g
u
r
e3
.
5
.Th
es
o
l
u
t
i
o
n
r
c
a
df
r
o
mt
h
cf
i
g
u
(
y
o
u
C
Zme
a
r
c
a
o
fs
o
l
u
t
i
o
o
)
i
s
x
j=
1
0

a
n
d
x
i=
7
.I
ti
sci
n
t
e
r
s
e
c
t
i
o
no
ft
h
ea
c
t
i
v
e
c
o
n
s
t
r
a
i
n
t
s(
3.3)d(
3.
4
)
.
l
ea
c
t
u
a
lv
a
l
u
e
sa
r
ex
j=
1
0.
48 dx
i=
6.
4
2
.S
i
n
c
ea
n
i
n
t
e
g
r
a
1numbero
fm
a
c
h
i
n
e
sh
a
v
et
ob
eo
r
d
e
r
e
dt
h
es
o
l
u
t
i
o
ni
sadjusteda
n
e
i
g
h
b
o
r
i
n
gi
n
t
e
g
e
rv
a
1u a
ts
a
t
i
s
f
i
e
sl
h
ec
o
n
t
r
a
i
n
t
s
.

M
i
n
i
m
i
z
e

j
{X
)
:-99O
xJ-900x
250
2-5

S
u
b
j
e
c
tt
o
: g
.
(
X
)
:0
.
4
x
.+0
.6x
)=
8
.
5
2+x

(
3
.
5
)
(
3
.
6
)

x
.-X
2+X
4=
25
g
2
(
X
)
:3

(
3
.
7
)

s=
70
g
3
(
X
)
:3
x
I+6x
2+X

(
3
.
8
)

x
.~O; x2~0; X3~0; X4~0; x5~0

(
3
.
9
)

I
nF
i
g
u
r
e3
.
6t
h
ec
o
n
s
t
r
a
i
n
t
sx
.~ 0a
n
dX
2~ 0a
r
ea
d
d
c
dt
ot
h
et
h
r
e
ef
u
n
c
t
i
o
n
a
l
c
o
n
s
t
r
a
i
n
tI
in
e
s(
3
.
6
)
(
3
.
8
)
.Theh
a
.
'
!
h
e
da
r
e
ai
st
h
cf
e
a
s
i
b
l
er
e
g
i
o
na
ti
st
h
d
e
s
i
g
n

112 L
lNEARPOGRAMMING

3
.
2 GRAPHICALSOLUTION

113

)
=

50

40

Th
i
swouldy
i
e
l
d1
0p
o
i
n
t
sf
o
ro
u
re
x
n
p
l
e
whichi
sshowni
nF
i
g
u
r
e3
.
6
.

30

BaslcV
a
r
l
a
b
l
e
s
:Thes
e
to
fv
a
r
i
a
b
l
e
si
nt
h
eb
a
s
i
cs
o
l
u
t
i
o
nt
h
a
th
a
v
en
o
n
z
e
r
ov
a
l
u
e
s
a
r
ec
a
l
l
e
db
a
s
i
cv
a
r
i
a
b
l
e
s
.C
o
r
r
e
s
p
o
n
d
i
n
g
l
y
t
h
es
e
to
fv
a
r
i
a
b
l
e
si
nt
h
eb
a
s
i
cs
o
l
u
t
i
o
n
h
a
v
et
h
ev
a
l
u
eo
fz
e
r
oa
r
ec
a
l
l
e
dn
o
n
b
a
s
i
cv
a
r
i
a
b
l
e
s
.Fort
h
ep
o
i
n
tPl
XJ d
X2
a
r
en
o
n
b
a
s
i
cv
a
r
i
a
b
l
e
sw
h
i
l
eX3
X4

andxsa
r
b
a
s
i
cv
a
r
i
a
b
l
e
s
.

!
i

20

10

BasicF
e
a
s
l
b
l
eS
o
l
u
t
l
o
n
:Thi
si
sab
a
s
i
cs
o
l
u
t
i
o
nt
h
a
ti
sa
l
s
of
e
a
s
i
b
l
e
.
'
he
s
ea
r
e

ep
o
i
n
t
sPl
P
2
P3
andP4i
nF
i
g
u
r
c3
.
6
.I
nLP
t
h
es
o
l
u
t
i
o
nt
ot
h
ep
r
o
b
l
e
m

i
fi
ti
s

mustbeab
a
s
i
cf
e
a
s
i
b
l
es
o
l
u
t
i
o
n
.Theb
a
s
i
cs
o
l
u
t
i
o
nc
a
na
I
sobec
o
n
s
i
d
e
r
e
d
u
n
i
q
u
e
caI
l
ya
sac
o
r
n
e
rp
o
i
n
to
ra
ne
x
t
r
e
m
ep
o
i
n
to
ft
h
ef
e
a
s
i
b
l
er
e
g
i
o
n
.
geome

-3
0
0

1
0
1
5
y
p
oA
Numbor0
1M
a
c
h
l
n
0
80
1T

20

ConvexPolyhedron:Thi
si
saboundedr
e
g
i
o
no
ft
h
ef
e
a
s
i
b
l
eds
i
g
nspacee
r
c
g
i
o
nd
e
f
i
n
e
dby q
u
a
d
r
i
l
a
t
e
r
a
lc
o
m
p
r
i
s
i
n
gt
h
ep
o
i
n
t
sP
I
P2
P3
a
n
dP4i
nF
i
g
u
r
e
3
.
6
.Thet
e
r
mc
O
l
l
v
e
xs
e
tr
e
p
r
e
s
e
n
t
sac
o
l
l
e
c
t
i
o
no
fp
o
i
n
t
so
rv
e
c
t
o
r
sh
a
v
i
n
gt
h
e
i
fa
l
lo
f
t
h
ep
o
i
n
t
s
f
o
l
l
o
w
i
n
gp
r
o
p
e
r
t
y
:F
o
ra
n
ylwop
o
i
n
t
si
nt
h
es
c
t(
o
rw
i
t
h
i
nar
c
g
i
o
n
)
t
wop
o
i
n
t
sa
l
s
oI
i
ei
ncsamer
e
g
i
o
n
t
h
er
e
g
i
o
ni
sac
o
n
v
e
x
ont
h
el
i
n
cc
o
n
n
e
c
t
i
n
g
s
e
t
.I
m
a
g
i
n
yl
i
n
ed
r
l
1
wni
nl
h
cr
e
g
i
o
nt
l
f
i
n
e
t
lbyeq
u
a
d
r
i
l
a
t
e
n
dwhosecomer
P2
P
3
andP
4
.FromF
i
g
u
r
e3
.
6i
ti
sc
1
e
a
rt
h
el
i
n
ewi
l
Is
t
i
l
lb
ewi
i
n
p
o
i
n
t
sa
r
ePl
t
h
er
e
g
i
o
ne
s
t
a
b
l
i
s
h
e
dby qua
iJa
t
e
r
-m
akingi
taconvexs
e
t
.

2
5

F
l
g
u
r
e3
.
6F
e
a
s
i
b
l
er
e
g
i
o
nE
x
a
m
p
l
e3
.
1
.
s
p
a
c
ei
nwhicha
l
lo
ft
h
ec
o
n
s
t
r
a
i
n
t
sa
r
es
a
t
i
s
f
i
e
d
.Thec
i
r
c
l
e
se
x
a
g
g
e
r
a
t
et
h
ep
o
i
n
t
so
f
i
n
t
e
r
s
e
c
t
i
o
no
ff
i
v
ec
o
n
s
t
r
a
i
n
t
st
a
k
e
ntwoa
tat
i
m
e
.Manyo
fthema
r
enumberedPl
P
2
.
.
.

P
6
.A
I
Io
ft
h
e
s
ep
o
i
n
t
so
fi
n
t
e
r
t
i
o
nc
a
nb
ea
s
s
o
c
i
a
t
e
dw
i
t
hac
e
r
t
a
i
np
r
o
p
e
r
t
y
concemingt
h
ev
a
l
u
e
so
ft
h
ev
a
r
i
a
b
l
e
s
.N
o
t
et
h
e
r
ea
r
ef
i
v
ed
c
s
i
g
nv
a
r
i
a
b
l
e
s(n=
5
)a
n
d
t
h
r
e
cf
u
n
c
t
i
o
n
a
lc
o
n
s
t
r
a
i
n
t
s(m=3
)
.F
o
re
n
p
l
c
:

OptlmumS
o
l
u
t
l
o
n
:This i
s af
e
a
s
i
b
l
es
o
l
u
t
i
o
n a
tm
i
n
i
m
i
z
e
st
h
eo
b
j
e
c
t
i
v
e
f
u
n
c
t
i
o
n
p
o
i
n
tP4
i
nF
i
g
u
r
e3
.
6
.I
nLP
t
h
eoptimums
o
l
u
t
i
o
nmustb
eab
a
s
i
c
f
e
a
s
i
b
l
es
o
l
u
t
i
o
n
.

P
I
:(
x1=
0
x
2=
0
X)=
8

5
.x
25
Xs
=70)
4=

B
a
s
l
s
:'
h
eb
a
s
i
s P s
e
n
t
scc
o
l
u
m
n
so
ft
h
ec
o
e
f
f
i
c
i
e
n
tm
a
t
r
i
xA coespond
e
ya
r
e
t
ot
h
eb
a
s
i
cv
a
r
i
a
b
l
e
s
.Theyf
o
r
mt
h
cb
a
s
i
so
ft
h
em
d
i
m
e
n
s
i
o
na
Is
p
a
c
e
.Th
t
e
r
m
e
dt
h
eb
a
s
i
sv
e
c
t
o
r
s
.

P
2
:(
x1=
0
~=
11
.67

X)=
1
.5
x
3
6
.
6
7
Xs
=0)
4=
P
5
:(
xJ=
1
0
.
7
x
2=
7
.
0
5
X)=
0
x
0
Xs
=-4.4)
4=
P
6
:(
x1=
0
x
1
4
.
1
7
x
0
x
3
9
.
1
7
X
s=1
5.
4
2
)
2=
4=
3=

hr
YY

BaslcS
o
l
u
t
l
o
n
:Ab
a
s
i
cs
o
l
u
t
i
o
ni
so
n
eo
b
t
a
i
n
e
dbys
e
t
t
i
n
gc
x
a
c
t
l
yn-mv
a
r
i
a
b
l
e
s
t
oz
e
r
o
.I
nF
i
g
u
r
e3
.
6
a
I
lo
ft
h
ep
o
i
n
t
si
d
e
n
t
i
f
i
e
dbyt
h
ec
i
r
c
l
e
sr
c
p
r
c
s
c
n
tb
a
s
i
c
f
o
rnd
e
s
i
g
n
s
o
l
u
t
i
o
n
s
.Thep
o
i
n
t
sc
h
o
s
c
na
b
o
v
ea
r
ea
l
lb
a
s
i
cv
a
r
i
a
b
l
e
s
.I
ng
e
n
e
r
a
l
o
n
s
i
n
t
s numhero
fb
a
s
i
cs
o
l
u
l
i
o
n
si
sg
i
v
c
nbyt
h
ec
o
m
b
i
n
a
t
i
o
n
v
a
r
i
a
b
l
e
sa
n
dmc

/
/
J
6
4
2
2
p
l
z/4124

a
r
i
a
b
l
e
sa
r
eo
b
t
a
i
n
e
da
se
i
n
t
e
t
i
o
no
f
t
h
ec
o
n
s
t
r
a
i
n
t
st
a
k
e
ntwo
Thev
a
l
u
e
so
fev
a
tat
i
m
e
.Inea
b
o
v
eI
is

tf
o
re
a
c
hp
o
i
n
t
e
x
a
c
t
l
ytwoo
ft
h
ev
a
r
i
a
b
l
e
sa
r
ez
e
r
o
.The
.nr
e
p
r
e
s
e
n
t
st
h
enumbero
fv
a
r
i
a
b
l
e
sa
n
d
number2c
o
r
r
e
s
p
o
n
d
st
ot
h
ev
a
l
u
eo
fn-m
mt
h
enumbero
fc
o
n
s
t
r
a
i
n
t
s
.P
o
i
n
t
sP5a
n
dP6a
r
ei
n
f
e
a
s
i
b
l
eb
e
c
a
u
s
eoneo
ft
h
e
v
a
r
i
a
b
l
e
sh
a
san
e
g
a
t
i
v
ev
a
l
u
e
.

CanonlcalForm:'
heb
a
s
i
sv
e
c
t
o
r
d
uc
t
ou
n
i
tv
e
c
t
o
r
st
h
r
o
u
g
hr
o
w
/
c
o
l
u
m
n
eb
a
s
i
cf
e
a
t
u
r
eo
fen
u
m
e
r
i
c
a
l
o
p
e
r
a
t
i
o
n
s(
o
rG
a
u
s
s
J
o
r
d
a
ne
I
im
i
n
a
t
i
o
n
)
. Th
e
t
i
t
i
v
ep
r
o
c
e
d
es
t
a
r
t
i
n
gfroma
ni
n
i
t
i
a
l
t
e
c
h
n
i
q
u
ef
o
rLP(
S
i
m
p
l
e
xm
e
t
h
o
d
)i
sa p
b
a
s
i
cf
e
a
s
i
b
l
es
o
l
u
t
i
o
n
a
n
dd
e
t
e
r
m
i
n
i
n
geb
e
s
tn
e
i
g
h
b
o
r
i
n
gb
a
s
i
cf
e
a
s
i
b
l
es
o
l
u
t
i
o
n
o
n
i
s
t
h
a
ti
m
p
r
o
v
e
st
h
eo
b
j
e
c
t
i
v
e
.Thep
r
o
c
e
d
u
r
ei
sc
a
r
r
i
e
donu
n
t
i
lt
h
eoptimums
o
l
r
e
a
c
h
d

o
ri
fi
ti
sd
e
t
e
r
m
i
n
e
d nos
o
l
u
t
i
o
ni
sp
o
s
s
i
b
l
e
.Thcc
a
n
o
n
i
c
a
lf
o
r
mi
su
s
e
d
f
o
rr
a
p
i
d
1yi
d
e
n
t
i
f
y
i
n
gt
h
es
o
l
u
t
i
o
n
.Eachi
t
e
r
a
t
i
o
ni
n p
r
o
c
e
d
u
r
ec
a
nb
ed
e
s
c
r
i
b
e
d
m
a
t
h
e
m
a
t
i
c
a
Il
ya
sf
o
l
l
o
w
s
.
Th
es
t
a
r
t
i
n
gc
o
n
s
t
r
a
i
n
t
s'co
r
g
a
n
i
z
e
da
s
[
A
]
I
I
[ 1=[
b
]

A
f
t
e
rt
h
eG
a
u
s
s
J
o
r
d
a
ne
l
i
m
i
n
a
t
i
o
n
Eq~Jation (
3
.
2
0
)i
sa
s
s
e
m
b
l
e
da
s

(
3
.
2
0
)

114

3
.
3N
U
M
E
R
I
C
A
LS
O
L
U
T
I
O
N
T
H
ES
I
M
P
L
E
XMETHOD 1
1
5

L
lN
E
A
APAOGAAMMING

p
o
i
n
ti
sn
o
tf
e
a
s
i
b
l
ew
i
t
hr
e
s
p
e
c
tt
oec
o
n
s
t
r
a
i
n
t8
2Th
e
r
e
f
o
r
e

t
h
e
r
ei
snop
o
i
n
t
i
sf
e
a
s
i
b
l
e
.Th
e
r
ei
snos
o
l
u
t
i
o
nt
oep
r
o
b
l
e
m
.

m[X]m +[
R
] n
m
)
[
X
]
(
s
)=[
b
]
e
s
i
g
nv
a
r
i
a
b
l
e
si
nt
h
ef
i
r
s
tt
e
r
m t
h
eb
a
s
i
cv
a
r
i
a
b
J
c
s
.Th
es
c
to
f(nThes
e
to
fm d
m)d
c
s
i
g
nv
a
r
i
a
b
J
e
si
ncs
c
c
o
n
dt
c
r
ma
r
e n
o
n
b
a
s
i
cv
a
l
u
c
s
.I
ncS
i
m
p
J
e
xmethod
t
h
cn
o
n
b
a
s
i
cv
a
r
i
a
b
l
c
sa
r
es
u
m
m
a
r
i
l
ys
e
tt
oz
c
r
o
.

I
nt
h
ea
b
o
v
ed
i
s
c
u
s
s
i
o
n
. e t
w
o
v
a
r
i
a
b
l
es
i
t
u
a
t
i
o
n
sp
r
o
v
i
d
e
da
no
b
v
i
o
u
s
c
Ja
s
s
i
f
i
c
a
t
i
o
no
ft
h
es
o
l
u
t
i
o
n
s
.I
np
r
a
c
t
i
c
e
.LPm
o
d
e
l
sa
r
el
a
r
g
ew
i
t
ho
v
e
rh
u
n
d
r
e
d
so
f

v
a
r
i
a
b
J
c
s
.M
o
d
e
l
i
n
ga
n
dt
r
a
n
s
c
r
i
p
t
i
o
ne
r
r
o
r
smayg
i
v
er
i
s
et
omanyo
ft
h
ea
b
o
v
e
s
i
t
u
a
t
i
o
n
s
.Th
eg
e
n
e
m
t
i
o
no
fs
o
l
u
t
i
o
ni
sb
a
s
e
donn
u
m
e
r
i
c
a
lt
c
c
h
n
i
q
u
c
so
fl
i
n
e
a
r
r
a
whichi
so
f
t
e
ns
e
n
s
i
t
i
v
et
oeq
u
a
l
i
t
yo
ft
h
em
a
t
r
i
xo
fc
o
e
f
f
i
c
i
e
n
t
s
.F
i
J
t
e
r
i
n
g
a
l
g
e
b
o
u
te
r
r
a
n
td
a
t
ai
su
s
u
a
l
l
yas
i
g
n
i
f
i
c
a
n
te
x
e
r
c
i
s
ei
nt
h
es
e
a
r
c
ho
fo
p
t
i
m
a
ls
o
l
u
t
i
o
n
s
.

3.
2
.
3 Di
ntSol
onTyp

Th
c
r
e
a
a
tmostf
o
u
rd
i
f
f
e
r
e
n
tr
e
s
u
J
t
s a
tc
a
nb
ee
x
p
c
c
t
e
df
o
rt
h
es
o
J
u
t
i
o
no
ft
h
eLP
p
r
o
b
J
e
m
:(
1
)au
n
i
q
u
es
o
l
u
t
i
o
n
(
2
)i
n
f
i
n
i
t
e
J
ymanys
o
J
u
t
i
o
n
s
(
3
)unboundeds
o
l
u
t
i
o
n
and(
4
)t
h
ep
o
s
s
i
b
i
l
i
t
ya
tt
h
e
r
ei
snos
o
J
u
t
i
o
n
.

THESIMPLEXMETHOD
3
.
3 NUMERICALSOLUTION-

Th
es
t
a
n
d
a
r
dn
u
m
e
r
i
c
a
lp
r
o
c
e
d
u
r
ei
sb
a
s
e
donea
l
g
o
r
i
t
h
mduet
oO
a
n
t
z
i
g
m
e
n
t
i
o
n
e
de
a
r
l
i
e
r
.I
ti
sr
e
f
e 1
0a
seS
i
m
p
l
e
xm
e
t
h
o
d
.Thep
r
o
c
e
d
u
r
ei
sr
e
l
a
t
e
d
t
o s
o
l
u
t
i
o
no
fas
y
s
t
e
mo
fI
ineare
q
u
a
t
i
o
n
s
.Th
ea
c
t
u
a
la
p
p
l
i
c
a
t
i
o
no
f
p
r
o
c
e
d
u
r
ec
a
nbea
s
s
o
c
i
a
t
e
dw
i
t
ht
h
eGauss-Jordanmethodf
r
o
ml
i
n
e
a
ra
l
g
e
b
r
a
.
wheret
h
ec
o
e
f
f
i
c
in
trowsa
r
et
r
a
n
s
f
o
r
m
e
dt
h
r
o
u
g
he
1ementarym
u
l
t
i
p
l
i
c
a
t
i
o
na
n
d
a
d
d
i
t
i
o
n
.Mostmainframecomputcri
n
s
t
a
l
l
a
t
i
o
n
su
s
u
a
l
l
yc
a
r
r
ys
o
f
t
w
a
r
et
h
a
tw
i
l
l
h
e
l
ps
o
l
v
eLPp
r
o
b
l
e
m
s
.MATLAsa
l
s
op
r
o
v
i
d
e
sp
r
o
c
e
d
u
r
e
s1
0s
o
l
v
eLPp
r
o
b
l
e
m
s
h
eSimplxmethodi
sa
p
p
l
i
e
dt
o
i
ni
t
sO
p
t
i
m
i
z
a
t
i
o
nT
o
o
J
b
o
x
.I
nt
h
i
ss
c
c
t
i
o
nt
s
i
m
p
l
ep
r
o
b
lm p
r
i
m
a
r
i
l
yt
uu
n
d
c
r
s
t
u
n
ut
h
cpl'Ogrumminga
n
dg
e
o
m
c
t
r
i
cf
e
a
t
u
r
e
s
.
Thi
sw
i
l
lb
eu
s
e
dwhenwed
i
s
c
u
s
sd
i
r
e
c
It
e
c
h
n
i
q
u
e
sf
o
rn
o
n
l
i
n
e
a
rp
r
o
b
l
e
m
s
.I
nt
h
e
n
e
x
ts
e
c
t
i
o
nt
h
eSimplexmethodi
si
n
t
r
o
d
u
c
e
di
nd
e
t
a
i
Jw
i
t
he
x
p
l
a
n
a
t
i
o
n
s
.I
n
s
u
b
s
e
q
u
e
n
ts
e
c
t
i
o
n
s
MATLAB.o
rs
p
r
e
a
d
s
h
e
e
tp
r
o
g
r
a
m
si
IkeExcelcana
l
s
obeu
s
e
d
t
oimplementt
h
eSimpJexm
e
t
h
o
d
.

(
1
)U
n
i
q
u
es
o
l
u
t
i
o
n
: l
eexampleu
s
e
df
o
rd
i
s
c
u
s
s
i
o
nh
a
saUsi
q
u
es
o
l
u
t
i
o
n
(
F
i
3
.
5
)
. l
ec
o
n
d
i
t
i
o
nn
e
c
e
s
s
yf
o
ri
st
oo
c
c
u
ri
s eo
b
j
e
c
t
i
v
e
f
u
n
c
t
i
o
na
n
dt
h
ec
o
n
s
t
r
a
i
n
t
sh
a
v
ed
i
s
s
i
m
i
l
a
rs
J
o
p
e
s
andt
h
ef
e
a
s
i
b
J
er
e
g
i
o
ni
s
b
o
u
n
d
c
d
l
c
Jo
s
e
d
.G
e
o
m
e
t
r
i
c
a
l
l
y
t
h
i
sc
a
nb
ev
i
s
u
a
l
l
ye
x
p
J
a
i
n
e
da
st
h
emovement
o
fel
i
n
er
e
p
r
e
s
c
n
t
i
n
gt
h
co
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
n
p u
l
l
c
J1
0i
t
s
e
J
f

i
naf
a
v
o
r
a
b
J
e
d
i
r
e
c
t
i
o
n

u
n
t
i
li
tr
e
m
a
i
n
sj
u
s
ti
nc
o
n
t
a
c
tw
i
t
honeo
ft
h
ef
e
a
s
i
b
J
ecomers(
b

i
c
i
b
J
er
e
g
i
o
ni
si
d
e
n
t
i
f
i
e
d
f
e
a
s
i
b
J
es
o
J
u
t
i
o
n
)o
ft
h
ef
e
a
s
i
b
l
er
e
g
i
o
n
.Thef
e
3
.
6
.A s
i
m
p
l
r
c
i
s
cu
fi
m
a
g
i
n
i
n
gd
i
f
f
e
o
b
j
e
c
t
i
v

u
i
s
t
i
n
c
t
l
yi
nF
i
g
f
u
n
c
t
i
o
nJ
i
n
e
ss
h
o
u
l
dc
o
n
v
i
n
c
et
h
ev
i
e
w
e
rt
h
a
ts
e
v
e
r
a
ld
i
f
f
e
r
e
n
to
b
j
e
c
t
i
v
e
s n

P
2
P3
o
rP
4
.T
h
i
s20geome

b
ed
e
f
i
n
e
dt
ol
o
c
a
t
et
h
eu
n
i
q
u
es
o
l
u
t
i
o
na
tP1
andc
o
n
s
t
r
u
c
t
i
o
nc
a
nbeu
s
e
dt
ou
n
d
e
r
s
t
a
n
dt
h
ee
x
t
e
n
s
i
o
nt
ond
i
m
e
n
s
i
o
n
s
.He

t
h
ei
In
e
swi
l
Jb
er
e
p
r
e
s
e
n
t
e
dbyl
I
y
p
e
r
p
l
a
n
e
s
.

(
2
)l
n
f
i
n
i
l
es
o
l
u
t
i
o
n
:
I
no
r
d
e
rf
o
ri
st
oo
c
c
u
r
eo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
nmustb
e
p
a
r
a
l
l
e
lt
ooneo
ft
h
ec
o
n
si
n
t
s
.F
o
re
x
a
m
p
l
e
i
nF
i
g
u
r
e3
.
6
l
e
tt
h
eo
r
i
g
i
n
a
l
h
eo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
ni
sp
a
r
a
l
l
e
lt
ot
h
ec
o
n
s
t
r
a
i
n
t
p
r
o
b
l
e
mber
e
d
e
f
i
n
e
ds
o t
h
Anyp
o
i
n
tont
h
ec
o
n
s
t
r
a
i
n
t

a
n
dl
y
i
n
gb
e
t
w
e
e
nt
h
el
i
n
es
e
g
m
e
n
td
c
f
i
n
e
dby
ep
o
i
n
t
sp2a
n
dP4

i
sa
no
p
t
i
m
a
ls
o
l
u
t
i
o
nt
ot
h
ep
r
o
b
l
e
m
.Th
e
yw
i
l
ly
i
e
J
de
s
a
r
n
ev
a
l
u
eo
f v
e
.

(
4
)Nos
o
l
u
l
i
o
n
:
F
i
g
u
r
e3
.
6i
su
s
e
dt
oe
x
p
l
a
i
ni
sp
o
s
s
i
bi
I
it
y
.C
o
n
s
i
d
e
r E

)
.Thef
e
a
s
i
b
l
c
d
i
r
e
c
t
i
o
no
fi
n
e
q
u
a
l
i
t
yi
n8
1i
sc
h
a
n
g
e
dt
ot
h
co
p
p
o
s
i
t
et
y
p
e(
c
s
p
e
c
tt
ot
h
ec
o
n
s
t
m
i
n
t
sg
la
n
d8
3i
st
ot
h
er
i
g
h
to
fX
I J
5
.
0
.T
h
i
s
r
e
g
i
o
nw
i r

I
nt
h
i
ss
e
c
t
i
o
n
.t
h
emachines
e
J
e
c
t
i
o
ne
x
n
p
J
ei
n
t
r
o
d
u
c
e
de
a
r
J
i
e
ri
ss
e
tupf
o
ra
p
p
l
y
i
n
g
o
d
.I
ti
si
n
s
t
r
u
c
t
i
v
et
on
o
t
ct
h
emethodi
si
t
e
r
a
t
i
v
e
.Givenas
t
a
r
t
i
n
g
t
h
eS
i
m
p
l
e
xmc

ti
tw
i
l
lmarchf
o
r
w
a
r
dt
h
r
o
u
g
hi
m
p
r
o
v
i
n
gd
e
s
i
g
n
sun
i
th
a
sf
o
u
n
dt
h
es
o
h
o
n
p
o
i
n
o
rc
a
n
n
o
tp
r
o
c
e
e
df
u
r
t
h
e
r
.F
o
rc
o
m
p
l
e
t
e
n
e
s
st
h
eo
r
i
g
i
n
a
lp
r
o
b
l
e
mi
sr
e
w
r
i
t
t
e
nh
e
r
e
:

4
5
F
i
r
2TIz-

o
l
l
l
l
i
o
n
: I
n
i
sc
a
s
et
h
ef
e
a
s
i
b
l
er
e
g
i
o
ni
sn
o
tb
o
u
n
d
e
d
.I
n
(
3
)Unboundeds
F
i
g
u
r
e3
.
6
i
ft
h
ec
o
n
s
t
r
a
i
n
t
sg
la
n
dg2w
e
r
en
o
tp
a
r
to
ft
h
ep
r
o
b
J
e
mf
o
r
m
u
J
a
t
i
o
n
e
nt
h
ef
e
a
s
i
b
J
er
e
g
i
o
ni
sn
o
tboundont
h
et
o
p
.R
e
f
e
r
r
i
n
gt
oF
i
g
u
r
c3
.
5
.E
i
v
ef
u
n
c
t
i
o
nc
a
nbes
h
i
f
t
e
d

u
n
l
i
l
T
Ut
e
dh
i
g
h
e
rv
a
l
u
e
s
.I
np
r
a
c
t
i
c
e
t
h
e
r
e
o
b
j t
a
n
g
eo
ft
h
ed
e
s
i
g
nv
a
r
i
a
b
J
e
s(
n
o
tp
t
o
f
s
t
a
n
d
d
w
i
l
lb
ea
nu
p
p
e
rboundon r
f
o
r
m
a
t
)t
h
a
twi
l
Jb
eu
s
e
dt
oc
Jo
s
ei
sr
e
g
i
o
n
.i
nwhichc
a
s
es
o
l
u
t
i
o
n
so
f
t
y
p
e(
1
)
ep
r
c
s
e
n
c
eo
fa
nunboundeds
o
l
u
t
i
o
na
1s
os
u
g
g
c
s
t
s
o
r(
2
)c
a
nb
er
e
c
o
v
e
r
e
d
.Th
tt
h
ef
o
r
m
u
l
a
t
i
o
no
f p
r
o
b
l
e
mmayb
el
a
c
k
i
n
g
.A
d
d
i
t
i
o
n
a
lm
e
a
n
i
n
g
f
u
l
c
o
n
s
t
r
a
i
n
t
(
s
)c
b
eaccommodatedt
od
e
f
i
n
et
h
es
o
J
u
t
i
o
n
.

3
.
3
.
1 FeaturesoftheSlmplexMethod

(
X
)
:9
90xl+9
0
O
x
250
Maximizc j
2+5

(
3
.
1
)

4
x
l+0
.
6
X
2:
s
;8.
5
S
u
b
j
e
c
tt
o
: 8
1
(
X
)
:O.

(
3
.
2
)

g
2
(
X
)
:3
x
I-X
2:
s
;25

(
3.
3
)

x
S
;7
0
8
3
(
X
)
:3xI+6
2:

(
3.
4
)

x1~O;

x2~0

Thep
r
o
b
l
e
mwasa
J
s
ot
r
a
n
s
f
o
r
m
e
d1
0t
h
es
t
a
n
d
a
r
df
o
r
m
a
ta
sf
o
l
l
o
w
s
.

116 L
lN
E
A
RPROGRAMMING

T
lO
NT
H
ES
I
M
P
L
E
XMETHOD 117
3
.
3N
U
M
E
R
I
C
A
LS
O
L
U

Ex
ample3.1

Witht
h
i
si
nmind
t
h
eS
i
m
p
l
e
xmethodi
sa
p
p
l
i
e
dt
ot
h
emachines
e
l
e
c
t
i
o
np
r
o
b
l
e
m
.
Th
cexamplew
i
l
lu
s
ea
n
dc
Ja
r
i
f
yt
h
ei
t
e
m
si
nt
h
ea
b
o
v
eI
is
.
t

J
(
X
)
:99
O
xI- 90
O
x
250
2- 5

M
i
n
i
m
i
z
e

S
u
b
j
e
c
tt
o
: g(
X
)
:0
.
4
x
l+0
.6x2+x
3 8
.
5

(
3
.
5
)
(
3
.
6
)

g
2
(
X
)
:3
x -x2+X4=25

(
3
.
7
)

g
3(
X
)
:3
x
I+6x
s 70
2+X

(
3
.
8
)

;x
;x
; Xso
x10
2~O; x
30
4 0

(
3
.
9
)

X
4
andX
sa
r
es
l
a
c
kv
a
r
i
a
b
l
e
s
.
X
3
TheS
i
m
p
l
e
xmethodi
su
s
e
dont
h
ep
r
o
b
l
e
mb
e
i
n
ge
x
p
r
e
s
s
e
di
nt
h
es
t
a
n
d
a
r
d
f
o
r
m
a
t
.
T
h
e
f
o
l
l
o
w
i
n
gi
n
f
o
r
m
a
t
i
o
ni
su
s
e
a
l
li
no
r
g
a
n
i
z
i
n
gt
h
ec
a
l
c
u
l
a
t
i
o
na
sw
e
l
la
s
c
o
g
n
i
z
i
n
gt
h
em
o
t
i
v
a
t
i
o
nf
o
rs
u
b
s
e
q
u
e
n
ti
t
e
r
a
t
i
o
n
s
.1
e
s
er
e
f
c
rt
oap
r
o
b
l
e
mf
o
r
whichau
n
i
q
u
es
o
l
u
t
i
o
ne
x
i
s
t
s
.Wh
i
l
em yo
ft
h
ei
t
e
m
sb
c
l
o
ww
c
r
ci
n
d
u
c
e
d
t
h
e
ya
r
er
e
f
e
r
e
n
c
e
dh
e
r
ef
o
rc
o
m
p
l
e
t
e
n
e
s
s
.
e
a
r
l
i
e
r
. i
n
c
l
u
d
e
st
h
es
l
a
c
kands
u
r
p
l
u
s
Thcnumbero
fv
a
r
i
a
b
l
e
si
nt
h
cp
r
o
b
l
e
mi
sn
v
a
r
i
a
b
l
e
s
.

3
.
3
.
2 AppllcatlonofSlmplexMethod

SimplexTable3
.
1
:I
nTable3.
1(SimplexTnble1)cf
i
r
s
trowi
n
d
i
c
a
t
c
st
h
e
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
n
.S
p
r
e
a
d
s
h
e
e
t
sa
r
ea
ne
f
f
i
c
i
e
n
t
v
a
r
i
a
b
l
en
a
m
e
s
.Thel
a
s
trowi
seo
wayt
op
r
o
s
st
h
ei
n
f
o
n
n
a
t
i
o
n
. i
ti
sb
e
i
n
gcmployed
e
n symbolfc
a
nb
e
a
s
tcolumna
n
dl
a
s
trowt
oa
l
l
o
wn
u
m
e
r
i
c
a
lc
a
1
c
u
l
a
t
i
o
n
s(
s
h
o
w
n
)
.
removedfromel
The
l
n
s
t
c
o
l
u
D1
n
i
s r
i
g
h
t
n
n
d
s
i
d
v
a
l
u
e
s
.Th
er
e
s
to
ft
h
cc
n
t
r
i
e
s t
h
ec
o
e
f
f
i
c
i
e
n
t
s
a
i
n
te
q
u
a
t
i
o
n
s
.
o
f
t
h
ec
o
n
s
Th
ec
u
r
r
e
n
ti
t
c
r
a
t
i
o
ni
so
v
e
ri
ft
h
et
a
b
l
ed
i
s
p
l
a
y
st
h
ec
a
n
o
n
i
c
a
lf
o
n
n
.I
np
r
a
c
t
i
c
ee
n
i
tv
c
c
t
o
r
si
nt
h
et
a
b
l
e
a
sw
e
l
l making
c
a
n
o
n
i
c
a
lfonnc
o
m
p
r
i
s

s
p
o
t
t
i
n
gt
h
emu
s
u
r
et
h
ee
n
e
su
n
d
e
rt
h
ebcolumn
e
x
c
e
p
tf
o
rt
h
er
o
w
sr
e
p
r
e
s
e
n
t
i
n
geo
b
j
e
c
t
i
v
e
r
en
o
n
n
e
g
a
t
i
v
eP
Q
)
.A glancea
tS
i
m
p
l
e
xT
a
b
l
e3.
1i
n
d
i
c
a
t
e
s tE
f
u
n
c
t
i
o
n
(
s
)a
c
a
n
o
n
i
c
a
lfonni
sp
r
e
s
e
nt
.
Thcu
n
i
tv
e
c
t
o
r
si
nct
a
b
l
ea
l
s
oi
d
c
n
t
i
f
yt
h
o
s
emv
a
r
i
a
b
l
c
sa
tw
i
l
lb
e
l
o
n
gt
ot
h
e
b i
s
.Th
o
s
ev
a
r
i
a
b
l
e
sw
i
l
lh
a
v
ean
o
n
z
e
r
ov
a
l
u
ef
o
ri
si
t
e
r
a
t
i
o
n
.Ther
e
m
a
i
n
i
n
g

n-m
n
o
n
b
a
s
i
cv
a
r
i
a
b
l
e
sa
r
es
c
tt
oz
e
r
o
.T
h
i
ss
o
l
u
t
i
o
ni
sd
i
r
e
c
t
l
yi
n
t
c
r
p
r
e
t
e
domTable
.Thes
o
l
u
t
i
o
nt
h
e
r
e
f
o
r
ei
s
3
.1

1
enumbcro
fc
o
n
s
t
r
a
i
n
t
si
sm(m<n
)
.

x
z

x =0
.
0
=0
.
0=8
.
5
x
5Xs
=70d
f=-5250
4=2

1
ep
r
o
b
l
e
mi
sa
1
wayst
om
i
n
i
m
i
z
et
h
eo
b
j
e
c
t
i
v
cf
u
n
c
t
i
o
n
f

s
a
m
ea
s numbero
fc
o
n
s
t
r
a
i
n
t
s
)
.
Thenumbero
fb
a
s
i
cv
a
r
i
a
b
l
e
si
sm (

Thenumberofnonbasicvariablcsisn-m.

Thes
e
to
fu
n
i
tv
e
c
t
o
r
se
x
p
e
c
t
e
di
se
" e2 em
Thecolumno
ft
h
er
i
g
h
t
h
a
n
ds
i
d
e(
b
)i
sp
o
s
i
t
i
v
ea
n
dg
r
e
a
t
e
r a
no
re
q
u
a
lt
o
z
e
r
o
.

Th
ec
a
1c
u
l
a
t
i
o
n
sa
r
eo
r
g
a
n
i
z
e
di
nat
a
b
l
e

w
h
i
c
hi
sc
a
l
l
e
dat
a
b
l
e
a
u
o
rS
i
m
p
l
c
x
tw
i
l
lb
er
e
f
e
r
r
e
dt
o t
h
et
a
b
l
ei
nt
h
i
sb
o
o
k
.
t
a
b
l
e
a
u
.I

Onlytheva1uesofthcc

i
c
i
e
n
t
sa
r
n
e
c
e
s
s
yf
o
rt
h
ec
a
l
c
u
l
a
t
i
o
n
s
.Th
et
a
b
l
e

crefore contains only coefficicnt va


1u
e
st
h
em
a
t
r
i
x[
A
]i
np
r
e
v
i
o
u
s

h
ec
o
e
f
f
i
c
i
e
n
t
si
nt
h
ec
o
n
s
t
r
a
i
n
te
q
u
a
t
i
o
n
s
.
d
i
s
c
u
s
s
i
o
n
s
.These t

Hast
h
eO
p
t
i
m
a
lS
o
l
u
t
i
o
nBeenOb
i
n
e
d
? Theo
p
t
i
m
n
ls
o
l
u
t
i
o
nr
e
q
u
i
r
e
st
h
a
tf
b
er
e
d
u
c
e
d mucha
sp
o
s
s
i
b
l
e
.F
o
rt
h
ec
u
r
r
e
n
ts
o
l
u
t
i
o
n v
a
l
u
eo
fX
Ii
sO
.From
E
q
u
a
t
i
o
n(
3
.
5
)i
fx w
e
r
ei
n
c
r
e
a
s
e
df
r
o
mz
e
r
o
e
nt
h
eo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
n
fd
e
c
r
e
a
s
e
s
f
u
r
t
h
e
ri
nv
a
l
u
e
.S
i
m
i
l
a
rr
e
a
s
o
n
i
n
gc
n
nbeu
s
e
df
o
rX
2T
h
e
r
e
f
o
r
e
t
h
es
o
l
u
t
i
o
nc
a
nb
e
f
u
r
t
h
e
ri
n

p
r
o
v
e
d
.S
o
l
u
t
i
o
nh
n
sn
o
tc
o
n
v
e
r
g
e
d
.Th
es
e
c
o
nd/n
e
x
ti
t
e
r
a
t
i
o
nw
i
l
l q
u
i
r
e
T
a
b
l
e3
.
2
.
n
s
s
e
m
b
l
i
n
ga
n
o
t
h
e
rt
a
b
l
e
Theo
n
l
ywayf
o
rx a
nd/
o
rX
2t
oh
a
v
eap
o
s
i
t
i
v
ev
a
l
u
ei
si
fyb
a
meab
a
s
i
c
someo
f
v
a
r
i
a
b
l
e
.S
i
n
c
et
h
enumbero
fb
n
s
i
cv
a
r
i
a
b
l
e
si
sp
r
e
s
c
r
i
b
e
df
o
rt
h
ep
r
o
b
l
e
m
ec
u
r
r
e
n
tb
n
s
i
cv
a
r
i
a
b
l
e
smustbecomen
o
n
b
a
s
i
c
.Fore
a
c
hi
t
e
r
a
t
i
o
neS
i
m
p
l
e
x
methoda
1
l
o
w
so
n
l
yo
n
ep
a
i
ro
fv
a
r
i
a
b
l
c
st
oc
o
n
d
u
c
ti
sb
a
s
i
s
n
o
n
b
a
s
i
se
x
c
h
a
n
g
e
.
i
ti
sn
e
c
e
s
s
a
r
y
Eachi
t
e
r
a
t
i
o
nt
a
k
e
sp
l
a
c
et
h
r
o
u
g
hanewt
a
b
l
e
.Toa
f
f
e
c
tt
h
ee
x
c
h
a
n
g

1
eo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
ni
se
l
a
s
t
r
o
w
i
ne
t
a
b
l
e
.1
e
c
o
n
s
i
n
tc
o
e
f
f
i
c
i
e
n
t
sa
r
e

w
r
i
t
t
e
n

t
.
Rowo
p
e
r
a
t
i
o
n
sc
o
n
s
i
s
to
fn
d
d
i
n
g(
s
u
b
t
r
a
c
t
i
n
g
)ad
e
f
i
n
i
t
em
u
l
t
i
p
l
eo
ft
h
ep
i
v
o
t
rowt
oo
t
h
E
rr
o
w
so
f
t
h
e
t
a
b
l
e
.
T
h
ep
i
v
o
t
r
o
wi
d
e
n
t
i
f
i
e
st
h
erowi
nwhicht
h
e
u
n
i
t
v
c
c
t
o
r(
i
na
.
c
c
r
t
a
i
nc
o
l
u
m
n
)w
i
l
lh
a
v
ct
h
ev
a
l
u
eo
f1
.
Rowo
p
e
r
a
t
i
o
n
sa
r
en
e
e
d
c
dt
oo
b
t
a
i
nt
h
ec
a
n
o
n
i
c
a
lf
o
n
n
.P
r
a
c
t
i
c
a
l
l
yt
h
i
s p
l
i
c
s
y
ous
h
o
u
l
dbEablet
os
p
o
tt
h
E
u
n
i
tv
e
c
t
o
r
si
nt
h
et
a
b
l
e
.
I
ft
h
i
shappens
t
h
e
nt
h
ec
u
r
r
e
n
ti
t
e
r
a
t
i
o
ni
scomplt
eandc
o
n
s
i
d
e
r
t
i
o
n
sf
o
rt
h
en
e
x
tonc
mustb
e
g
i
n
.

T
a
b
l
e3
.
1S
i
m
p
l
e
xT
a
b
l
e3
.
1
X
I

0
.
4
3
3
9
9
0

X
2
0
.
6
1
6
9

X
J

X
4

X~

b
8
.
5
2
5
7
0
5
2
8
0

118

3
.
3 NUMERICALSOLUTIONTHESIMPLEXMETHOD

L
lNEARPROGRAMMING

t
od
e
t
e
r
m
i
n
ew
h
i
c
hi
st
h
ev
a
r
i
a
b
l
e w
i
l
le
n
t
e
rt
h
eb
a
s
i
s(
X
Io
rX2)
E
n
t
e
r
i
n
gB
a
s
i
c
V
a
r
i
a
b
l
e(
E
B
V
)
.A
l
s
ot
obee
s
t
a
b
l
i
s
h
e
di
swhicho
fev
a
r
i
a
b
l
(
X
3
'X4'o
rx
s
)w
i
l
l
l
e
a
v
e b
a
s
i
s

L
e
a
v
i
n
gB
a
s
i
cV
a
r
i
a
b
l
e(LBV
)
.I
nt
e
r
m
so
ft
h
en
u
m
e
r
i
c
a
lc
a
l
c
u
l
a
t
i
o
n
s
i
nt
h
enewt
a
b
l
e

t
h
eu
n
i
tv
e
c
t
o
ru
n
d
e
rt
h
eLBVcolumnwi
l
Ibet
r
a
n
s
f
e
r
r
e
dt
ot
h
eEBV
columnt
h
r
o
u
g
he
l
e
m
e
n
t
a
r
yr
o
w
/
c
o
l
u
m
no
p
e
r
a
t
i
o
n
s
.

E
n
t
e
r
i
n
g8
a
s
i
cV
a
r
i
a
b
l
e
:TheEBVi
sc
h
o
s
e
nb
yi
t
sa
b
i
l
i
t
yt
op
r
o
v
i
d
ef
o
rt
h
el
a
r
g
e
s
t
d
e
c
r
l
i
nt
h
o
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
ni
nec
u
r
r
e
n
ti
t
e
r
a
t
i
o
n
l
t
a
b
l
e
. l
i
si
sd
e
t
e
r
c
n
i
n
e
dby
c
i
e
n
ti
nt
h
erowa
s
s
o
c
i
a
t
e
dw
i
t
ht
h
eo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
n
.From
t
h
el
a
r
g
e
s
tn
e
g
a
t
iv
ec
o
ei
T
a
b
l
3
.
1
t
h
i
sv
a
l
u
ei
s-990andc
o
p
ondstoev
a
r
i
a
b
l
cX
IEBVi
sX!.A t
i
ec
a
n
beb
r
o
k
c
na
r
b
i
t
r
a
r
i
Jy
.
Leaving8
a
s
i
cV
a
r
i
a
b
l
e
:TheLBVi
sa
l
s
oc
h
o
s
e
nbyi
t
sa
b
i
l
i
t
y
l
p
r
o
v
ee
o
t
i
v
e
.Th
eLBVi
so
n
l
yd
e
t
e
r
m
i
n
e
da
f
t
e
rt
h
eEBVh
a
sb
e
e
nc
h
o
s
e
n
. i
s
s
e
x
a
I
n
e
dt
oe
s
t
a
b
I
is
h LBV.
i
m
p
o
r
t
a
n
tb
e
c
a
u
s
eo
n
l
yt
h
ecolumnu
n
d
e
reEBVi
'
a
t
i
oo
ft
h
ev
a
l
u
e
so
fer
i
g
h
t
h
a
n
ds
i
d
e(bcolumnv
a
l
u
e
s
)t
ot
h
e
Theminimuml
c
o
r
r
c
s
p
o
n
d
i
n
gc
o
e
f
f
i
c
i
e
n
t
si
nt
h
eEBVcolumnp
r
o
v
i
d
e
dcc
o
e
f
f
i
c
i
c
n
tv
a
l
u
ci
s

c
i
d
c
seLBV.T
h
i
sr
n
i
n
i
m
u
mv
a
l
u
ci
d
c
n
t
i
f
i
e
st
h
eroww
h
i
c
hi
sa
I
s
ot
h
e
p
o
s
i
t
i
v
cd
p
i
v
o
tr
o
w
.I
nt
h
i
srow
t
h
ecolumn
whichc
o
n
t
a
i
n
st
h
eu
n
i
t
yv
a
l
u
eo
f someu
n
i
t
v
e
c
t
o
r(
o
rac
u
r
r
e
n
tb
a
s
i
cv
a
r
i
a
b
l
e
)
i
st
h
ccolumno
fcI
e
a
v
i
n
gb
a
s
i
cv
a
r
i
a
b
l
e
. l
i
s
i
stheLBV.
I
nTable3.1eEBVh
a
sa
l
r
e
a
d
yb
e
e
ni
d
e
n
t
i
f
i
e
da
sX
l
'Th
er
a
t
i
o
su
n
d
e
rt
h
i
scolumn
a
r
e21
.258
.
3
3
3
3and23.
33
3
3
.Th
el
e
a
s
tv
a
l
u
ei
s8
.
3
3
3
3
.lIsc
o
r
r
p
ondst
oe
Sc
ondrowo
fT
a
b
l
e3
.
1
.Theu
n
i
tv
e
c
t
o
ru
n
d
e
rt
h
eX4columni
si
ni
srow(
r
o
w2
)
.
seLBV des
e
c
o
n
drowi
st
h
ep
i
v
o
trowf
o
rT
a
b
l
e3
.
2
.T
h
i
s
T
h
c
r
e
f
o
r
eX4i
impI
i
e
s s
e
o
ndroww
i
l
lb
eu
t
i
l
i
z
e
dt
oc
h
a
n
g
c
u
r
r
e
n
tX
Icolumn(
E
BV
c
o
l
u
m
n
)i
i
l
t
oau
n
i
tv
e
c
t
o
r a
li
sc
u
r
r
e
n
t
l
yu
n
d
e
rX4
T
h
i
sm
e
c
h
a
n
i
c
a
lp
r
o
c
e
d
u
r
ei
s
b
a
s
e
dont
h
ef
o
l
l
o
w
i
n
gr
e
a
s
o
n
i
n
g
.l
ev
a
l
u
eofEBVmustb
es
u
c
ht
h
a
ti
tw
i
l
ln
o
tc
a
u
s
e
t
h
e
rb
a
s
i
cf
e
a
s
i
b
l
ev
a
r
i
a
b
l
c
s
.The
c
o
n
s
t
r
a
i
n
tv
i
o
l
a
t
i
o
nw
i
t
ht
h
ec
u
r
r
e
n
tv
a
l
u
e
so
feo
a
p
p
l
i
c
a
t
i
o
no
ft
h
eGauss-Jo
r
d
a
np
r
o
c
e
s
ss
h
o
u
l
ds
h
i
f
tt
h
ecolumnu
n
d
e
rt
h
eLBVt
oe
EBVl
e
a
v
i
n
gt
h
ecolumno
ft
h
eo
t
h
e
ru
n
i
tv
e
c
t
o
r
s(
t
h
a
ta
r
ep
a
r
to
ft
h
eb
a
s
i
s
)u
n
a
l
t
r
e
d
.
T
a
b
l
e3
.
2i
st
h
e
nc
o
n
s
t
r
u
c
t
e
du
s
i
n
gt
h
cp
i
v
o
trow2
.

25

Thev
a
l
u
eo
ft
h
ef
i
r
s
te
l
e
m
e
n
ti
nt
h
enewrowwi
1
lb
e1
.Therowi
sm
o
d
i
f
i
e
db
y
d
i
v
i
d
i
n
gt
h
r
o
u
g
hby3
.Th
enews
e
c
o
n
drowi
nT
a
b
l
e3
.
2w
i
l
lb
e

0.3333

0.
33
33

3333
8.

I
nTable3
.
2

ef
i
r
s
te
l
e
m
e
n
to
ft
h
ef
i
r
s
trow(T
a
b
l
e3
.
1
)mustb
er
e
d
u
c
e
dt
oO
.T
h
i
s
i
sa
l
w
a
y
sa
c
h
i
e
v
e
dbya
d
d
i
n
g
/
s
u
b
t
r
a
c
t
i
n
g
f
r
o
mt
h
erowb
e
i
n
gm
o
d
i
f
i
e
dana
p
p
r
o
p
r
i
a
t
e
m
u
l
t
i
p
l
eo
f
t
h
ep
i
v
o
tr
o
w
.Them
o
d
i
f
i
c
a
t
i
o
nt
oef
i
r
s
trowi
s
row1-[
0
.
4"(
p
iv
o
trow2
)
}t
oy
i
e
l
d
0
.
7
3
3
3
1 0
.
1
3
3
3 0 5
.
1
6
6
7

Thet
h
i
r
droww
i
l
lbeo
b
t
a
i
n
e
dby
row3-{
3"
(
p
i
v
o
trow3
)
}
r
owwi
1
lbeo
b
t
a
i
n
e
dby
Thefour

row4+[
(
9
9
0"(
p
i
v
o
trow2
)I
T
a
b
l
e3
.
2i
so
b
t
a
i
n
e
da
f
t
e
rt
h
e
s
ec
a
l
c
u
l
a
t
i
o
n
s
.
c
i
e
n
ti
nT
a
b
l
e3
.
2i
nrow4i
n
d
i
c
a
t
e
s wen
e
e
d
Then
e
g
a
t
i
v
ev
a
l
u
eo
ft
h
ec
o
e
:
a
tl
e
a
s
tonemorei
t
e
r
a
t
i
o
n
l
t
a
b
l
e
.TheEBVi
sX2Byexamining p
o
s
i
t
i
v
er
a
t
i
o
so
f
ev
a
1u
e
si
n bcolumnt
ot
h
ec
o
e
f
f
i
c
i
e
n
ti
n X2column

t
h
eLBVi
si
d
t
i
f
i
e
da
s
X
s
Thep
i
v
o
trowi
srow3
.T
a
b
l
e3
.
3wi
1
lbec
o
n
s
t
r
u
c
t
e
dbyt
r
a
n
s
f
o
r
m
i
n
gt
h
er
o
w
so
f
T
a
b
l
e
3
.
2
.
SimplexTable3
.
3(
T
a
b
l
e3
.
3
)
p
i
v
o
trow3=(
r
o
w3
)

newrow1=row1-{
0
.
7
3
3
3
3 (
p
i
v
o
t
r
o
w3
)
}
0
.
3
3
3
3*
(
p
i
v
o
trow3
)
}
newrow2=row2+{
1
2
3
0(p
i
v
o
trow3
)
}
newrow4=row4+{

3i
so
b
t
a
i
n
e
db
a
s
e
dont
h
ea
b
o
v
eo
p
e
r
a
t
i
o
n
s
.
T
a
b
l
e3.

X4

X2

X)

nunU U

5
.
1
6
6
7
8
.
3
3
3
3
4
5
1
3
5
3
0

X
I

'

3
3
0

nunUEnu

nununu

0
.
1
3
3
3
0
.
3
3
3
3

X
s

nunu--AU

.
t
)
-

nu--nun

0
.
7
3
3
3
3
3
3
3
-0.
7
1
2
3
0

nu---nunu

X
l

SimplexTable3
.
2(T
able3
.
2
)
:Thep
i
v
o
trowi
so
b
t
a
i
n
e
df
r
o
mo
r
i
g
i
n
r
ow20f
.Th
ef
o
l
l
o
w
i
n
gi
s o
r
i
g
i
n
a
lrow
T
a
b
l
e3
.1

T
a
b
l
e
3
.
3S
i
m
p
l
e
xT
a
b
l
e3
.
3

i
m
p
l
e
xT
a
b
l
e3
.
2
2S
Table3.
X
I

119

X4

0
.
0
2
8
5
0
.
2
8
5
7
-0.
1
4
2
8
1
5
4
.
2
8

X
s

0
.
0
4
7
0
.
0
4
7
6
0
.
1
4
2
8
1
7
5
.
7
1

0
.
4
5
2
4
1
0
.
4
7
6
2
.
4
2
8
5
6
2
1
4
3
7.
14

1
2
0

L
lN
E
A
RP
R
O
G
R
A
M
M
I
N
G

T
a
b
l
e3.
3d
o
e
sn
o
th
a
v
e yn
e
g
a
t
i
v
ev
a
l
u
e eo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
nr
o
w
.A
s
o
l
u
t
i
o
nh
a
sb
e
e
na
c
h
i
e
v
e
d
.Th
eb
a
s
i
cv
a
r
i
a
b
l
e
sh
a
v
et
h
eu
n
i
tv
e
c
t
o
ri
nt
h
e
i
rc
o
l
u
m
n
.
Then
o
n
b
a
s
i
cv
a
r
i
a
b
l
e
sa
r
es
e
tt
oz
e
r
o
.FromT
a
b
l
e3
.
3t
h
eo
p
numv
a
l
u
e
sc
a
nb
o
b
s
e
r
v
e
d:

3
.
3N
U
M
E
R
I
C
A
LS
O
L
U
T
I
O
N
T
H
ES
I
M
P
L
E
XM
E
T
H
O
D 1
2
1

l
1
l

;
x=10.4762xi=6.4285.X
;=0.4524I=-21437.14

-17/1980
-51198

A(:
6
).
/
A
(
:
l
)
ans
85/4
25/3
70/3
-16/3

format short
H1U4y

4ur

'To find LBV divide last column by coeff. in EBV'i


'column';
'Take the minimum of the positive values';

0
0

A=[O.4 0.6 1 0 0 8.5;3 -1 0 1 0 25;3 6 0 0 1 70;-990


-900 0 o0 5280]
A
2/5 3/5
1 o 0 17/2
3
1
0
25
1
0
3
6
0 o 1
70
9
90 -900
0 o 0 5280
'This is Table 1 - note the canonical form';
'EBV is x1 - first c
olumn';

0
0

o
0
0
-7/99
0 0 0 16/3
'something not right - The above division should have';
'been an elemen byelement one'i

3
.
3
.
3 SolutlonUslngMATLAB

Ex
a
m
p
l
e3_l
.
t
x
t
format compact
format rational
'Set up the coefficient matrix and b as matrix A';

A(:
6)/A(:
1
)
ans

N
o
t
e

t
h
ev
a
l
u
e
o
f
j
"+21437.14=Ofromw
h
i
c
ht
h
s
o
l
u
t
i
o
nf
o
r
t
h
eo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
n
(
f
.
)c
a
nb
eo
b
t
a
i
n
e
d
.

Thet
a
b
l
e
si
nt
h
ep
r
e
v
i
o
u
sc
o
m
p
u
t
a
t
i
o
nc
a
nb
eg
e
n
e
r
a
t
e
dq
u
i
t
ee
f
f
o
r
t
l
e
s
s
l
yu
s
i
n
ga
s
p
r
e
s
h
e
e
t.l
i
s would b
ea
c
c
o
m
p
l
i
s
h
e
d by s
e
t
t
i
n
g up .
em
a
t
h
e
m
a
t
i
c
a
l
m
a
n
i
p
u
l
a
t
i
o
ni
no
n
ec
e
l
la
n
dc
o
p
y
i
n
ga
n
dp
gt
oeo
t
h
e
rc
e
l
l
s
.S
i
n
c
et
h
i
sbook
i
s
t
si
nMATLABp
r
o
g
r
n
ming

t
h
ei
l
l
u
s
a
t
i
o
nh
e
r
ew
i
l
la
l
l
o
w o
p
p
o
r
t
u
n
i
t
yt
o
a
l
s
o s
p
r
a
c
t
i
c
et
h
eu
s
eo
fMATLABf
o
rm
a
t
r
i
xm
a
n
i
p
u
l
a
t
i
o
n
s
.Thec
o
d
ep
r
e
s
e
n
t
e
dh
e
r
ei
s
p
t
u
r
e
du
s
i
n
gM
A
T
L
A
B
'
Sd
i
a
r
ycommand.Thecommandr
e
q
u
i
r
e
st
h
enameo
fa
f
i
l
ew
h
i
c
hr
e
c
o
r
d
sa
l
lc
o
m
m
u
n
i
c
a
t
i
o
nw
i
l
hMATLABa
sw
e
l
litsr
e
s
p
o
n
s
ed
u
r
i
n
g
i
n
t
e
r
a
c
t
i
v
es
e
s
s
i
o
n(
s
e
eh
e
l
pd
i
a
r
y
)
.Th
et
e
x
tf
i
l
ei
scommentedl
i
b
e
r
a
l
l
y(
f
o
ry
o
u
r
u
n
d
e
r
s
t
a
n
d
i
n
g
)a
n
du
s
i
n
gt
h
es
e
m
i
c
o
l
o
na
tee
n
do
ft
h
es
t
r
.
g
sp
r
e
v
e
n
t
se
i
r
e
c
h
o
.
hi
sf
i
l
et
h
ecommentsa
r
eq
u
o
d
s
n
g
.Thec
o
d
er
e
p
r
e
s
e
n
t
s d
e
t
a
i
l
so
fa
na
c
t
u
a
l
s
e
s
s
i
o
n
.Th
e
s
es
t
a
t
e
m
e
n
t
sa
r
et
y
p
e
da
e
reMATLABp
r
o
m
p
t t
h
eCommand
w
i
n
d
o
w
.Th
i
si
st
h
e
r
e
f
o
r
ea
ni
n
t
e
r
a
c
t
i
v
es
e
s
s
i
o
n
.V
a
l
u
e
st
oer
i
g
h
to
fa
n
s=i
st
h
ee
c
h
o
o
fr
e
s
u
l
to
fee
x
u
t
i
o
no
feMATLABcommand.N
o
t
e
:commentsn
e
e
dn
o
tbe
t
y
p
e
d
.

'The row identifies the pivo row to create he next


table';

A(:
6) ./A(:
l)

ans =
21.2500
8.3333
23.3333
-5.3333
'Second row is the pivot row and x4 is LBV';
'Constructing Table 2';
'note the scaling facor for the matrix'i
A
A

1.0e+003 *
0.0004 0.0006 0.0010
0
0.0030 -0.0010
o
0.0010
0.0030 0.0060
o
0
-0.9900 -0.9000
o
0
'The element at A(2
1
) must be 1';
A(2:) = A(2:)/A(21
)

o
o

0.0085
0.0250
0.0010 0.0700
o 5.2800

0.0085

0.00 3

1.0e+003 *
0.0004 0.0006 0.0010
o
0.0010 -0.0003
0 0.0003
0.0030 0.0060
o 0.0010
0
-0.9900 -0.9000
0
0
o

0.0700
5.2800

3
.
3N
U
M
E
R
I
C
A
LS
O
L
U
T
lO
NT
H
ES
I
M
P
L
E
XMETHOD 1
2
3

122 L
I
N
E
A
RPROGRAMMING

2
) must have a value of 0';
'A(2
A(2:) =A(2:) - A(22)*A(3:);
'A(4
2
) must have a va1ue of 0';
A(4:)
A(4:) - A(42)*A(3:);

l
) must be a 0 '
;
'The element at A(l
A(l:) =A(l:) 0.4*A(2:)
A

1.0e+003

----

nunUFO
nu nU2d
nunununHE
nununU

nununu 4

nununup
nununu

4
U
U
U

nunununu

nunununu

6.4286
-11.0000
'Pivot row is third row and LBV is x5'
ans
Pivot row is third row and LBV is x5
'Oops forgot the semicolon';
'Calcu1ation of LBV';
'Construction of Table 3.3 -no echo of calculations';
2
) must have value 1';
'A(3
A(3:) = A(3:)/A(32);
'A(1
2
) must have a value of 0';
A(l:) - A(12)*A(3:);
A(l:)

7.0455
2
5.0000

-11/105 19/42
1/21 220/21
1/7
45/7
1230/7 150060/7
'No further iterations necessary';
diary off
--

ans

A(:
6
)./A(:
2
)

Rdrf

7
o
'
'
'

'
'
a
u
-anu

'
4

4nununu

1.0e+004 *
0.0001 0.0001 0.0000
o 0.0005
0.0001 -0.0000
0.0000
O 0.0008
o 0.0007
o -0.0001 0.0001 0.0045
0
.1230
0.0330
o 1.3530
'Table 2 complete - and canonical form is present';
2)';
'Solution is not converged because of A(4

EBV is x2';
'Calculation of LBV';

A
nunU 4
nu

0.0083
0.0450
5.2800

----

0.0052

format rational
nu
4nunu

nununU
nununuEJ
nununU
nunununu

0.0007 0.0010 0.0001


0.0010 -0.0003
0.0003
o
o 0.0070
-0.0010 0.0010
-0.9900 -0.9000
o
o
o
'Element at A(4
1
) must be 0';
A(4:)
A(4:) - A(41)*A(2:)

ununu

1.Oe+003 *

0.0001
0
o 0.0001

nunununu

A'
"
1.0e+004 *

nu
nu
nu

nu

00..00005823
50..20870000

0.0007 0.0010 -0.0001


o
0.0010 0.0003
o 0.0003
0.0030 0.0060
o 0.0010
0.9900 -0.9000
o
o
1
) must be a zero';
'Element at A(3
A(3:) =A(3:) - A(31)*A(2:)

Th
el
a
s
tfewcommandss
u
g
g
s
lI
h
cp
o
s
s
i
b
i
l
i
t
yo
fd
e
v
e
l
o
p
i
n
gaf
o
rl
o
o
pf
o
rt
h
erow
saproblema
tt
h
ee
n
do
fec
h
a
p
t
e
r
.
m
a
n
i
p
u
l
a
t
i
o
n
s
.I
ti
ss
u
g
g
e
s
t
e
da
onToolbox
3
.
3
.
4 SolutlonUsingM
A
T
L
A
8
'
SOptimiza

T
h
i
ss
e
c
t
i
o
ni
su
s
e
f
u
lf
o
rt
h
o
s
ct
h
a
th
a
v
ea
c
c
e
s
s1
0t
h
eO
p
t
i
m
i
z
a
t
i
o
nT
o
o
l
b
o
.
xfrom

MA'
1.AB(
V
e
r5
.
2
)
.
S
t
a
r
tMA
T
I
.
.
A
B
.I
nt
h
eCommandwindowt
y
p
eh
e
l
pl
p
.T
h
i
sw
i
l
lp
r
o
v
i
d
ei
n
f
o
r
m
a
t
i
o
n
o
nt
h
eu
s
eo
ft
h
el
i
n
e
a
rp
r
o
g
r
a
m
m
i
n
gr
o
u
t
i
n
e1
0s
o
l
v
et
h
ep
r
o
b
l
e
m
.H
e
r
e
.t
h
es
t
a
n
d
a
r
d
u
s
ei
se
m
p
l
o
y
e
d
.I
ti
n
v
o
l
v
e
ss
p
e
c
i
f
y
i
n
gv
a
1u
e
sf
o
rt
h
ec
o
s
lc
o
e
f
f
i
c
i
e
n
t
s(

v
e
c
t
o
r
)
.t
h
e
x
)
.a
n
dt
h
er
i
g
h
t
h
a
n
ds
i
d
ev
e
c
t
o
r(
b
v
e
c
t
o
r
)
.
c
o
e
f
f
i
c
i
e
n
t
l
c
o
n
s
t
r
a
i
n
lm
a
t
r
i
x(A-ma
bef
o
l
l
o
w
i
n
gs
e
q
u
e
n
c
eo
fs
t
e
p
si
l
l
u
s
a
t
e
seu
s
eo
ft
h
ep
r
o
g
r
a
m
.

U
s
i
n
gt
h
eL
in
e
a
rProgrammingR
o
u
t
i
n
e
9
90;900];
>> f
>>A
[0.4 0.6 3 -1; 3 6
)
>>b
[8.5 25 70)'
>> x '
" 1p(f
A
b)

Thes
o
l
u
t
i
o
na
sp
o
s
l
e
di
nt
h
eCommandw
i
n
d
o
w
:

x
10.4762
6.4286
>> sol "
' f'*x

124

L
lN
E
A
RPROGRAMMING

3
.
4

801

A
O
D
I
T
I
O
N
A
LE
X
A
M
P
L
E
S 125
(
3
.
2
2
)

x+y~ 64O

1
.6157e+004

d
d
e
d c
o
n
s
t
a
n
t-5280whichwaso
m
i
t
t
e
di
np
r
o
b
l
e
m
Tot
h
i
ss
o
l
u
t
i
o
nmustbea
d
e
f
i
n
i
t
i
o
nf
o
rMATLAB.T
h
e
r
ea
r
m
anyd
i
f
f
e
r
e
n
twayst
ou
s
e l
i
n
e
a
rprogramming
f
u
n
c
t
i
o
ni
nMATLAB.

Thes
i
d
ec
o
n
s
t
r
a
i
n
t
sonxa
n
dya
r
et
h
r
e
s
p
e
c
t
i
v
ew
a
r
e
h
o
u
s
ei
I
m
i
t
s
.Thes
h
i
p
p
i
n
gc
o
s
l
I
S
.
5*
y+6
.
0*
(
2
0
0
0-x-y)+
Cost=4.2 x+4

4
.
7 (
8oo-x)+4.(10-y
)+5.
1(
x+y-6
4
0
)

3
.
4 ADDITIONALEXAMPLES

.
4
'
"x-y+1
9120
Cost=1

h i
ss
e
c
t
i
o
na
d
d
i
t
i
o
n
a
le
x
a
m
p
l
e
s ep
s
e
n
t
e
d
.Th
e
s
ee
x
a
m
p
l
e
si
l
l
u
s
t
r
a
t
ee
re
q
u
a
l
e
x
t
e
n
s
i
o
n
/
m
o
d
i
f
i
c
a
t
i
o
no
ft
h
eS
i
m
p
l
e
x method t
oh
a
n
d
l
eg
r
e
a
t
e
r o
c
o
n
s
t
r
a
i
n
t
s
n
e
g
a
t
i
v
ev
a
l
u
e
sf
o
rt
h
ed
e
s
i
g
nv
a
r
i
a
b
l
e
s
e
q
u
a
l
i
l
yc
o
n
s
t
r
a
i
n
t
s

a
n
ds
oo
n
.
I
na
l
Iec
a
s
e
st
h
ep
r
o
b
l
e
mi
st
r
a
n
s
f
o
r
m
e
da
p
p
r
o
p
r
i
a
t
e
l
ya
n
desameS
i
m
p
l
e
x
melhodi
st
h
e
na
p
p
l
i
e
d
.

Assemb
1
ingt
h
ep
r
o
b
l
e
m(
E
x
a
m
p
l
e3
.
2
)

Mbz
e f
(
x
y
)
:1
.4x-y+1
9
1
2
0

(
3
.
2
4
)

Subject81(X

y
)
:x+y:
52000

(
3
.
2
5
)

3
.
4
.
1 Example3.2TransportationProblem

8
2
(
X

y
)
:x+y6
40

I
kc
perativesuppi
1
e
smi
I
ki
ng
a
l
l
o
nj
u
g
sfromi
t
stwow
a
r
e
h
o
u
s
e
s
TheF
r
e
s
hMi
ewY
o
r
k
)a
n
dWi
l
Ii
a
m
s
p
o
r
t(
P
e
n
n
s
y
l
v
a
n
i
a
)
.1
th ac
a
p
a
c
i
t
yo
f
l
o
c
a
t
e
di
nB
u
f
f
a
l
o(N
2000g
a
l
l
o
n
sp
e
rd
a
ya
tB
u
f
f
a
l
oand1600ga
Il
o
n
sp
e
rdaya
tWi
1
1i
a
m
s
p
o
r
t
.I
td
e
l
i
v
e
r
s

u
i
r
e
s1
4
4
0
800gallons/dayRochesr(NewY
o
r
k
)
.S
y
r
a
c
u
s
e(NewY
o
r
k
)r
g
a
l
l
o
n
s
l
d
a
y
a
n
der
e
m
a
i
n
d
e
r(
1
3
6
0ga
1l
o
n
s
)a
r
et
r
u
c
k
e
dt
oNewYorkC
i
t
y
.l
ec
o
s
t
4
.
t
os
h
i
pl
h
em
i
l
kt
oe
a
c
ho
fl
h
ed
e
s
t
i
n
a
t
i
o
n
si
sd
i
f
f
e
r
e
n
ta
n
di
sg
i
v
e
ni
nT
a
b
l
e3.
Est
a
b
i
1s
hes
h
i
p
p
i
n
gs
t
e
g
yf
o
rminimums
t
.

(
3
.
2
3
)

(
3
.
2
6
)

0$x:
58 0:
5y:
51
4
4
0

(
3
.
2
7
)

F
i
g
u
r
e3
.
7i
I
Iu
s
t
r
a
t
e
st
h
eg
r
a
p
h
i
c
a
ls
o
l
u
t
i
o
nt
oExample3
.
2
.Thel
i
n
er
e
p
r
e
s
e
n
t
i
n
gt
h
e
19
1
2
0
)
.
c
o
n
s
t
a
n
to
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
nv
a
l
u
e
sl
e
r
md
o
e
sn
o
ti
n
c
l
u
d
et
h
ec
o
n
s
t
a
n
tt
e
r
m(
Thes
o
l
u
t
i
o
ni
sa
tt
h
ei
n
t
e
r
s
e
c
l
i
o
no
fc
o
n
s
t
r
i
n
tg
la
n
dt
h
eu
p
p
e
ri
1m
i
tonev
a
l
u
eo
f
xaseo
b
j
e
c
t
i
v
ei
st
od
e
c
r
e
a
s
e
f
a
smucha
sp
o
s
s
i
b
l
e
.Fromt
h
ef
i
g
u
r
et
h
es
o
l
u
t
i
o
ni
s

2
0
0

f=168ooor$168.oo
x=800;y=1

(
3
.
2
8
)

ProblemF
o
r
m
u
l
a
t
/
o
n
:Le
txbeenumbero
fga
I
l
o
n
ss
h
i
p
p
e
dfromB
u
f
f
a
l
ot
o
t
ybeenumbero
fg
a
l
l
o
n
ss
h
i
p
p
e
dmB
u
f
f
a
l
ot
oS
y
r
a
c
u
s
e
.
R
o
c
h
e
s
t
e
r
.Le

ogallons/day.Therefore

Thew
a
r
e
h
o
u
s
i
n
gc
o
n
s
i
n
ta
tB
u
f
f
a
I
oi
s2

x+y$2000

(
3
.
2
1
)

Amounts
h
i
p
p
e
domW
i
l
l
i
a
m
s
p
o
r
tt
oR

hes =8oo-x
Am
o
u
n
ts
h
i
p
p
e
domW
i
l
l
i
a
m
s
p
o
r
tt
oS
y
r
a
c
u
s
e=1
4
4
0-y
Amounts
h
i
p
p
e
df
r
o
mW
i
l
l
i
a
m
s
p
o
r
tt
oNewYorkC
i
t
yi
s
1
6
0
0-(
8
0
0-x)(
1 O
y
)o

Minimize

a
1
0
B
u
f
f

f
t
.
x
y
)
:1
.4x-y+19120

S
u
b
j
e
c
tt
o
: g
l
(
X

y
)
:x+y+S
I=20

T
a
b
l
e3
.
4S
h
i
p
p
i
n
gCost(Cen p
e
r
G
a
l
l
o
n
)

Wi
I
l
i
n
s
p
o
r
t

Two-PhaseS/mplexMethod:Themaind
i
f
f
e
r
e
n
c
eb
e
l
w
e
e
nExamples3
.
1a
n
d
o
n
s
a
i
n
ti
nt
h
eI
a
t
t
e
r
.S
i
n
c
et
h
es
t
a
n
d
a
r
dLPp
r
o
b
l
e
mo
n
l
yr
e
q
u
i
r
e
st
h
e
3
.
2i
se c
)er
i
g
h
t
h
a
n
dc
o
n
s
t
r
a
i
n
t
sonxa
n
dyh
a
v
et
o
d
e
s
i
g
nv
a
r
i
a
b
l
e
sbes
e
m
i
p
o
s
i
t
i
v
e 0
b
eaccommodatedt
h
r
o
u
g
ha
d
d
i
t
i
o
na
Ii
n
e
q
ua
Ii
t
yc
o
n
s
t
r
a
i
n
t
s
.A
p
p
l
y
i
n
gt
h
er
e
g
u
l
a
r
t
eap
r
o
b
l
e
mi
nr
e
c
o
g
n
i
z
i
n
g
S
i
m
p
l
e
xmethodw
i
t
h
o
u
ta
d
d
i
t
i
o
na
Ip
r
o
c
e
s
s
i
n
gwouldc
r
ec
o
n
i
1f
o
r
m
. Example 3
.
2e
x
p
r
e
s
s
e
di
n es
t
a
n
d
a
r
df
o
r
m
a
to
fl
i
n
e
a
r
programmmgI
S

R
o
c
h
e
s
t
e
r

S
y
r
a
c
u
s
e

NewY
o
r
kC
i
t
y

4
.
2
4
.
7

4
.
5
4
.
6

6
.
0
5
.
1

(
3
.
2
4
)
(
3
.
2
9
)

g
2
(
X

y
)
:x+y-S2=6
40

(
3
.
3
0
)

g
3
(
X

y
)
:x+S)=8
00

(
3
.
3
1
)

8
4
(
x

y
)
:Y+S4=1
4
4
0

(
3.
32
)

126

L
lN
EARPROGRAMMING

3
.
4A
D
D
I
T
I
O
N
A
LE
X
A
M
P
L
E
S 127

Jv
a
r
i
a
b
l
e
s
. di
fi
t
sva
Ju
ei
sz
e
r
o
i
si
m
p
l
i
e
s
o
b
j
e
c
t
i
v
ed
e
p
e
n
d
so
n
l
yont
h
ea
r
t
i
f
i
c
ia
r
t
i
f
i
c
ia
Jv
a
r
i
a
b
l
e
sa
r
el
Io
n
b
a
s
i
cv
a
r
i
a
b
l
e
s
.Thi
sa
l
s
os
u
g
g
e
s
t
st
h
a
tt
h
e
s
e
t
h
a
t a
v
a
r
i
a
b
l
e
swercb
a
s
i
cv
a
r
i
a
b
l
e
sa
tt
h
es
t
a
r
to
ft
h
ep
r
o
c
e
d
u
r
c
.WhenP
h
a
s
e1i
sc
o
m
p
l
e
t
e
d
e
nb
o
t
ht
h
ea
r
t
i
f
i
c
ia
Jo
b
jc
t
i
v
ef
u
n
c
t
i
o
nandt
h
ea
r
t
i
f
l
c
i
a
lv
a
r
i
a
b
l
e
sa
r
ed
i
s
c
a
r
d
e
d
m

t
a
b
l
eandP
h
a
s
e1/b
e
g
i
n
s
.
P
h
a
s
eI
Ii
st
h
es
t
a
n
d
a
r
dS
i
m
p
l
e
xt
e
c
h
n
i
q
u
ea
p
p
l
i
e
dt
ot
h
et
a
b
l
efromt
h
eendo
f
r
t
i
f
i
c
i
a
le
l
e
m
e
n
t
s
.Th
et
a
b
l
es
h
o
u
l
db
ei
nc
a
n
o
n
i
c
a
lf
o
r
m
.
P
h
a
s
e1n
e
g
l
e
c
t
i
n
ga
l
l a
A
d
d
i
t
i
o
n
a
lt
a
b
l
e
sa
r
eo
b
t
a
i
n
e
da
sn
e
c
e
s
s
a
r
yu
n
t
i
lt
h
es
o
l
u
t
i
o
ni
sr
e
a
c
h
e
d
.
I
nExample3
.
2s
e
tupe
a
r
l
i
e
r

t
h
e
r
ew
i
l
lb
eonea
r
t
i
f
i
c
ia
Jv
a
r
i
a
b
l
ea1
anda
na
r
t
i
f
i
c
ia
J
c
o
s
tf
u
n
c
t
i
o
nA"I
nt
h
eexampleo
n
l
yE
q
u
a
t
i
o
n(
3.
3
0
)w
i
l
lbea
f
f
e
c
t
e
da
s

25

2
2

~
E

5
0
0

g
2
(
X

y
)
:x+y-s2+al=640
i
4
0

1
0
0
0
1
5
0
0

1
0
0

2
0
0

3
0
0
4
0
0
5
0
0
1
0
1
1
0
m
B
f
o
l
o
R
h 1
0
1
x
:
g

6
0
0

7
0
0

(
3
.
3
4
)

Thea
r
t
i
f
i
c
ia
Jc
o
s
tf
u
n
c
t
i
o
nw
i
l
lb
e
(
3.
35)

Af=al

Hhh

--

g
o
=

8
0
0

T
a
b
l
e3
.
5r
e
p
s
e
n
t
s f
1
f
S
tt
a
b
l
ei
nP
h
a
s
eI
.Thef
1
r
s
tf
o
u
rrowsr
e
p
s
e
n
tE
c
o
n
s
t
r
a
i
n
t
s
.Thef
1f
t
hrow p s
e
n
t
seo
r
i
g
i
n
a
lo
b
j
ef
u
n
c
t
i
o
n
.Th
el
a
s
trowi
se
a
r
t
i
f
i
c
ia
Jo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
n
.Thi
si
s rowu
S
e
dt
o i
v
i
t
e
r
a
t
i
o
n
si
nP
h
a
s
e1
.
T
a
b
l
e3
.
5i
sn
o
ti
nc
a
n
o
n
i
c
a
Jf
o
r
m
.b
u
tr
e
p
l
a
c
i
n
gt
h
es
i
x
t
hrowbyt
h
er
e
s
u
l
to
f
s
u
b
t
r
a
c
t
i
n
ges
e
c
o
n
drowfromt
h
es
i
x roww
i
l
lp
r
o
v
i
d
eac
a
n
o
n
i
c
a
lf
o
r
m
.T
a
b
l
e
3
.
6i
l
l
u
s
t
r
a
t
e
st
h
er
e
s
u
l
to
fs
u
c
harowm
a
n
i
p
u
l
a
t
i
o
n
.

(
3
.
3
3
)

eEBV(
e
n
t
e
I
gb
a
s
i
cv
a
r
i
a
b
l
e
)a
sb
o
t
ho
fthemhaveac
o
e
f
f
i
c
i
e
n
to
f-1i
nel
a
s
t

F
l
g
u
r
e3
.
7G
r
a
p
h
l
c
a
l8
0
l
u
l
l
o
n
E
x
a
m
p
l
a3
.
2
.

SlmplexMe
od

PhBse1
:I
nT
a
b
l
e3
.
6t
h
e
r
ea
r
etwoc
h
o
i
c
e
s(
x

y
)a
v
a
i
l
a
b
l
ef
o
r
;y~ 0
;s
l
's
2
's
3
's
4~ 0
x~ 0

H
e
r
e
s
l
S
3
's
4a
r
et
h
es
l
a
c
kv
a
r
i
a
b
l
e
sandS
2i
s s
u
r
p
.
l
(
s
i
m
i
l
a
rt
oaslack-usedf
o
r
~ c
o
n
s
t
r
a
i
n
t
s
)v
a
r
i
a
b
l
e
.I
nExample3
1
. whent
h
ef
1
r
s
tS
i
m
p
l
e
xt
a
b
l
ewass
e
tup
(
S
i
m
p
l
e
xT
a
b
l
e3.
1
)
t
h
es
l
a
c
kv
a
r
i
a
b
l
e
swereh
e
l
p
f
u
li
ni
d
e
n
t
i
f
y
i
n
gt
h
ec
a
n
o
n
i
c
a
l
f
o
r
m
.Herei
nExample3
.
2
i
tdo
n
o
tworkout twayb
e
c
a
u
s
et
h
ec
o
e
f
f
l
c
i
e
n
to
fS
2
i
s1
.M
u
l
t
i
p
l
y
i
n
gEq
u
a
t
i
o
n(
3.
30
)by-1m
i
g
r
a
t
e
st
h
en
e
g
a
t
i
v
es
i
g
nt
ot
h
er
i
g
h
t
h
a
n
d
s
i
d
e
.whichi
sa
l
s
od
i
sa
Jl
oweda
s:
f
a
ser
e
c
o
g
n
i
t
i
o
no
fec
a
n
o
n
i
c
a
lformi
s
c
o
n
c
e
r
n
e
d
.T
h
i
smeansa
d
d
i
t
i
o
n
a
lp
r
o
b
l
e
m
s
p
e
c
i
f
i
cp
r
e
p
r
o
c
e
s
s
i
n
gmustt
a
k
ep
l
a
c
et
o
o
di
su
s
e
dt
oh
a
n
d
l
el
a
r
g
e
i
d
e
n
t
i
f
yt
h
ei
n
i
t
i
a
lc
a
n
o
n
i
c
a
lf
o
r
m
.S
i
n
c
et
h
eS
i
m
p
l
e
xme
p
r
o
b
l
e
m
sw
i
t
hs
e
v
e
r
a
lv
a
r
i
a
b
l
e
s
i
ti
smorec
o
n
v
e
n
i
e
n
tt
oa
p
p
l
yt
h
eS
i
m
p
l
e
xp
r
o
c
e
d
u
i
nac
o
n
s
i
s
t
e
n
tw
a
y
.T
h
i
si
sa
c
c
o
m
p
l
i
s
h
e
dbya
p
p
l
y
i
n
gt
h
esameS
i
m
p
l
e
xt
e
c
h
n
i
q
u
ei
n
at
w
o
p
a
r
ts
e
q
u
e
n
c
e
.
l
ef
1
r
s
tp i
sr
e
c
o
g
t
z
e
da
sP
h
a
s
e1
.H
ereanewk
i
n
do
fv
a
r
i
a
b
l
e
.ana
r
t
i
f
i
c
i
a
l
v
a
r
i
a
b
l
e
i
sd
e
f
i
n
e
df
o
re
a
c
hs
u
r
p
l
u
sv
a
r
i
a
b
l
ei
nt
h
esamee
q
u
a
t
i
o
n
.Also
anew
o
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
nc
a
l
l
e
da
na
r
t
i
f
i
c
i
a
lc
o
s
tf
u
n
c
t
i
o
l
lo
r a
r
t
i
f
i
c
i
a
lo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
n
i
si
n
t
r
o
d
u
c
e
d
.Thea
r
t
i
f
l
c
i
a
lo
b
j
e
c
t
i
v
ei
sd
e
f
i
n
e
d asumo
fa
Jl
t
h
ea
r
t
i
f
i
c
i
a
lv
a
r
i
a
b
l
e
s
i
nt
h
ep
r
o
b
l
e
m
.I
nP
h
a
s
e1
ea
r
t
i
f
l
c
i
a
lo
b
j
e
c
t
iv
ei
s d
u
c
e
dt
oz
e
r
ou
s
i
n
gt
h
es
t
a
n
d
a
r
d
S
i
m
p
l
c
xp
r
o
c
e
d
u
r
e
.Whent
h
i
si
sa
c
c
o
m
p
l
i
s
h
e
d
P
h
a
s
e1i
sc
o
m
p
l
e
t
e
.Ifea
r
t
i
f
i
c
ia
J

r
o
w
.W
h
i
l
et
h
ec
h
o
i
c
ec
a
nbea
r
b
i
t
r
a
r
yxi
sagoodc
h
o
i
c
eb
e
c
a
u
s
ei
th
a
sal
a
r
g
e
r
Jo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
n
. l
eLBV(
l
e
a
v
i
n
gb
a
s
i
c
n
e
g
a
t
i
v
ec
o
e
f
f
l
c
i
e
n
ti
nt
h
eo
r
i
g
i
na
v
a
r
i
a
b
l
e
)i
si
d
e
n
t
i
f
i
e
dt
h
r
o
u
g
ht
h
eminimump
o
s
i
t
i
v
eva
Ju
eo
ft
h
er
a
t
i
oo
ft
h
ev
a
l
u
e
si
n
t
h
ebcolumnt
o ev
a
l
u
e
su
n
d
e
r xcolumn. l
e
s
cr
a
t
i
o
sa
r
e2
0
0
0
/
16
4
0
1
1and
8
0
0
11
.Thes
e
l
e
c
t
i
o
ni
d
e
n
t
i
f
i
e
ses
e
c
o
n
drowa
s p
i
v
o
trowanda1 eLBV.
i
v
o
trowt
h
eu
n
i
tv
e
c
t
o
r[
01000O
]
Thtobe
U
s
i
n
gt
h
es
e
c
o
n
drow ep
5 E
xample3
.
2
:I
n
i
t
i
a
lT
a
b
l
e

P
h
a
s
e1
T
a
b
l
e3.

S
I

S
2

0
1

S
3

S
4

f4


A
l

1
.

2
0
0
0
640
800
1
4
4
0

1
9
1
2
0
f

L
lN
E
A
RP
R
O
G
R
A
M
M
I
N
G

T
a
b
l
e3
.
6E
x
a
m
p
l
e3
.
2
:S
l
m
p
l
e
xT
a
b
l
e1

P
h
n
s
e(
C
a
n
o
n
l
I
F
o
r
m
)

nuhunu'

o
o

2
0
0
0
640

8
0
0
1
4
4
0

f-1
9
1
2
0

y
0

0
.
4

$2

o
1
.
4

$)

$4

nununu'a

$4

nun'anunu

$)

A
D
D
I
T
I
O
N
A
LE
X
A
M
P
L
E
S 1
2
9

p
l
e
xT
a
b
l
e1

P
h
a
s
es(
C
a
n
o
n
i

IF
o
r
m
)
T
a
b
l
e3
.
8E
x
u
m
p
l
e3
.
2
:m

-EAUAununu

a
l

hunu-anunuhu

$2

AU'anu u
nunu

-shununuhunu

1
.
4

$
1

1lull-

3
.
4

u
'
aU
nunU

b
V L F iji E
-E

1
2
8

1
3
6
0
0

1
6
0
1440

f-18224

Af-640

SimplexMethod

Phase1
1
:Thec
o
l
u
m
nw
i
t
ht
h
eg
r
t
e
s
tn
e
g
a
t
i
v
ec
o
e
f
f
i
c
i
e
n
t
(

1
.4
)i
ss
2
l
i
si
seEBV.TheLBVi
ss
3l
ei
r
drowi
st
h
e
r
e
f
o
r
et
h
ep
i
v
o
tr
o
w
.

l
s
f
e
e
df
r
o
mt
h
ealc
o
l
u
m
nt
ot
h
exc
o
l
u
m
n
.C
o
n
s
t
r
u
c
t
i
o
no
f
T
a
b
l
e3
.
7s
t
t
sw
i
t
h

Th
eu
n
i
tv
e
c
t
o
r[
0010O
]
Tm
u
s
tb
ea
n
s
f
e
r
r
e
df
r
o
mt
h
eS3c
o
l
u
m
nt
o S2c
o
l
u
m
n
t
h
r
o
u
g
hr
o
wm
a
n
i
p
u
l
a
t
i
o
n
su
s
i
n
gt
h
ep
i
v
o
tr
o
w
.se
x
e
r
c
i
s
er
e
s
u
l
t
si
nT
a
b
l
e3
.
9
.
Then
e
g
a
t
i
v
ec
o
e
f
f
i
c
i
e
n
tu
n
d
e
rt
h
eyc
o
l
u
m
ns
u
g
g
e
s
t
swea
r
en
o
tf
m
i
s
h
e
dy
et
.F
o
r
t
h
en
e
x
ti
t
e
r
a
t
i
o
n
EBVi
sy
.LBVi
sSI'Th
ef
i
r
s
tr
o
wi
sep
i
v
o
tr
o
w
.Th
eu
n
i
tv
e
c
t
o
r
[
10000]
Tm
u
s
tb
ec
o
n
s
t
r
u
c
t
e
du
n
d
e
rt
h
eyc
o
l
u
m
nu
s
i
n
gr
o
wm
a
n
i
p
u
l
a
t
i
o
nu
s
i
n
g
t
h
ep
i
v
o
tr
o
w
.Thi
sr
e
s
u
l
t
si
nT
a
b
l
e3.
10
.
Th
e
r
ea
r
en
on
e
g
a
t
i
v
ec
o
e
f
f
i
c
i
e
n
t
si
nt
h
el
a
s
tr
o
w
.Thes
o
l
u
t
i
o
nh
a
sb
e
e
nr
e
l
l
c
h
e
d
.
Fromi
n
s
p
e
c
t
i
o
no
fT
a
b
l
e3
.
1
0
t
h
es
o
l
u
t
i
o
ni
s

t
h
es
e
c
o
n
drowi
nT
a
b
l
e3
.
6
w
h
i
c
hi
sa
l
s
ot
h
es
e
c
o
n
dr
o
wo
f
T
a
b
l
e3
.
7b
e
c
a
u
s
eo
f
t
h
e
c
o
ei
c
i
e
n
to
f1i
nt
h
es
e
c
o
n
drowu
n
d
e
rt
h
exc
o
l
u
m
n(
n
os
c
a
l
i
n
gi
sn
e
c
e
s
s
a
r
y
)
.Th
i
s
i
st
h
ep
i
v
o
tr
o
w
.Too
b
t
a
i
na0i
nt
h
ef
i
r
s
tr
o
wi
nt
h
exc
o
l
u
m
n
t
h
ep
i
v
o
trowi
s
c
t
e
df
r
o
m r
s
trowi
nT
a
b
l
e3
.
6
.Th
enewf
i
r
s
trowi
s
s
u
b
t
r

1
3
6
0

Too
b
t
a
i
na0i
nt
h
exc
o
l
u
m
ni
nt
h
et
h
i
r
dr
o
w
.t
h
ep
i
v
o
tr
o
wi
ss
u
b
t
r
a
c
t
e
df
r
o
mt
h
e
t
h
i
r
drowo
f
T
a
b
l
e3
.
6
.Thef
o
u
r
t
hrowh
a
sa0i
np
l
a
c
ea
n
dh
e
n
c
ci
sc
o
p
i
e
d'OmTable
cf
i
f
t
hrowi
so
b
t
a
i
n
c
db
ya
d
d
i
n
ga1
.4m
u
l
t
i
p
l
eo
ft
h
ep
i
v
o
tr
o
wt
ot
h
ef
it
1
r
o
w
3
.
6
.Th
i
nT
a
b
l
e3
.
6
.Th
el
a
s
trowi
st
h
ea
d
d
i
t
i
o
no
ft
h
ep
i
v
o
tr
o
wt
ot
h
el
a
s
tr
o
wf
r
o
mT
a
b
l
e
3
.
6
.T
a
b
l
e3
.
7i
sc
o
m
p
i
l
e
da
sf
o
l
l
o
w
s
:

8
0
0
Y=
1
2
0
0
.S2=
1
3
6
0
.S4=
240
f=1
6
8
0
0
x=

I
nt
h
i
se
x
n
p
l
e
t
h
es
a
m
eS
i
m
p
l
e
xm
e
t
h
o
dwasr
e
p
e
a
t
e
d
l
ya
p
p
l
i
e
da
l
t
h
o
u
g
ht
h
e
h
a
s
e
s
.Thef
i
r
s
tp
h
a
s
ew
a
sap
r
e
p
r
o
c
e
s
s
i
n
gp
h e

p
r
o
b
l
e
mw
a
ss
o
l
v
e
du
s
i
n
g op
movet
h
es
u
r
p
l
u
sv
a
r
i
a
b
l
e
sawayf
r
o
m i
n
i
t
i
a
ls
e
to
fb
a
s
i
cv
a
r
i
a
b
l
e
s
.l
i
sw
a
s
a
c
h
i
e
v
e
db
yi
n
t
r
o
d
u
c
i
n
ga
d
d
i
t
i
o
n
n
lv
a
r
i
a
b
l
e
sf
o
re
a
c
hs
u
l
u
sv
a
r
i
a
b
l
ea
n
da
n
a
d
d
i
t
i
o
n
a
lc
o
s
tf
u
n
c
t
i
o
nt
h
a
td
r
o
v
eei
t
e
r
a
t
i
o
n
so
ft
h
ef
r
r
t
p
h
a
s
e
.
f
o
r~ a
n
d=c
o
n
s
t
r
a
i
n
t
s
a
r
t
i
f
i
c
i
a
lv
a
r
i
a
b
l
e
sa
n
d a
r
t
i
f
i
c
i
a
lc
o
s
t
Tos
u
m
m
a
r
i
z
e
f
u
n
c
t
i
o
n ei
nd
u
di
n
t
ot
h
ep
r
o
b
l
e
mf
o
rc
o
n
v
e
n
i
e
n
c
e
.l
eS
i
m
p
l
e
xm
e
t
h
o
di
s
a
p
p
l
i
e
di
nt
w
op
h
a
s
e
s
.l
ef
i
r
s
tp
h
a
s
ei
st
e
r
m
i
n
a
t
e
dwhent
h
ea
r
t
i
f
i
c
i
a
lv
a
r
i
a
b
l
e
sc
l
l
n
b
ee
l
i
m
i
n
a
t
e
df
r
o
m p
r
o
b
l
e
m
.

T
a
b
l
e
2

P
h

1(
C
a
n
o
n
l
c
o
lF
o
r
m
)
T
a
b
l
e3
.
7E
x
a
m
p
l
e3
.
2
:S
l
m
p
l
e

$
1

$2

$)

$4

nunuhu'a

1
3
6
0
6
4
0

1
6
0
1440

.
2
:S
l
m
p
l
e
xT
a
b
l
e2

P
h
a
s
n
(Ca
n
o
n
l
c
a
lF
o
r
m
)
T
a
b
l
e3
.
9E
x
a
m
p
l
e3

f-18224

Af

$2
nu'anunu
u

$
'Enunununu

AUAUnu

FromT
a
b
l
e3
.
7
.ev
a
l
u
eo
fAfi
sO
.a
n
dali
sn
o
tab
a
s
i
cv
a
r
i
a
b
l
eT
h
i
si
d
e
n
t
i
f
i
e
s
t
h
ee
n
do
fP
h
a
s
e1
.Th
ea1c
o
l
u
m
na
n
d l
a
s
tr
o
wa
r
ed
i
s
c
a
r
d
e
da
n
dP
h
a
s
e1
1i
ss
t
a
r
t
e
d
w
i
t
hT
a
b
l
e3
.
8
.

--

'anusaE E
a

o
o

u
nu'anununu

1
.4

'a'E'anu-a

'anununununu

a
a

0 1 1 0 0

nu'AUnununu

a
l

(
3
.
3
6
)

$)

$4

o
o
o
0
1
.4

1
2
8

1
6
0
1
4
4
0
f-180

130 UNEARPROGRAMMING

3
.
4A
D
O
I
T
I
O
N
A
lEXAMPLES

T
a
b
l
e3
.
1
0E
xample3
.
2
:S
l
m
p
l
e
xT
a
b
l
e3(
F
i
o
a
l
)

P
h
a
s
cD
x

SI

S2

S)

1
o
0
.
4

Maximize
S4

f-16800

g
.
: xl+ Xz ~5
g
2
:2
x
.+x
4
2=
g
3
:x
1+X2~ 1
x1~ 0

x
su
n
r
e
s
t
r
i
c
t
e
di
ns
i
g
n
2i

.
8
u
s
i
n
gt
h
edrawLine.mm
f
i
l
e

i
sag
r
a
p
h
i
c
a
ld
e
s
c
r
i
p
t
i
o
no
ft
h
ep
r
o
b
l
e
mand
F
i
g
u
r
e3
t
h
es
o
l
u
t
i
o
n
.1
es
o
l
u
t
i
o
ni
sa
t
x
l 3a
n
dX2 -2i
n f
i
g
u
r
e

wherec
o
n
v
e
n
i
e
n
t
l
y e
o
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
ni
sa
l
s
op
a
s
s
i
n
gt
h
r
o
u
g
h
.

StandardFormat:Ast
h
es
t
a
n
d
a
r
df
o
r
m
a
te
x
p
e
o
n
l
yn
o
n
n
e
g
a
t
i
v
ev
a
r
i
a
b
l
e
s
.t
h
e
v
a
r
i
a
b
l
ex2i
sr
e
p
l
a
c
e
dbyap
a
i
ro
fv
a
r
i
a
b
l
e
si
nt
h
ef
o
r
m
u
l
a
t
i
o
n
:
6

0o

10

.
3wi
l
1i
nc
1udeane
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
ta
n
di
J
lu
s
t
r
a
t
e
st
h
et
e
c
h
n
i
q
u
ef
o
r
Example3
h
a
n
d
l
i
n
gv
a
r
i
a
b
l
e
s a
r
eu
n
r
e
s
i
c
t
e
di
ns
i
g
n(
c
a
nh
a
v
en
e
g
a
t
i
v
ev
a
l
u
e
s
)
.V
a
r
i
a
b
l
e
s
l
i
k
ep
r
o
f
i

tt
e
m
p
e
r
a
t
u
r
e
andn
e
tincomec
a
nben
e
g
a
t
i
v
eq
u
i
t
eo
f
l
1
.The u
a
l
i
t
y
i
n
ti
sh
a
n
d
l
e
dbyi
n
d
u
c
i
n
g i
f
i
c
i
a
lv
a
r
i
a
b
l
ea
n
da
p
p
l
y
i
n
getw
o
p
h
a
s
e
c
o
n
s
S
i
m
p
l
e
xl
t
i
q
u i
l
l
u
s
t
e
di
nExample3
.
2
.Wh
ent
h
i
sc
o
r
r
e
s
p
o
n
d
i
c
i
a
l
v
a
r
i
a
b
l
ei
sz
e
r
o
(n
do
fP
h
a
s
e1
)

c
o
n
s
t
r
a
i
n
th
a
sb
e
e
nme
t
.Then
e
g
a
t
i
v
ev
a
l
u
ef
o
r
i
f
f
l
r
e
n
c
eb
e
t
w
e
e
ntwop
o
s
i
t
i
v
ev
a
r
i
a
b
l
e
s
.Once
av
a
r
i
a
b
l
ei
ss
i
m
u
l
a
t
e
dt
h
r
o
u
g
hed
a
g
a
i
n
t
h
e
s
ec
h
a
n
g
s
p
r
e
s
e
r
v
et
h
eS
i
m
p
l
e
xmethodi
ni
t
so
r
i
g
i
n
a
lf
o
r
m
.Th
ef
o
l
l
o
w
i
n
g
i
sa
ni
m
a
g
i
n
a
r
yp
r
o
b
l
e
m
i
l
l
u
s
t
r
a
t
ea
d
d
i
t
i
o
n
a
lf
e
a
t
u
r
e
ss
osomei
L
b
e
r
t
yi
st
a
k
e
ni
nt
h
e
f
o
r
m
u
l
a
t
i
o
n
.

j
{
xl
X2)
:2
x1

Thec
o
n
s
t
r
a
i
n
t
sc
a
nber
e
c
o
g
n
i
z
e
dfromt
h
es
t
a
t
e
m
e
n
to
fep
r
o
b
l
e
md
i
r
e
c
t
l
ya
s

1
2

8
0
0
1
3
6
0
2
4
0

3
.
4
.
2 Example3 -EqualltyConst
i
n
t
sandU
n
r
e
s
t
r
i
c
t
e
dVarlables

1
3
1

TheProblem:Today
af
u
l
l
t
i
r
n
es
t
u
d
e
n
toncampusi
sa
l
w
a
y
sd
r
i
v
e
nt
omaximize
g
r
a
d
e
s
.As
t
r
o
n
gi
n
f
l
u
e
n
c
ef
o
raf
a
v
o
r
a
b
l
eoutcomei
seamounto
fi
n
v
e
s
t
made
i
nh
i
t
t
i
n
g b
o
o
k
sv
e
r
s
u
set
i
m
es
p
e
n
tp
l
a
y
i
n
gp
i
n
b
a
l
l
.I
nat
y
p
i
c
a
lday

a
tl
e
a
s
to
n
e
h
o
u
ri
sd
e
f
i
n
i
t
e
l
ye
x
t
e
n
d
e
dt
ot
h
ep
u
r
s
u
i
to
fl
e
a
r
n
i
n
go
rp
r
e
s
s
i
n
gt
h
es
i
d
e
so
ft
h
e
c
o
i
n
o
p
e
r
a
t
e
dm
a
c
h
i
n
e
.Notmoret
h
a
nf
i
v
eh
o
u
r
si
sa
v
a
i
l
a
b
l
f
o
rs
u
c
hd
i
s
p
e
r
s
i
o
n
.
e
s

af
a
t
i
g
u
et
h
r
e
s
h
o
l
do
ff
o
u
ru
n
i
t
s
.b
a
s
e
donac
o
m
b
i
n
a
t
i
o
no
f two
Overey
a
c
t
i
v
i
t
i
e
sh
a
sb
e
e
ne
s
t
a
b
l
i
s
h
e
da
s a norm f
o
ra
c
c
e
p
t
a
b
l
ep
e
r
f
o
r
m
c
e
.T
h
i
s
c
o
m
b
i
n
a
t
i
o
ni
st
h
esumo
ft
h
eh
o
u
r
ss
p
e
n
ta
tt
h
ep
i
n
b
a
l
lm
a
c
h
i
n
eandt
w
i
c
t
h
e
h
o
u
r
s

a
nacknowledgmentt
h
a
th
i
t
t
i
n
geb
o
o
k
si
ss
t
r
e
s
s
f
u
l
.An
e
g
a
t
i
v
e
s
p
e
n
tonhomework
h
o
u
ro
fp
i
n
b
a
l
lp
l
a
y
i
n
gwi
1
lgot
oi
n
c
r
e
a
s
et
h
et
i
m
es
p
e
n
ts
t
u
d
y
i
n
gwhichw
i
l
l
c
o
n
t
r
i
b
u
t
ep
o
s
i
t
i
v
e
l
yt
ot
h
eg
r
a
d
e
s
.Theo
v
e
r
a
l
lg
r
a
d
ei
sd
e
t
e
r
m
i
n
e
da
sal
i
n
e
a
r
c
o
m
b
i
n
a
t
i
o
no
ft
w
i
c
et
h
et
i
m
es
p
e
n
ta
c
a
d
e
m
i
c
a
l
l
ya
n
ds
u
b
t
r
a
c
t
i
n
get
i
m
es
p
e
n
t
p
l
a
y
i
n
g
.Thep
r
o
b
l
e
me
r
e
f
o
r
ei
st
od
i
s
t
r
i
b
u
t
et
h
et
i
m
et
omaximizet
h
eg
r
a
d
e
s
o
b
t
a
i
n
e
d
.
ProblemF
o
r
m
u
l
a
t
i
o
n
:Therea
r
etwoo
r
i
g
i
n
a
ld
e
s
i
g
nv
a
r
i
a
b
l
e
si
nt
h
ep
r
o
b
l
e
m
.XI
i
st
h
enumbero
fh
o
u
r
ss
p
e
n
ts
t
u
d
y
i
n
ga
n
dX2i
st
h
et
i
m
es
p
e
n
te
n
j
o
y
i
n
gt
h
egameo
f
p
i
n
b
a
l
lwhent
h
s
amet
i
m
ec
a
nbea
d
v
a
n
t
a
g
e
o
u
s
l
ys
p
e
n
tm
a
i
n
t
a
i
n
i
n
gh
i
g
hg
r
a
d
e
s
.
'
he
o
b
j
t
i
v
ef
u
n
c
t
i
o
nc
a
nbee
x
p
r
e
s
s
e
da
s

Z ;

F
l
g
u
r
e3
.
8G
r
a
p
h
l
c
a
ls
o
l
u
t
i
o
nE
x
a
m
p
l
e3
.
3
.

1
3
2

L
lN
E
A
RP
R
O
G
R
A
M
M
I
N
G

3
.
4

Phase"T
a
b
l
e2
:T
a
b
l
e3
.
13i
s1
h
es
e
c
o
n
dt
a
b
l
eu
n
d
e
rP
h
a
s
e.
1Thev
a
Ju
eo
ft
h
e

s
oP
h
a
s
e1i
sn
o
ty
e
to
v
e
r
.TheEBVi
sX22
TheLBVisal
The
a
r
t
i
f
i
c
i
a
lf
u
n
c
t
i
o
ni
s2
s
e
c
o
n
drowi
s1
hep
i
v
o
tr
o
w
.T
h
i
sr
o
wwi
I
Ib
eu
s
e
df
o
rrowm
a
n
i
p
u
l
a
t
i
o
n
s
y
i
e
l
dt
h
e
n
e
x
tt
a
b
l
e
.

(
3
.
3
7
)
x
2
1- X22
2=x
i
n
t
r
o
d
u
c
i
n
gs
l
a
c
k

Thes
t
a
n
d
a
r
df
o
m
l
a
ta
f
t
e
rc
o
n
v
e
r
t
i
n
gt
oam
i
n
i
n
u
z
a
t
i
o
np
r
o
b
l
e
m
s
u
r
p
l
u
sda
r
t
i
f
i
c
i
a
lv
a
r
i
a
b
l
e
sa
n
di
n
c
l
u
d
i
n
gE
q
u
a
t
i
o
n(
3.
37
)r
e
s
u
l
t
si
n
(
3.
38
)

Mi
n
r
i
z
e f-2x1+x21-X22
Su
j
e
c
t
t
o
:

gl:X1+X21-X22+
8
1=5

.
3
:T
a
b
l
c2

P
h
a
s
cI
T
a
b
l
e3
.
1
3E
x
a
m
p
l
e3

X
)

g
l
:2x1+X21-X +at=4
g
3
:X1+X
2
1-X22- 8
2+a2=1
a
Jf
u
n
c
t
i
o
nt
ob
ed
r
i
v
e
nt
oz
e
r
oi
s
Thea
r
t
i
f
i
c
i

(
3
.
3
9
)

Af=al+~

A
D
D
I
T
I
O
N
A
LE
X
A
M
P
L
E
S 1
3
3

X
2
)

X22

$)

U)

8
2

a2


--22
3 -3

4
2
l

2
3

/+2
Ar-2

S
i
m
p
l
e
xT
a
b
l
e0
:T
a
b
l
e3
.
1
1 p1h
es
d
a
r
df
O
m18
t
i
nat
a
b
l
e
.I
ti
sc
a
I
I
e
dT
a
b
l
e

Phase"T
a
b
l
e3
:T
a
b
l
e3
.
1
4i
st
h
enewt
a
b
l
ea
f
t
e
rt
h
erowo
p
e
r
a
t
i
o
n
s
.Thev
a
l
u
eo
f

n
beob
r
v
e
d
.A
d
d
i
n
ge
s
e
c
o
n
d

a
r
t
i
f
i
c
ia
Jf
u
n
c
o
ni
s0a
sb
o
1
hala
n
da2 en
o
n
b
a
s
i
cv
a
r
i
a
b
l
e
sw
i
t
hava
Ju
eo
f0
f
o
ri
si
t
e
r
a
t
i
o
n
.lIss
i
g
f
i
e
see
n
do
fP
h
a
s
e1
.P
h
a
s
enw
i
l
ls
t
a
r
tw
i
t
ht
h
el
a
s
trow
a
1
n
c
t
i
o
n
)a
n
dt
h
et
w
oc
o
l
u
m
n
st
h
a
tr
e
p
r
e
s
e
n
tt
h
ea
r
t
i
c
i
a
lv
a
r
i
a
b
l
e
sr
e
m
o
v
e
d
(
a
r
t
i
f
i
c
i

o
t
oa
I
I
owf
o
rp
r
e
p
r
o
c
e
s
s
i
n
gw
i
1
hr
e
s
p
e
c
t

t
h
ea
r
t
i
f
i
c
i
a
1v
a
r
i
a
b
l
e
ss
ot
h
a
tac
a
n
o
n
i
c
a
1f
OlID

di
r
d
r
o
w d
s
u
b
t
r
a
c
ge
r
e
s
u
l
tomel
a
s
t
rowa
n
dr
e
p
l
a
c
i
n
gt
h
el
a
s
troww
i
1
ht
h
i
sc
o
m
p
u
t
a
t
i
o
nwi
l
1y
i
e
l
dt
h
en
e
x
tt
a
b
l
e
.

f
r
o
mt
h
ec
u
r
r
e
n
tt
a
b
l
e
.

T
a
b
l
c3
.
1
1 Exam
p
l
c3
.
3
:T
a
b
l
e0
P
h
a
s
cI
X
)

X
2
)

X
2
2

1
I
1

-2

8
)

Tb
l
c
3
.
1
4 Ex
m
p
l
c3
.
3
:T
a
b
l
e3

P
h
a
s
eI

1
0
1
1
1
)
a
)

S
2

a2

b
5
4
1

A
r

Phase1
T
a
b
l
e1
:T
a
b
l
e3
.
1
2p
r
o
v
i
d
e
ses
t
a
r
to
feS
i
m
p
l
e
xm
e
t
h
o
d
.I
nP
h
a
s
e1
t
h
m
o
t
i
v
a
t
i
o
ni
st
od
r
i
v
et
h
ea
r
t
i
f
i
c
i
a
lf
u
n
c
t
i
o
nt
o0

w
h
i
c
hh
a
sav
a
l
u
eo
f5
.TheEBV
i
S
X
lTheLBVi
sa2'Th
et
h
i
r
drowi
st
h
ep
i
v
o
tr
o
w
.Rowm
a
n
i
p
u
l
a
t
i
o
n
sw
i
t
ht
h
ep
i
v
o
t
rowl
e
a
dt
oT
a
b
l
e3.
13
.

Ph
s
c
l
T
l
l
b
l
c3
.
1
2E
x
a
m
p
l
c3
.
3
:T
a
b
l
e1

X
2
'

X22

a)

$
2

a2

'Aq

--

12123

5
o

o
f



Ar-5
4

X
2
)

X
2
2

o

o

$)

1
o
3

8
2

o
z

b
4
2
3
/+8
A
r

Phase1
Table1
:T
a
b
l
e3
.
1
5r
e
p
r
e
s
e
n
t
se
t
a
b
l
t
os
t
a
r
tt
h
ei
t
e
m
t
i
o
n
sf
o
rP
h
a
s
e1
1
.
However

t
h
e
r
ea
r
enon
e
g
a
t
i
v
ec
o
e
f
f
i
c
i
e
n
t
si
n l
a
s
trow(
r
o
wc
o
r
r
e
s
p
o
n
d
i
n
gt
ot
h
e
ef
u
h
e
r
1
h
ec
a
n
o
n
i
c
a
lf
o
r
mc
a
nb
eo
b
s
e
r
v
e
d
.
o
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
n
)
.S
c
a
n
n
i
n
gt
h
e l
l
e
r
e
f
o
r
e

t
h
es
o
l
u
t
i
o
nh
a
sb
e
e
no
b
t
a
i
n
e
d
.l
es
o
l
u
t
i
o
ni
s
T
a
b
l
c3
.
1
5E
x
a
m
p
l
c3
.
3
:T
a
b
l
e1

P
h
a
s
c1
1

-1

X
)

8
2

4
2
3

/+8

1
3
4 L
J
N
E
A
RP
R
O
G
R
A
M
M
I
N
G

3
.
4 ADDITIONALE
X
A
M
P
L
E
S

xl
=3
;
x
;s
l=4;s2=0;/=-8
22=2
2
1=0;x

1
3
5

+9x
x
O
x
0
0
g
l
:4
xI
4S6
3+I
2+7

i
1
it
yo
ft
h
el
i
t
h
o
g
r
a
p
h
i
cm
a
c
h
i
n
ec
a
nb
ed
e
v
e
l
o
p
e
da
s
T
h
ea
v
a
i
l
a
b

Thev
a
l
u
eo
fX2i
s
X
0-2
)=-2
2=(X21-x
2
2)=(
w
h
i
c
hw
a
si
d
e
n
t
i
f
i
e
de
a
r
l
i
e
rg
r
a
p
h
i
c
a
l
l
y
.
l
ea
b
o
v
ee
x
a
m
p
l
ei
l
l
u
s
t
r
a
t
e
dt
h
eS
i
m
p
l
e
xm
e
t
h
o
df
o
rh
a
n
d
l
i
n
gn
e
g
a
t
i
v
ev
a
r
i
a
b
l
e
s
a
n
de
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
t
s
.Not
a
p
p
l
i
c
a
t
i
o
no
ft
h
eS
i
m
p
l
e
xt
e
c
h
n
i
q
u
ei
t
s
e
l
fd
i
dn
o
t
c
h
a
n
g
e
.Ani
n
t
e
r
e
s
t
i
n
gf
e
a
t
u
r
eo
fep
r
o
b
J
e
mwast
h
a
tP
h
a
s
eI
Iw
a
si
m
m
e
d
i
a
t
l
y
'P
h
a
s
e1
w
i
t
h
o
u
ta
n
yf
u
r
t
h
e
ri
t
e
r
a
t
i
o
n
w
h
i
c
hs
u
g
g
e
s
t
s t
h
es
e
a
r
c
hf
o
r
o
b
t
a
i
n
e
da
f
t
el
t
h
eo
p
t
i
m
u
mi
st
a
k
i
n
gp
l
a
c
ei
nP
h
a
s
1e
v
e
nt
h
o
u
g
hef
o
c
u
si
so
nt
h
ea
r
t
i
f
i
c
i
a
l
v
a
r
i
a
b
l
e
s
.
ample3
.4AF
o
u
r
V
a
r
l
a
b
l
eProblem
3
.
4
.
3 Ex
r
o
c
e
d
u
r
ei
si
d
e
n
t
i
c
a
lt
ot
h
e
E
x
a
m
p
l
e3
.
4p
r
e
s
e
n
t
sap
r
o
b
l
e
mw
i
t
hf
o
u
rv
a
r
i
a
b
l
e
s
.Thep
e
x
c
e
p
ta
te
r
ei
sn
og
r
a
p
h
i
c
a
ls
o
l
u
t
i
o
nt
ot
h
ep
r
o
b
l
e
m
.
o
n
ei
nE
x
a
m
p
l
e3
.
1

g
2
:xl+x2+3x
x
;420
4s
3+8
J
ep
a
c
k
a
g
i
n
gc
o
n
s
t
r
a
i
n
ti
s

g
3
:3
0
xI
+40
.
+20x3+I
O
x
0
0
4S8
A
l
ld
e
s
i
g
nv
a
r
i
a
b
l
e
sa
r
p
t
e
dt
ob
eg
a
t
e
r z
e
r
o
.Asf
o
r
m
u
l
a
t
e
da
b
o
v
ee
r
e
f
o
u
rd
e
s
i
g
nv
a
r
i
a
b
l
e
sa
n
do
n
l
y
p
r
o
b
l
e
ms
u
g
g
e
s
t
sad
e
g
r
e
eo
fi
n
c
o
m
p
l
e
t
e
n
e
s
s
.

n
e
r
ea
t
h
r
e
ec
o
n
s
t
r
a
i
n
t
s
.l
en
u
m
b
e
ro
fv
a
r
i
a
b
l
e
si
nt
h
eb
a
s
i
sc
a
no
n
l
yb
ee
.H
e
n
c
ea
t
a
r
i
a
b
l
e
sm
u
s
th
a
v
eav
a
l
u
eo
fz
e
r
o
.S
e
v
r
a
lu
s
e
f
u
la
d
d
i
t
i
o
n
a
l
l
e
a
s
to
n
eo
fev
c
o
n
s
t
r
a
i
n
t
sc
a
ns
t
i
Ub
ei
n
c
l
u
d
e
dt
od
e
f
i
n
eav
a
l
i
do
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
e
mw
i
t
han
o
n
z
e
r
o
s
o
l
u
t
i
o
n
.l
i
si
snowam
o
d
e
l
i
n
gi
s
s
u
e
.F
o
rp
r
o
b
l
e
m
sw
i
t
hal
i
m
i
t
e
dn
u
m
b
e
ro
fd
e
s
i
g
n
a
y
i
n
ga
t
t
e
n
t
i
o
n1
0t
h
ep
r
o
b
l
e
md
e
v
e
l
o
p
m
e
n
ta
l
l
o
w
sa
n
t
i
c
i
p
a
t
i
o
no
ft
h
e
v
a
r
i
a
b
l
e
sp
s
o
l
u
t
i
o
na
sw
e
l
l t
h
eo
p
p
o
r
t
u
n
i
t
yt
o o
u
b
l
e
s
h
o
o
td
e
c
i
s
i
o
n
sf
r
o
mp
r
a
c
t
i
c
a
l
t
u
d
e
n
tt
oe
x
p
l
o
t
h
i
sp
r
o
b
l
e
mf
u
r
t
h
e
r
.
c
o
n
s
i
d
e
r
a
t
i
o
n
s
.I
ti
sl
e
f
tt
oes
StandardFormat: T
he o
b
j
e
c
t
i
v
f
u
n
c
t
i
o
nr
e
q
u
i
r
e
s a minimum f
o
r
m
u
l
a
t
i
o
n
.

TheProblem: T
heRITs
t
u
d
e
n
t
r
u
nm
i
c
r
o
e
l
e
c
t
r
o
n
i
cf
a
b
r
i
c
a
t
i
o
nf
a
c
i
l
i
t
yi
st
a
k
i
n
g

tc
a
nb
eu
s
i
n(
1
)t
o
u
c
h
o
r
d
e
r
sf
o
rf
o
u
ri
n
d
i
g
e
n
o
u
s
l
yd
e
v
e
l
o
p
e
dASICc
h
i
p
sa
s
e
n
s
o
r
s(
2
) LCD(
3
)p
r
e
s
s
u
r
es
e
n
s
o
r
sa
n
d(
4
)c
o
n
t
r
o
l
l
e
r
s
.The
r
ea
r
es
e
v
e
r
a
l
q
u
i
p
m
e
n
ta
v
a
i
l
a
b
i
l
i
t
ys
t
u
d
e
n
th
o
u
r
s
c
o
n
s
t
r
a
i
n
t
so
nt
h
ep
r
o
d
u
c
t
i
o
nb
a
s
e
do
ns
p
a
c
ee
a
n
dt
h
ef
a
c
ta
tep
r
i
m
a
r
ym
i
s
s
i
o
no
fef
a
c
i
l
i
t
yi
ss
t
u
d
e
n
tt
r
a
i
n
i
n
g
.F
i
r
s
te
i
n

to
u
t
s
i
d
eo
fp
r
o
c
e
s
s
i
n
gf
o
ra
l
lc
h
i
p
si
s6
0
0h
o
u
r
s
.T
o
u
c
h
h
a
n
d
l
i
n
gt
i
m
ec
o
n
s
s
e
n
s
o
r
sr
e
q
u
i
r
e4h
o
s
LCD9h
o
u
r
sp
r
e
s
s
'
es
e
n
s
o
r
s7h
o
u
r
s
a
n
dc
o
n
o
l
l
e
r
s1
0
t
h
et
i
m
ea
v
a
i
l
a
b
l
eo
nt
h
el
i
t
h
o
g
r
a
p
h
i
cm
a
c
h
i
n
e
si
sa
b
o
u
t4
2
0h
o
u
r
s
.
h
o
u
r
s
.S
e
c
o
n
d
u
i
r
e1h
o
u
r

p
r
e
s
s
u
r
es
e
n
s
o
r
s3h
o
u
r
s
a
n
dc
o
n
t
r
o
l
l
e
r
s8
T
o
u
c
hs
e
n
s
o
r
sa
n
dLCDr
h
o
u
r
s
.P
a
c
k
a
g
i
n
gc
o
n
s
i
d
e
r
a
t
i
o
n
sp
l
a
c
et
h
emaximuma
t8
0
0v
o
l
u
m
eu
n
i
t
s
.T
o
u
c
h
s
e
n
s
o
r
sr
e
q
u
i
r
e3
0v
o
J
u
r
n
eu
n
i
t
s
LCD40v
o
l
u
m
eu
n
i
t
s
p
r
e
s
s
'cs
e
n
s
o
r
s2
0u
n
i
t
s
a
n
d
c
o
n
o
l
l
e
r
1
0u
n
i
t
sb
e
c
a
u
s
eo
fe
i
rc
o
m
p
a
c
ts
i
z
e
.A
l
lec
o
n
s
a
i
n
t
sa
b
o
v
ea
r
e
i
n
d
i
t
e
dp
e
rweeko
fo
p
e
r
a
t
i
o
no
ft
h
ef
a
c
i
l
i
t
y
. en
e
tr
e
v
e
n
u
ei
s$
6
$
1
0

$
9
d$20
f
o
rt
h
et
o
u
c
hs
e
n
s
o
r

LCDp
s
s
u
r
es
e
n
s
o
r

a
n
dc
o
n
t
r
o
l
l
e
r

r
e
s
p
e
c
t
i
v
e
l
y
.l
ef
a
c
i
l
i
t
y
k
et
od
e
t
e
r
m
i
n
et
h
er
i
g
h
t
i
si
n
t
e
r
e
s
t
e
di
nm
a
x
i
m
i
z
i
n
gr
e
v
e
n
u
ep
e
rweeka
n
dw
o
u
l
di
mixof f
o
u
rd
e
v
i
c
e
s
.
ProblemFormulation: T
he f
o
r
m
u
l
a
t
i
o
ni
ss
t
r
a
i
g
h
t
f
o
r
w
a
r
db
a
s
e
do
nt
h
e
e
t
x( r
e
p
s
e
n
tt
h
en
u
m
b
e
ro
ft
o
u
c
hs
e
n
s
o
rc
h
i
p
sp
e
rw k
X2E
s
t
a
t
e
m
e
n
t
sa
b
o
v
e
.L

x3t
h
en
u
m
b
e
ro
fp
r
e
s
s
u
r
es
e
n
s
o
r
s

a
n
dX4en
u
m
b
e
ro
fc
o
n
o
l
l
e
r
s
.
n
u
m
b
e
ro
fLCD
o
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
ni
s
M
a
x
i
m
i
z
e
:

f
.6x.+lOx2+9x3+20x4

Theh
a
n
d
l
i
n
gt
i
m
ec
o
n
s
t
r
a
i
n
tc
a
nb
ee
x
p
r
e
s
s
e
d

Henc

M
i
n
i
m
i
z
e

f
.-6x1-1O~ -9x3- 20x4

Thes
y
m
b
o
l/h
a
sb
e
e
nr
e
t
a
i
n
e
df
o
rc
o
n
v
e
n
i
e
n
c
ee
v
e
nt
h
o
u
g
ht
h
ed
I
I
c
t
i
o
no
f
ec
o
n
s
i
n
t
sa
r
es
e
tu
pu
s
i
n
gs
l
a
c
kv
a
r
i
a
b
l
e
sS
I
S2
a
n
d
s
o
p
t
i
m
i
z
a
t
i
o
nh
a
sc
h
a
n
g
e
d
. l

3
g
.
:4
xI
+9x
x
0
x
.=6
0
0
2+7
4+S
3+1
.+~ +3x
x
20
g
2
:x
4+s2=4
3+8
0
x
.+40x
0x
O
x4+S3=8
0
0
g
3
:3
2+2
3+l
Al
lv
a
r
i
a
b
l
e
sa
r
e~ O
. l
ef
o
r
m
u
l
a
t
i
o
na
b
o
v
es
u
g
g
e
s
t
st
h
es
t
a
n
d
a
r
dS
i
m
p
l
e
xm
e
t
h
o
d
.

S '
P
1
e
xTable1
:T
a
b
l
e
3
.
1
6
i
se
i
n
i t
a
b
l
ef
o
rE
x
a
m
p
l
e3.
4
.Thec
a
n
o
n
ic
a
If
o
r
m
i
so
b
s
e
r
v
a
b
l
ef
r
o
mt
h
et
a
b
l
ea
n
dt
h
ei
n
i
t
i
a
lb
a
s
i
cf
e
a
s
i
b
l
es
o
l
u
t
i
o
nc
a
nb
ed
e
t
e
r
m
i
n
e
d
.
T
a
b
l
e3
.
1
6E
x
a
m
p
l
c3
.
4
:S
i
m
p
l
e
xT
a
b
l
e1
X
I
4

X2

X
J
7

3
0

4
0

3
2
0

1
0

X4

1
0
8
1
0
2
0

8
1

J
"

S
J

4
2
0

8
0
0
f

3
.
4A
D
D
I
T
I
O
N
A
LE
X
A
M
P
L
E
S 137

L
lN
E
A
RP
R
O
G
R
A
M
M
I
N
G

136

UsingM
A
T
L
A
B

T
a
b
l
e3
.
1
7E
x
a
m
p
l
e3
.
4
:S
i
m
p
l
e
xT
a
b
l
e2

X2

X3

2
.
7
5
0
.
1
2
5
2
8
.
7
5

7
.
7
5
0
.
1
2
5
3
8
.
7
5

3
.
2
5
0
.
3
7
5
1
6
.
2
5

3
.
5

7
.
5

1
.5

X4

8
J

8
2

7525.5
2
7
5

1
.
2
5
0
.
12
5

1
.
2
5

2
.
5

f+1050

S
I
S
2
.a
n
d$
3a
r
et
h
eb
a
s
i
cv
a
r
i
a
b
l
e
s
.TheEBVi
sX4a
n
d eLBVi
s$
2w
h
i
c
hi
sE
min
l
umo
f{
6
0
.5
2
.
58
0
}
.Thep
i
v
o
trowi
st
h
es
e
c
o
n
drowu
s
e
df
o
rt
h
erow
m
a
n
i
p
u
l
a
t
i
o
n
st
ol
e
a
dt
oT
a
b
l
e3
.
1
7
.

S
i
m
p
l
e
xT
a
b
l
e2
:Usingt
h
ep
i
v
o
trowi
d
e
n
t
i
f
i
e
di
nel
a
s
tt
a
b
l
eeu
n
i
lv
e
c
t
o
r[
0

o1O]Tunderthe$2columnneedstobetransferredtoeX4 columnthrough

a
n
o
n
i
c
a
lf
o
r
ma
f
t
e
rc
o
m
p
l
e
t
i
o
no
f
e
l
e
m
e
n
t
a
r
yrowo
p
e
r
a
t
i
o
n
s
.T
a
b
l
e3
.
1
7showsec
eEBVi
sX2a
n
deLBV
a
l
lt
h
eo
p
e
r
a
t
i
o
n
s
.Theb
a
s
i
cv
a
r
i
a
b
l
e
sa
r
eX4.$" dS
Th
3
i
sS
3l
ep
i
v
o
trowi
st
h
ei
r
dr
o
w
.Theo
b
j
e
c
t
i
v
eo
ft
h
erowm
a
n
i
p
u
l
a
t
i
o
n
si
st
o
t
r
a
n
s
f
e
rt
h
eu
n
i
tv
e
c
t
o
r[
001O
]
Tf
r
o
mt
h
e$
3columnt
o e
x
z
c
o
l
u
m
n
.

S
i
m
p
l
e
xT
a
b
l
e3
:T
a
b
l
e3
.
1
8d
e
n
o
t
e
st
h
er
e
d
u
c
e
dt
a
b
l
ew
i
t
h
ec
a
n
o
n
i
ca
Jf
o
r
ma
f
t
e
r
eX2.x4

a
n
dS
IT
h
e
r
e
t
h
er
e
q
u
i
r
e
drowo
p
e
r
a
t
i
o
n
sa
r
ec
o
m
p
l
e
t
e
d
.Theb
a
s
i
sv
a
r
i
a
b
l
m non
e
g
a
t
i
v
ec
o
e
f
f
i
c
i
e
n
t
si
nt
h
el
a
s
trows
u
g
g
e
s
t
i
n
gt
h
es
o
l
u
t
i
o
nh
a
sb
e
e
no
b
t
a
i
n
e
d
.
Fromet
a
b
l
e s
o
l
u
t
i
o
ni
s
XJ=0
x2=7
.
0
9
6

x3=0

x4=5
1
.6
1/=1
1
0
3
.
2
2
Thes
o
l
u
t
i
o
na
b
o
v
ei
sn
o
ts
a
t
i
s
f
a
c
t
o
r
yb
e
c
a
u
s
et
h
ea
c
t
u
a
ld
e
c
i
s
i
o
nwi
I
li
n
v
o
l
v
ei
n
t
e
g
e
r
v
a
l
u
e
sf
o
rt
h
ed
e
s
i
g
nv
a
r
i
a
b
l
e
s
.Att
h
i
st
i
m
ei
n
t
e
g
e
r
p
r
o
g
r
a
m
m
i
n
gi
sn
o
ta
no
p
t
i
o
n
.The
a
d
j
u
s
t
e
ds
o
l
u
t
i
o
ni
s
X1=O
x

X4=5
2/=1
1
1
0
2=7
3=O
sc
h
o
i
c
es
a
t
i
s
f
i
e
sec
o
n
s
t
r
a
i
n
t
sg
3
.g
2howevern
e
e
d
saI
i
t
t
l
ee
l
a
s
t
i
c
i
t
y(
l
i
m
i
t
Whilei
i
s4
2
3
)t
ob
es
a
t
i
s
f
i
e
d
.Not
a
ts
Ji
sab
a
s
i
cv
a
r
i
a
b
l
es
oc
o
n
s
t
r
a
i
n
tg
Js
h
o
u
l
dn
o
tb
ea
p
r
o
b
l
e
m
.

X2

X
J

3
0
.
0
3
2
2
5
80 0
.
3
2
2
5
8
1
0
.
7
4
1
9
3
51 0
.
4
1
9
3
5
5
2
.
0
6
4
5
1
60 1
.
6
4
5
1
6
1

X4

h
e
l
pl
p
LP Linear programming.
X=LP(f
A
b) solves the linear programming problem:
min f 'x subject t
o: Ax <= b
A
b
VLB
VUB) defines a set of lower and upper
X=LP(f
bounds on the designvariablesXso that the solution
is always in the range VLB <= X <= VUB.
X=LP(f
A
b
VLB
VUB
XO) sets the initial starting point
to XO.
X=LP(f
A
b
VLB
VUB
XO
N) indicates that the first N
constraintsdefinedbyA andb areequalityconstraints.
A
b
VLB
VUB
XO
N
DISPLAY) controls the level of
X=LP(f
warning messages displayed. Warning messages can be
turned off with DISPLAY -1.
LAMBDAJ= LP(f
A
b) returns the se of Lagrangian
[
x
multipliers
LAMBDAat the solution.
[
X
LAMBDA
HOW]
LP(f
A
b) also returns a string how
that indicates error conditions at the final iteration.
LPproduceswarningmessageswhenthe solutioniseither
unbounded or infeasible.
c= [
6 -10 -9 -20];
A= [
4 9 7 10; 1 1 3 8; 30 40 20 10];
b = [600 420 800];
x=[cAb]

x =lp(cAb)

Warning: The solution is unbounded and at infinity;


the constraints are not restricive enough.

T
a
b
l
e3
.
1
8E
x
a
m
p
l
e3
.
4
:S
i
m
p
l
c
xT
n
b
l
e3
x

T
h
i
si
sas
t
a
n
d
a
r
dLPp
r
o
b
l
e
m
.T
h
i
ss
e
c
t
i
o
nu
s
e
sMATLABi
na
ni
n
t
e
r
a
c
t
i
v
es
e
s
s
i
o
nt
o
s
o
I
v
e ep
r
o
b
l
e
m
.l
ef
o
l
l
o
w
i
n
gi
sad
i
a
r
yf
i
l
ef
o
rt
h
eMATLABs
e
s
s
i
o
nf
o
ri
s
n
p
l
e
.N
o
t
et
h
ewamingf
r
o
mMAT
L
A
B
.a
swe
Iles
u
b
s
e
q
u
e
n
tu
s
eo
ft
h
eLP
e
x
p
r
o
g
r
a
mu
s
i
n
gb
o
u
n
d
soned
e
s
i
g
nv
a
r
i
a
b
l
e
s
:l
eb
o
l
di
t
a
l
i
c
i
z
e
ds
t
a
t
e
m
e
n
t
sa
r
e
.
t
commandsu
s
e
dt
oo
b
t
a
i
nt
h
es
o
l
u
t
i
o
n
.Th
eMATLABp
r
o
m
p
ti
se
d
i
t
e
do
u

x
8
2

$
)

0
.
2
2
0
0
.
0
0
3
2
3 56
9
0
3
2
2 o 02
.032258 0.025806 7.019269774
2.25806451
.
1
-0

1
.

0
.
1
9
3
5
4
8 f+1
1
0
3
.
2
2

1.0e+015 *
-0.0000
3.7067
8
.8391
2.8513

138 L
lN
EARPROGRAMMING

3
.
5A
D
D
I
T
I
O
N
A
LT
O
P
I
C
SI
NL
lN
EARPROGRAMMING

139

o
r
m
a
lmaximump
r
o
b
l
e
m
.
T
h
i
si
sa
l
s
or
e
f
e
r
r
e
d1
0 an
l
ed
u
a
lo
ft
h
eabovep
r
o
b
l
e
mi
sd
e
f
i
n
e
da
s

vlb= [0 0 0];
vub [100100100];
x= lp(cAbvlbvub)

;
:

x
0.0000
7.0968

51.6129
T
h
i
si
st
h
cs
o
l
u
t
i
o
no
b
t
a
i
n
e
dbyw
o
r
k
i
n
gt
h
r
o
u
g
ht
h
eS
i
m
p
l
e
xm
e
t
h
o
d
.

Minimize

w:b
l
Y
I+blY
2+.
.
.
+b
"
J
'
r
n

S
u
b
j
e
c
t1
0
: a
2
1
Y
2+..
.
+a
r
n
l
Y
m~ c
J
l
l
Y
I+a

(
3.
43
)

(
3
.
4
4
)

a
l2
Y1+a
2
2
Y
2+.
.
.
+a
2
Y
m~ c
2
m
a
l
J
'1.
.
.a
2
"
Y
2+.
.
.
+atuJ'" C"
YI~O

lNEARPROGRAMMING
3
.
5 ADDITIONALTOPICSINL
Thi
ss
e
c
t
i
o
nl
o
o
k
sb
r
i
e
f
l
ya
ta
d
d
i
t
i
o
na
Ji
d
e
a
sa
s
s
o
c
i
a
t
e
dw
i
t
h
eLPp
r
o
b
l
e
m
.F
i
r
s
t
e
r
ei
sad
i
s
c
u
s
s
i
o
no
ft
h
ed
u
a
lproblema
s
s
o
c
i
a
t
e
dw
i
t
ha
nLPp
r
o
b
l
e
m
.D
u
a
l
i
t
yi
s
impo
a
n
t
i
n d
i
s
c
u
s
s
i
o
no
fs
e
n
s
i
t
i
v
i
t
yanalysis-var
i
o
ni
no
p
t
i
m
i
z
a
t
i
o
ns
o
l
u
t
i
o
n
d
u
et
ov
a
r
i
a
t
i
o
ni
nt
f
t
eo
r
i
g
i
l
1
p
a
r
a
m
e
t
e
r
so
ft
l
t
ep
r
o
b
l
e
m
.Followingi
si
sa
d
i
s
c
u
s
s
i
o
no
fs
e
v
e
r
a
lp
i
e
c
e
so
fi
n
f
o
r
m ont
h
a
tc
a
nb
ee
x
c
t
e
dfromef
i
n
a
Jt
a
b
l
e
o
ft
h
eS
i
m
p
l
e
xm
e
t
h
o
d
.

i=1
2
...m

(
3
.
4
5
)

R
e
l
a
t
i
o
n
s(
3.
43
)(
3.
4
5
)d
e
s
c
r
i
b
ean
o
r
m
a
lminimump
r
o
b
l
e
m
.Therei
sa
l
s
oa
ni
n
v
e
r
s
e
r
e
l
a
t
i
o
n
s
h
i
pb
e
t
w
e
e
n d
e
f
i
n
i
t
i
o
no
ft
h
e two p
r
o
b
l
e
m
s above a
sf
a
ra
s
i
d
e
n
t
i
f
i
c
a
t
i
o
no
ft
h
ep
r
i
m
a
landd
u
a
lp
r
o
b
l
e
m
.I
f l
a
t
t
e
ri
sc
o
n
s
i
d
e
r
e
dap
r
i
m
a
l
probleme
nt
h
ef
o
r
m
e
ri
sec
o
r
r
e
s
p
o
n
d
i
n
gd
u
a
lp
r
o
b
l
e
m
.
l
ef
o
l
l
o
w
i
n
go
b
s
e
r
v
a
t
i
o
n
sc
b
emadew
i s
p
e
c
t1
0t
h
ep
a
i
ro
fp
r
o
b
l
e
m
s
e
s
t
a
b
l
i
s
h
e
da
b
o
v
e
:
1
.Thenumbero
fd
u
a
lv
a
r
i
a
b
l
e
si
s same enumbero
fp
r
i
m
a
lc
o
n
s
t
r
a
i
n
t
s
.

3
.
5
.
1 PrlmalandDualProblem

2
.Thenumbero
fd
u
a
lc
o
n
s
t
r
a
i
n
t
si
st
h
csamea
st
h
enumbero
fp
r
i
m
a
lv
a
r
i
a
b
l
e
s
.

e
v
e
r
yLPp
r
o
b
l
e
m
whichw
i
l
lber
e
f
e
r
r
e
dtoap
r
i
m
a
lproblem
A
s
s
o
c
i
a
t
e
dwi
t
h
e
r
ei
s ac
o
r
r
e
s
p
o
n
d
i
n
gd
i
1
p
r
o
b
l
e
m
.I
nm y waysedua
J pl'Oblemi
sa
s
p
o
s
i
t
i
o
noft
h
ep
r
i
m
a
Jp
r
o
b
J
e
m
.lIsi
smeantt
oi
m
p
J
yt
h
a
ti
ft
h
ep
r
i
m
a
lprobJem
i
n
i
m
i
z
a
t
i
o
nprobJemedua
Jp
roblemw
i
l
lbeam
a
x
i
m
i
z
a
t
i
o
no
n
e
.l
ft
h
eprim
a
J
i
sam
p
r
o
b
l
e
mh
a
s1
1v
a
r
i
a
b
l
e
st
h
ed
u
a
lp
r
o
b
l
e
mw
i
l
lh
a
v
enc
o
n
s
t
r
a
i
n
t
s
.l
ft
h
ep
r
i
m
a
l
p
r
o
b
l
e
mh
a
smc
o
n
s
t
r
a
i
n
t
s
t
h
edua
Jp
r
o
b
l
e
mw
i
l
lhavemv
a
r
i
a
b
l
e
s
.Fort
h
ed
i
s
c
u
s
s
i
o
n
t
a
n
d
a
r
dp
r
i
m
C
l
lp
r
o
b
l
e
mi
sg
e
n
e
r
a
l
l
yd
e
f
i
n
e
da
sam
a
x
i
m
i
z
a
t
i
o
n
o
fd
u
a
l
i
t
yes
problemw
i
t
h~ i
n
u
a
l
i
t
i
e
s(
t
h
o
u
g
hes
t
a
n
d
a
r
dp
r
o
b
l
e
mi
nt
h
eLPprogramming
d
i
s
c
u
s
s
i
o
n was am
i
n
i
m
i
z
a
t
i
o
n problem [
s
Eq
u
a
t
i
o
n
s(
3.
10
)
(
3.
12
)
] d e
ns Intswere ualities).Thes rdprimaJprobleminthissectioni
sd
e
f
i
n
e
da
s[
3
]

Maximize

(
3.
4
0
)

z
:CIXI+CIX2+...+c

S
u
b
j
e
c
t1
0
: a
x
l
2
2+.
.
.
+a
l
.
x
n~ b
l
l
x
l+a

(
3.
41
)

a
2
1
x
I+a2~2 +.
.
.
+az"xn~ b2

"I
X
I+a
naX
2+ +a
n
"

x
n~bm
a

O
j=1

2.....n

(
3
.
4
2
)

3
. Thec
o
e
f
f
i
c
i
e
n
tm
a
t
r
i
xA o
ft
h
ep
r
i
m
a
lproblemi
st
r
a
n
s
p
o
s
e
dt
op
r
o
v
i
d
eE
c
o
e
f
f
i
c
i
e
n
tm
a
t
r
i
xo
ft
h
ed
u
a
lp
r
o
b
l
e
m
.
u
a
l
i
t
i
e
s'Cr
e
v
e
r
s
e
di
nd
i
r
e
c
t
i
o
n
.
4
.Thein

5
.ThemaximizationprobJemoft
h
ep
r
i
ma
Jp
roblembecomesam
i
n
i
m
i
z
a
t
i
o
n
probJemi
nt
h
d
u
a
lp
r
o
b
l
e
m
.
6
.Th
ec
o
s
tc
o
e
f
f
i
c
i
e
n
t
so
ft
h
cp
r
i
m
a
lp
r
o
b
l
e
mbecomet
h
er
i
g
h
t
h
a
n
ds
i
d
eva
Ju
e
s
o
ft
h
ed
u
a
Jpl'Ob
l
e
m
.Ther
i
g
h
t
h
a
n
ds
i
d
ev
a
Ju
e
so
ft
h
ep
r
i
m
a
lbecomet
h
ec
o
s
t
c
o
e
f
f
i
c
i
e
n
t
so
ft
h
ed
u
a
lp
r
o
b
l
e
m
.

7
.Thep
r
i
m
a
landdua
Jv
a
r
i
a
b
l
e
sb
o
t
hs
a
f
yt
h
en
o
n
n
e
g
a
t
i
v
i
t
yc
o
n
d
i
t
i
o
n
.
Example3.5l
ef
o
l
l
o
w
i
n
gexamplea
d
a
p
t
e
dandm
o
d
i
f
i
e
dfromR
e
f
e
r
e
n
c
e3
i
s
u
s
e
dt
oi
I
Iu
s
t
r
a
t
et
h
ep
r
i
m
a
l
/
d
u
a
lv
e
r
s
i
o
n
so
f
t
h
esamep
r
o
b
l
e
m
.Theproblemi
ss
o
l
v
e
d
u
s
i
n
gt
e
c
h
n
i
q
u
e
sd
e
v
e
l
o
p
e
di
nt
h
i
sc
h
a
p
t
e
r
.
eP
roblem:Thel
o
c
a
lc
o
l
l
e
g
e
'
sS
c
h
o
o
lo
fAm
e
r
i
c
a
nC
r
a
f
t
s
m
a
nh
a
sd
e
c
i
d
e
d1
0
p
a
r
t
i
c
i
p
a
t
es
e
r
i
o
u
s
l
yi
na1
0
ca
Jc
h
a
r
i
t
ye
v
e
n
tbym
a
n
u
f
a
c
t
u
r
i
n
gs
p
e
c
i
a
lcommemorative
d
e
s
k
s
t
a
b
l
e
s
andc
h
a
i
r
s
.Each1
y
p
eo
ff
u
m
i
t
u
r
er
e
q
u
i slumbera
sweI
la
stwot
y
p
e
s
o
fs
k
i
l
l
e
dl
a
b
o
r
:f
m
i
s
h
i
n
g dc

'
P
e
n
t
r
y
.T
a
b
l
e3
.
1
9documentst
h
er
e
s
o
u
r
c
e
sn
e
e
d
e
d
.
Jl
umberc
o
o
p
e
r
a
t
i
v
eh
a
sd
o
n
a
t
e
d48b
o
a
r
df
e
e
to
flumb
r
.F
a
c
u
l
t
ya
n
ds
t
a
f
f
The1
0
ca
h
a
v
ev
o
l
u
n
t
d2
0f
i
n
i
s
h
i
n
gh
o
u
r
sand8c
a
r
p
e
n
t
r
yh
o
u
r
s
.Thed
e
s
kw
i
l
lb
es
o
l
df
o
r

3
.
5A
D
D
1
T
1
0
N
A
lT
O
P
1
C
SI
NUNEAAPAOGAAMM1NG 1
4
1

140 l
l
N
E
A
RPROGAAMM1NG
l
i
l1
l

T
n
b
l
c3
.
1
9 Example3
.
5
:Rources
R
e
s
o
u
r
c
e
s
d
f
e
e
t
}
L
u
m
b
e
r
(
b
o
F
i
n
i
s
h
i
n
gh
o
u
r
s(
h
o
u
r
s
)
C
a
r
p
e
n
t
r
y(
h
o
u
r
s
)

De
s
k

T
a
b
l

8
4
2

6
2
1
.5

C
h
a
i
r

21 Examplc3
.
5
:P
r
b
n
a
lP
r
o
b
l
c
m

S
i
m
p
l
c
xT
a
b
l
c2
T
a
b
l
c3.

X2

X3

0
.
7
5

0
.
5
0
.
2
5

$
2

$
3

$
4

1
.5
0
.
5

1
5

4
-2
0
.
5

-5

3
0

b
1
6
4
4
5
f+240

$60
et
a
b
l
ef
o
r$30
a
nd c
h
a
i
rf
o
r$
2
0
.Thes
c
h
o
o
ld
o
e
sn
o
te
x
p
e
c
tt
os
e
J
1more
v
et
a
b
l
e
s
.Thes
c
h
o
o
lwouldl
i
k
et
omaximiz
r
e
v
n
uec
oJ
1e
c
t
e
d
.

i
s8
.Not
a
b
l
e
sw
i
l
lbemade.l
et
o
t
a
Jr
e
v
e
n
u
e
o
p
t
i
m
a
lnumbero
fc
h
a
i
r
st
obemade
i
s
$
2
8
0
.
0
0
.

8
t
a
n
d
a
r
dForma P
r
i
m
a
lP
r
o
b
l
e
m
:Defining X
l a
st
h
e number o
fd
e
s
k
s
p
r
o
d
u
c
e
d

X2a
st
h
enumbero
ft
a
b
l
e
sm
a
n
u
f
a
c
t
u
r
e
d

andX3a
st
h
enumbero
fc
h
a
i
r
s
m
a
n
u
f
a
c
t
u
r
e
d
t
h
ep
r
i
m
a
lp
r
o
b
l
e
mc
a
nbee
x
p
r
e
s
s
e
di
ns
t
a
n
d
a
r
df
o
r
m
a
t

.
1ndardFormat
D
ualP
r
o
b
l
e
m
:Whil
et
h
ed
ua
Jp
roblemc
a
nbx
p
r
e
s
s
e
d d
s
e
t
u
pm h i
c
a
1
ly
i
ti
s c
u
l
tt
oa
s
s
o
c
i
a
t
eed
e
s
i
g
nv
a
r
i
a
b
l
e
si
nad
i
r
e
c
tm n
e
r
aswasdone t
h
eprim
a
Jp
r
o
b
l
e
m
.Usingy
"Y2Y3andY4a
st
h
ed
e
s
i
g
nv
a
r
i
a
b
l
e
s
t
h
e
problemc
a
nbef
o
r
m
u
l
a
t
e
da
si
nE
q
u
a
t
i
o
n
s(
3.
48
)and(
3
.
4
9
)belowu
s
i
n
gt
h
e
e
g
i
n
n
i
n
go
ft
h
i
ss
e
c
t
i
o
n
.Somea
s
s
o
c
i
a
t
i
o
n
sa
r
ep
o
s
s
i
b
l
ei
nl
i
g
h
to
f
d
e
f
i
n
i
t
i
o
na
teb
t
r
a
n
s
p
o
s
i
t
i
o
no
fec
o
s
tc
o
e
f
f
i
c
i
e
n
t
sandt
h
ec
o
n
s
t
r
a
i
n
t
s
.Forexample

t
h
ev
a
r
i
a
b
l
ey
c
a
nbea
s
s
o
c
i
a
t
e
dw
i
t
ht
h
el
u
m
b
e
rc
o
n
s
t
r
a
i
n
tb
e
c
a
u
s
ei
nt
h
ed
u
a
lproblemi
t
sc
o
s
t
c
o
e
f
f
i
c
i
e
n
ti
st
h
el
u
m
b
e
rc
o
n
s
t
r
a
i
n
to
fep
r
i
m
a
lp
r
o
b
l
e
m
.I
nas
i
m
i
l
a
rmanner

Y2i
s
a
s
s
o
c
i
a
t
e
dw
i
t
ht
h
ef
i
n
i
s
h
i
n
gh
o
u
r
s
Y3w
i
t
h c
a
r
p
e
n
t
r
yh
o
u
r
s
andY4w
i
l
ht
h
es
p
e
c
ia
J
t
a
b
l
ec
o
n
s
t
r
a
i
n
t. o
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
no
ft
h
ed
u
a
lproblemi
se
x
p
r
e
s
s
e
di
nE
q
u
a
t
i
o
n
(
3.
48
)
.I
nviewofeformoft
h
ee
x
p
r
e
s
s
i
o
no
ft
h
eo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
nw
E
i
sa
p
o
s
s
i
b
i
l
i
t
yt
h
a
ted
e
s
i
V i
a
b
l
i
nt
h
edu
a
Jp
roblemc
a
nhave e
c
o
n
o
r
n
i
c
i
m
p
l
i
c
a
t
i
o
na
sf
ol
1
o
w
s
:

(
3
.
4
6
)

M
a
x
i
m
i
z
ez
:60x +30X2+20X3

(
3.
4
7
a
)

Su
e
c
tt
o
: g 8x +6
x
:
:
;4
8
2+X3:

.5X3:
:
:
;2
0
g
2
:4x +2x
2+1

(
3.
4
7
b
)

x
+1
.5x2+0.5X3:
:
;8
g
3
:2

(
3
.
4
7
c
)

1
!
4
:X2::;5

(
3.
4
7
d
)

Xl
>X2
x3~O

unitpriceforaboardfootoflumber

Jp
r
o
b
l
e
mi
sa
v
a
i
1a
b
l
ei
nT
a
b
l
e
s3
.
2
0
3
.
2
2
.T
a
b
l
e3
.
2
0i
st
h

Thes
o
l
u
t
i
o
nt
ot
h
ep
r
i
ma
i
n
i
t
ia
Jt
a
b
l
e
.TheEBVi
sx a
n
dt
h
eLBVi
s$
3Thep
i
v
o
trowi
st
h
ei
r
dr
o
w
.The
s
u
b
s
e
q
u
e
n
trowo
p
e
r
a
t
i
o
n
sd
e
t
e
r
m
i
n
eT
a
b
l
e3
.
2
1
.I
nt
h
i
st
a
b
l
eeEBVisx3'theLBV

zmdt
h
ep
i
v
o
trowi
st
h
eS
E
c
o
n
dEow.Table322ist
h
ea
n
a
lt
a
b
l
ee
s
t
a
b
l
i
s
h
i
n
gE
i
ss
z
o
p
t
i
m
a
lv
a
l
u
e
sf
o
rt
h
ev
a
r
i
a
b
l
e
s
.x
teoptimaJnumberofdesksequals2.~e

y
Y
2
Y
3
y
4

T
a
b
l
e3
.
2
0 Exnmple3
.
5
:P
r
i
m
a
lP
r
o
b
l
c
m

S
i
m
p
l
e
xT
a
b
l
e1

X2

B
4
2

6
2
1
.5

-60

3
0

X)

1
.5
0
.
5

-20

$
2

S)

$
4

o

o

b
4
8
2
0
8
5
f+O

u
n
i
tp
r
i
c
ef
o
raf
i
n
i
s
h
i
n
ghour
u
n
i
tp
r
i
c
ef
o
ra
nhouro
fc
a
r
p
n
t
r
y
i
scannotbeeu
n
i
tp
r
i
c
ea
sed
e
s
kh
a
sap
r
i
c
edefmedi
nt
h
ep
r
i
m
a
J
problem-maybeas
p
e
c
ia
Jp
r
i
c
ef
o
rt
h
ed
e
s
kr
e
s
o
u
r
c
e

S
i
m
p
l
e
xT
a
b
l
e3
T
a
b
l
e3
.
2
2 Example3
.
5
:P
r
i
m
a
lProblem

o
o

X2

-2
-2
1
.2
5
5

X3

$
2

.
1
'
3

$
4

2
4
8
2
5

f+280


o
o

2
2
0
.
5

-8
4
1
.5

1
0

1
0

1
4
2

lN
E
A
RP
R
O
G
R
A
M
M
I
N
G 1
3
.
5A
O
O
I
T
I
O
N
A
LT
O
P
I
C
SI
NL
4
3

L
lN
E
A
RP
R
O
G
R
A
M
M
I
N
G

T
a
b
l
e3
.
2
4E
x
a
m
p
l
e3
.
5
:D
u
a
lP
r
o
b
l
e
m

P
b
a
s
e1

T
a
b
l
e2

Thed
u
a
lc
a
nb
ee
x
p
r
e
s
s
e
da
s
Min
i
m
i
z
ew
:4
8
Y
I+2
0
Y
2+8
Y
3+5
Y
4

(
3.
48
)

S
u
b
j
e
c
tt
o
: hl
:
8
Y
I+4
Y
2+2
Y
3+O
Y
46
0

(
3.
4
9
a
)

h
:
6
Y
I+2
Y
2+1
.
5
Y
3+l
Y
4~ 3
0
2

(
3.
4
9
b
)

:l
h
Y
I+1
.5
Y
2+0
.
5
Y
3+O
Y
4~ 2
0
3

(
3
.
4
9
c
)

YI>Y
2

Y
3~ 0

M
i
n
i
r
n
i
z
ew
:4
8
Y
I+2
0
Y
2+8
Y
3+5
Y
4

(
3.
48
)

:
8
Y
l+4
S
u
b
j
e
c
t
t
o
: hl
Y
2
+2
Y
3-S
I+a
l=60

(
3
.
5
1
a
)

h
:
6
)
'
1+2
Y
2+1
.5
Y
3+l
Y
4-S
2+a
2=3
0
2

(
3
.
5
I
b
)

.5
Y
2+0.
5
Y
3+O
Y
4-S
3+a
3=2
0
h
:l
Y
l+1
3

(
3
.
5
1
c
)

[
nP
h
a
s
e1o
ft
h
et
w
o
p
h
a
s
ea
p
p
r
o
a
c
hea
r
t
i
f
i
c
i
a
lv
a
r
i
a
b
l
e
sa
r
er
e
m
o
v
e
df
r
o
mE
b
a
s
i
su
s
i
n
ga
na
r

c
i
a
lc
o
s
tf
u
n
c
t
i
o
nAd
e
f
i
n
e
da
s
(
3
.
5
2
)

M
i
n
i
m
i
z
e
A
: al+a2+a3

Th
eP
h
a
s
e[c
o
m
p
u
t
a
t
i
o
n
sa
r
ed
i
s
p
l
a
y
e
di
nT
a
b
l
e
s3
.
2
3
3
.
2
6
.T
a
b
l
e3
.
2
3i
si
d
e
n
t
i
f
i
e
d
s
i
m
p
l
e
xT
a
b
l
e1

a
n
di
ti
n
c
1
u
d
e
ssomep
r
e
p
r
o
c
c
s
s
i
n
gb
yw
h
i
c
ha"a
2
a
n
da
3 e
a
st
h
madeeb
a
s
i
sv
a
r
i
a
b
l
e
s
.Asar
e
m
i
n
d
e
r
t
h
el
a
s
trow p
r
e
s
e
n
t
sea
r
t
if
i
c
i
a
ls
t
f
u
n
c
t
i
o
n
.Fromi
n
s
p
e
c
t
i
o
no
f
T
a
b
l
e3
.
2
3
Y
Ii
st
h
eEBV.TheLBVi
sa
2a
n
d p
i
v
o
t
rowi
st
h
es
e
c
o
n
dr
o
w
.F
o
l
l
o
w
i
n
gt
h
es
t
a
n
d
a
r
drowo
p
e
r
a
t
i
o
n
s
T
a
b
l
e3
.
2
4i
so
b
t
a
i
n
e
d
.
t
h
eEBVi
sY
2(
el
a
r
g
e
s
tn
e
g
a
t
i
v
ec
o
e
f
f
i
c
i
e
n
t
)
.

l
eLBVi
s
a
3(
E i
r
d
[
nt
h
i
st
a
b
l
e

Y
3

8
6

4
2
1
.
5

4
8
1
5

2
0
7
.
5

1
.
5

0
.
5
8

Y
4

5
1

5
2

3333 0
3
30
.
2
5
03
o
6
6
60
.
2
5
4
1
.
.
3
1
.
1

o2
.
5

-0
.
2
5

Y
4
1
.3
3
3
1

0.
16
6
6
0
.
1
6
6
7
3
1
.5

5
1

5
2

S
)

1
.5

a
2

a
)

3333
2
0
0
.
1
6
6
6 5
oo -8.1666 o w-152
2.5 A-35

1
.3
3
3
3
o -0.
1
6
6
6 o
0
.
1
6
6
61
8

a
1

1
.
-0

S)

r
o
wc
o
n
t
a
i
n
st
h
er
n
i
n
i
m
u
mo
f(
[
2
0
/1
.3
3
3
]
[
5
1
0
.
3
3
3
3
][
1
5
/1
.
6
6
6
7
]
}
)
. l
ep
i
v
o
tr
o
wi
s
r
o
w3
.T
a
b
l
e3
.
2
5r
e
p
s
e
n
t
st
h
n
e
x
tt
a
b
l
e
. leEBViss
o
t
ei
nt
h
i
sc

t
h
e
r
ea
r
et
w
o
N
2
S
2

S
3
)
'S
2i
sc
h
o
s
e
na
r
b
i
t
r
a
r
i
l
y
.l
eLBVi
sa
l
'Thep
i
v
o
tr
o
wi
sE
c
a
n
d
i
d
a
t
e
sf
o
rEBV(
t
r
o
w
.A
f
t
e
r
t
h
e[u
i
r
e
dr
o
wm
a
n
i
p
u
l
a
t
i
o
n
s
T
a
b
l
e3
.
2
6i
sg
e
n
e
r
a
t
e
d
.S
c
a
n
n
i
n
gt
h
el
a
s
t
r
o
w

t
h
e
a
r
eo
n
l
yp
o
s
i
t
i
v
ec
o
e
f
f
i
c
i
e
n
t
s
.P
h
a
s
e1i
snowc
o
m
p
l
e
t
e
.Th
ea
r
t
i
f
i
c
i
a
lv
a
r
i
a
b
l
e
s
a
n
d t
i
f
i
c
i
a
lc
o
s
tf
u
n
c
t
i
o
nc
a
nb
er
e
m
o
v
e
df
r
o
mt
h
eLPp
r
o
b
l
e
m
.
E
l
i
r
n
i
n
a
t
i
n
gea
r
t
i
f
i
c
i
a
lv
a
r
i
a
b
l
e
sa
n
dt
h
ea
r
t
i
f
i
c
i
a
lc
o
s
tf
u
n
c
t
i
o
nf
r
o
mt
h
ep
r
o
b
l
e
m
P
h
a
s
eI
Ii
ss
t
a
r
t
e
dw
i
t
hT
a
b
l
e3
.
2
7
.N
o
t
t
h
a
tT
a
b
l
e3
.
2
7c
o
n
t
a
i
n
sn
oi
n
f
o
r
m
a
t
i
o
nt
h
a
t
to
n
l
yp
r
o
v
i
d
e
sa
nu
nc
1u
t
t
e
r
e
dt
a
b
l
ef
o
r
i
sn
o
ta
l
r
e
a
d
ya
v
a
i
l
a
b
l
ei
nT
a
b
l
e3
.
2
6
.I
c
o
n
t
i
n
u
i
n
gt
h
a
p
p
l
i
c
a
t
i
o
no
ft
h
eS
i
m
p
l
e
xl
e
c
h
n
i
q
u
e
a
n
dt
h
e
r
e
f
o
r
ei
sn
o
tr
e
a
l
l
y
n
e
c
e
s
s
a
r
y
.I
nT
a
b
l
e3
.
2
7
t
h
eEBVi
sS
3
TheLBVi
sY
la
f
t
e
rat
o
s
sb
e
t
w
e
e
nY
la
n
ds
l
'
Thep
i
v
o
trowi
st
h
es
e
c
o
n
dr
o
w
.T
h
i
srowi
su
s
e
dt
oo
b
t
a
i
nT
a
b
l
e3
.
2
8
.I
nt
h
i
st
a
b
l
e

n
o
t
et
h
a
tt
h
ev
a
l
u
eo
ft
h
eb
a
s
i
cv
a
r
i
a
b
l
eS
Ii
sO
.T
h
i
si
sc
a
l
l
e
dad
e
g
e
n
e
r
a
t
eb
a
s
i
c
eEBVf
o
rt
h
en
e
x
tt
a
b
l
ei
sY
3T
heLBViss
hep
i
v
o
trowi
ses
o
n
d
v
a
r
i
a
b
l
e
.Th
T
3
row(
n
o
t
e
:t
h
ep
i
v
o
tr
o
wc
a
nb
eeroww
i
t
ht
h
ed
e
g
e
n
e
r
a
t
eb
a
s
i
cs
o
l
u
t
i
o
ni
ft
h
e
tr
o
w
)
.T
a
b
l
e3
.
2
9i
sef
i
n
a
lt
a
b
l
e
c
o
l
u
m
no
ft
h
eEBVh
a
sap
o
s
i
t
i
v
ec
o
e
f
f
i
c
i
e
n
ti
na
a
n
dc
o
n
t
a
i
n
st
h
eo
p
t
i
m
a
ls
o
l
u
t
i
o
n
.Thef
i
n
a
ls
o
l
u
t
i
o
no
ft
h
ed
u
a
lp
r
o
b
l
e
mi
s
Y
;=0
=
O
ly
;=1
0

=0
s;=5
w
=280
t
h
es
o
l
u
t
i
o
no
ft
h
ep
r
i
m
a
lp
r
o
b
l
e
mi
s
Tos
u
m
m
a
r
i
z
e
x
;=2

x
i=0
x
;=8

s
;
p=2
4

p=5
z*=280

f
h
es
u
b
s
c
r
i
p
t
p
"i
sa
d
d
e
dt
oa
s
s
o
c
i
a
t
et
h
es
l
a
c
kv
a
r
i
a
b
l
ew
i
t
hep
r
i
m
a
lp
r
o
b
l
e
m
.

T
a
b
l
e3
.
2
3 Exam
p
l
e3
.
5
:D
u
a
lProb
l
e
m

P
h
o
s
e1

T
a
b
l
e1
Y
2

Y
)

Y
2

(
3
.
5
0
)

l
n
t
r
o
d
u
c
i
n
gs
u
r
p
l
u
sv
a
r
i
a
b
l
e
s(
SI>S
2S
3
)a
n
da
r
t
i
f
i
c
i
a
lv
a
r
i
a
b
l
s
(
a
l 2
'a
3
)ed
u
a
l
p
r
o
b
l
e
mi
se
x
p
r
e
s
s
e
da
sas
t
a
n
d
a
r
dLPp
r
o
b
l
e
m
:

Y
1

Y
1

Thes
o
l
u
t
i
o
no
ft
h
ed
u
a
lp
r
o
b
l
e
mi
s

Q2

a)

6
0

3
200
o w
5
o 110
A-

;
Y

=1
0
=1
0

Y
4=0
s
i=5

W
=280
Y
;=0
SomeFormalObservations
T
h
e
s
eo
b
s
e
r
v
a
t
i
o
n
sa
r
ej
u
s
t
i
f
i
e
dm
o
r
cf
o
r
m
a
l
l
yi
nmanyb
o
o
k
so
nl
i
n
e
a
rp
r
o
g
r
a
m
m
i
n
g
someo
fw
h
i
c
hh
a
v
eb
e
e
ni
n
c
1u
d
e
di
nt
h
er
e
f
e
r
e
n
c
e
sa
tt
h
ee
n
do
ft
h
i
sc
h
a
p
t
e
r
.H
e
r
e
t
h
ef
o
r
m
a
lr
e
s
u
l
t
sa
r
ee
l
a
b
o
r
a
l
e
du
s
i
n
gt
h
es
o
l
u
t
i
o
n1
0t
h
ee
x
a
m
p
l
ed
i
s
c
u
s
s
e
di
nt
h
i
s

Table3
.
2
5 Example3.
5
:DualProblemPhase1Table3

Y
2

Y
3

Y
4

-0.
2
8
5
7
0
.
1
7
8
6
0
.
2
1
4
3

1
.
14
2
9

-4
.
8
5
7
1

2
.
4
2
8
6

0
.
2
8
5
7

1
.
14
29

0
.
2
1
4
3
-0
.
1
4
2
9

S
3

$2

1
.
14
29
-0.
2
1
4
3
0
.
1
4
2
9

1
.
14
2
9
02857
0
.
8
5
7
1

7.
42
86

3.
4
2
8
6

1
.
14
29

1
.
14
2
9

a2

a
3

0
.
2
1
4
3
0
.
1
4
2
9

l
.
429
-0.
2
8
5
7
0
.
8
5
7
1

2
.
8
5
7
1
0
.
7
1
4
3
1
2
.
8
5
7

7
.
4
2
8
6

3
.
4
2
8
6

w-291
.43

2
.
1
4
2
9

1429
2.

1
.
14
29

A-2
.
8
5
7

r
ob
l
e

m Phase1Table4
T
a
b
l
e3.
2
6 Example3.
5
:DualP

Y
2

Y
3

Y
4



0
.
2
5
0
.
1
2
5
0
.
2
5

3
.
0
0
0
1

6
E
O

4E
-0
5

4
.
9
9
9
8

$
-0.
8
7
5
-0.
1
8
7
5
0
.
1
2
5
6
.
4
9
9
8

$2

6E-06
-4
E-05

S
3

0
.
5


0
.
0
0
0
2

3
.
9
9
9
8

0
.
8
7
5
0
.
1
8
7
5
0
.
1
2
5
6
.
4
9
9
8

-6E

4E
-0
5
0
.
02

a
3

2
.
4
9
9
9
1
.2
5
1
2
.
5

-0.
5
3
.
9
9
9
8

w-3IO

146

:
1
1

L
lN
EARPROGRAMMING

Y2

Y3

Y4

0
.
2
5
0
.
1
2
5
0
.
2
5

S
I

0
.
8
7
5
0
.
1
8
7
5
0
.
1
2
5
.
5
6

S2

S
3

0
.
5

2
.
5
1
.2
5
1
2
.
5
w-31
O

s
e
c
t
i
o
n
.Nop
r
o
o
f
s '
eg
i
v
e
n
.A
lo
ft
h
eo
b
s
e
r
v
a
t
i
o
n
smayn
o
tb
er
e
l
e
v
a
n
tt
oExample
3.
5d
i
s
c
u
s
s
e
da
b
o
v
e
b
u
ta
r
ei
nc
Iu
d
e
dh
e
r
ef
o
rc
o
m
p
l
e
t
e
n
e
s
s
.
1
. Thep
r
i
m
a
la
n
dd
u
a
lp
r
o
b
l
e
m
sh
a
v
et
h
esam
v
a
l
u
ef
o
rt
h
eo
p
t
i
m
a
lo
b
j
e
c
t
i
v
e
f
u
n
c
t
i
o
n
7
.
"=w"=2
8
0
.Thi
si
st
r
u
e
however
o
n
l
yi
f
t
h
ep
r
o
b
l
e
m
sh
a
v
eo
p
t
i
m
a
l
s
o
l
u
t
i
o
n
s
.
2
.(
fxi
sa
n
yf
e
a
s
i
b
l
es
o
l
u
t
i
o
nt
ot
h
ep
r
i
m
a
lp
r
o
b
l
c
m
.andyi
s yf
e
a
s
i
b
l
es
o
l
u
l
i
o
n
.
(
x
)
.T
h
i
sf
e
a
t
u
r
ep
r
o
v
i
d
e
sa
ne
s
t
i
m
a
t
co
ft
h
e
t
ot
h
cd
u
a
lp
r
o
b
l
c
m
. w(y)~ 7
b
o
u
n
d
so
ft
h
ed
u
a
1o
p
t
i
m
a
lv
a
l
u
ei
faf
c
a
s
i
b
1
ep
r
i
m
a
ls
o
l
u
t
i
o
ni
sknown.T
h
i
s
r
c
s
u
l
ta
l
s
oh
o
l
d
sf
o
rt
h
er
e
v
e
r
s
ec
a
s
ewhenaf
e
a
s
i
b
l
d
u
a
li
sknown.

f
eprima
1problemi
su
n
b
o
u
n
d
e
d
.t
h
ed
u
a
lp
r
o
b
l
e
mi
si
n
f
e
a
s
i
b
l
e
.The
3
.I
unboundedp
roblemi
m
p
l
i
e
s t
h
eo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
nv
a
1uecanbepushedt
o
i
n
f
i
n
i
t
el
i
m
i
t
s
. l
i
sh
a
p
p
e
n
si
ft
h
ef
e
a
s
i
b
l
edomaini
sn
o
tc
1o
s
e
d
.Ofc
o
u
r
s
c
p
r
a
c
t
i
c
a
lc
o
n
s
i
d
e
r
a
t
i
o
n
sw
i
l
l
l
i
m
i
tt
h
co
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
nv
a
l
u
et
oc
o
r
r
c
s
p
o
n
d
i
n
g
I
im
i
t
sont
h
ed
e
s
i
g
nv
a
r
i
a
b
l
e
s
.Thei
n
v
e
r
s
er
e
l
a
t
i
o
n
s
h
i
ph
o
l
d
st
o
o
.I
ft
h
edua
1
p
r
o
b
l
e
mi
sunbounded
t
h
ep
r
i
m
a
li
si
n
f
e
a
s
i
b
l
e
.
t
hp
r
i
m
a
lc
o
n
s
t
r
a
i
n
ti
sa
nequa
1i
t
yc
o
n
s
t
r
a
i
n
t
t
h
ei
t
hd
u
a
lv
a
r
i
a
b
l
ei
s
4
.I
ft
h
ei
ft
h
ep
r
i
m
a
lv
a
r
i
a
b
l
ei
su
n
r
c
s
t
r
i
c
t
e
d
u
n
r
e
s
l
r
i
c
t
e
di
ns
i
g
n
.Ther
c
v
e
r
s
eh
o
l
d
st
o
o
.I
i
ns
i
g
n
.e
nt
h
cd
u
a
lc
o
n
s
t
r
a
i
n
ti
sa
ne
q
u
a
l
i
t
y
.

5
.O
b
t
a
i
n
i
n
gp
r
i
m
a
ls
o
l
u
t
i
o
nf
r
o
md
u
a
ls
o
l
u
t
i
o
n
:
(
i
)I
ft
h
ei
t
hd
u
a
lc
o
n
s
t
r
a
i
n
ti
sas
t
r
i
c
ti
n
e
q
u
a
l
i
t
yt
h
e
nei
p
r
i
m
a
lv
a
r
i
a
b
l
ei
s
n
o
n
b
a
s
i
c(
f
o
roptimums
o
l
u
t
i
o
n
so
n
l
y
)
.Fromt
h
ed
u
a
ls
o
l
u
t
i
o
n
t
h
es
e
c
o
n
d

P
h
o
s
e1
1
T
n
b
l
e2
T
o
b
l
e3
.
2
8 Example3
.
5
:D
u
a
lP
r
o
b
l
e
m
Y
I

2
2
2
8

Y2

Y
.
l

Y4

0
.
5
0
.
2
5
0
.
5

S
I

0
.
5
0
.
3
7
5

0
.
2
5
5

S2

147

5
:Du
a
1P
r
o
b
1
e
m

P
h
o
s
e1
1
T
o
b
1
e3
T
a
b
1
e3
.
2
9 Examp1e3.

T
a
b
1
e3
.
2
7 Exomp1e3.
5
:D
u
o
1P
r
o
b
1
e
m
P
h
a
s
cD
T
a
b
1
c1
Y
I

3
.
5A
D
D
I
T
I
O
N
A
lT
O
P
I
C
SI
NLlNEARPROGRAMMING

5
J



2
.
5
1
5

3
0
0

IV-

Y
I

2
8
2
~i

2
4

Y2

Y
.
l

Y4



o

5

.
f
l

1
.2
5
1
.5

0
.
5
2

S
2

SJ

2
4
-2
8

5
1
0
1
0

.
f-280

c
o
n
s
t
r
a
i
n
t(
3
.
5
1
b
)i
ss
a
t
i
s
f
i
e
da
sa
ni
n
e
q
u
a
l
i
t
y
.T
h
e
r
e
f
o t
h
es
e
c
o
n
dp
r
i
m
a
l
v
a
r
i
a
b
l
ex2i
sn
o
n
b
a
s
i
c di
si
se
v
i
d
e
n
ti
nT
a
b
l
e3.
2
2
.
(
i
i
)I
ft
h
ei
t
hdua
1v
a
r
i
a
b
l
ei
sb
a
s
i
ct
h
e
nt
h
ei
t
hp
r
i
m
a
lc
o
n
s
t
r
a
i
n
ti
sas
t
r
i
c
t
e
q
u
a
l
i
t
y
.FromT
a
b
l
e3
.
2
9
Y2a
n
dY3a
r
eb
a
s
i
cv
a
r
i
a
b
l
e
st
h
e
r
e
f
o
r
e(
3
.
4
7
b
)
d(
3.
47
c
)mustb
ee
q
u
a
l
i
t
i
e
s
.T
h
i
si
si
n
d
e
c
dt
r
u
c
.

Thescr
e
l
a
t
i
o
n
sc
a
na
l
s
ob
cc
s
t
a
b
l
i
s
h
c
dbyv
a
l
u
c
so
fs
l
a
c
kv
a
r
i
a
b
l
e
s
.
6
.R
c
c
o
v
e
r
i
n
gp
r
i
m
a
ls
o
l
u
t
i
o
nf
r
o
mf
i
n
a
ld
u
a
lt
a
b
l
e
: Whenl
h
ep
r
i
m
a
la
n
dd
u
a
l
p
r
o
b
l
e
r
n
sa
r
ei
nt
h
es
t
a
n
d
a
r
df
o
r
mt
h
ev
a
l
u
eo
ft
h
ci
t
hp
r
i
m
n
lv
a
r
i
a
b
l
ee
q
u
a
l
s
t
h
ec
o
e
f
f
i
c
i
e
n
ti
nt
h
el
a
s
trowo
ft
h
ef
i
n
a
lt
a
b
l
c
)
t
h
er
e
d
u
c
e
dc
o
e
f
f
i
c
i
e
n
t(
t
h
a
ti
s
o
ft
h
es
l
a
c
k
l
s
u
r
p
l
u
sv
a
r
i
a
b
l
ea
s
s
o
c
i
a
t
e
dw
i
t
ht
h
ei
t
hd
u
a
lc
o
n
s
t
r
a
i
n
t
.I
nT
a
b
l
e
3
.
2
9t
h
ev
a
l
u
e
so
ft
h
er
e
d
u
c
e
dc
o
s
tc
o
e
f
f
i
c
i
e
n
tc
o
r
r
e
s
p
o
n
d
i
n
gt
ot
h
es
u
r
p
l
u
s
I
S
2
'a
ndS3 a
r
e20and8r
e
s
p
e
c
t
i
v
e
l
y
.Th
ea
r
ep
r
e
c
i
s
e
yt
h
e
v
a
r
i
a
b
l
e
sS
o
p
t
i
m
a
lv
a
l
u
e
so
ft
h
ep
r
i
m
a
lv
a
r
i
a
b
l
e
s
.
7
.l
fei
t
hd
u
a
lv
a
r
i
a
b
l
ei
sn
o
n
b
a
s
i
ct
h
ev
a
l
u
eo
fi
t
sr
e
d
u
c
e
dc
o
s
tc
o
e
f
f
i
c
i
e
n
ti
s
t
h
ev
a
l
u
eo
f s
l
a
c
k
l
s
u
r
p
l
u
sv
a
r
i
a
b
l
eo
ft
h
ec
o
r
r
e
s
p
o
n
d
i
n
gp
r
i
ma
1c
o
n
s
t
r
a
i
n
.
t
I
nTable3
.
2
9
Y
Ia
ndY
4a
r
en
o
n
b
a
s
i
cv
a
r
i
a
b
l
e
sw
i r
e
d
u
c
e
dc
o
s
tc
o
e
f
f
i
c
i
e
n
t
so
f
24a
n
d5r
e
s
p

t
i
v
e
l
y
.Thesea
r
ct
h
ev
a
l
u
e
so
fS
I(
S
l
p
)a
ndS4(S4p)i
nT
a
b
l
e3
.
2
2
.
8
.O
b
t
a
i
n
i
n
gdua
Is
o
l
u
t
i
o
nf
r
o
mp
r
i
m
a
ls
o
l
u
t
i
o
n
: Whent
h
ep
r
i
m
a
landd
u
a
l
p
r
o
b
l
e
r
n
sa
r
ei
nt
h
es
t
a
n
d
a
r
dform

t
h
ev
a
l
u
eo
ft
h
ei
t
hd
u
a
lv
a
r
i
a
b
l
ee
q
u
a
l
st
h
e
r
e
d
u
c
c
dc
o
e
f
f
i
c
i
e
n
t( i
s
t
h
ec
o
e
f
f
i
c
i
e
n
ti
nt
h
el
a
s
trowo
ft
h
ef
i
n
a
lt
a
b
l
e
)o
f
t
h
es
l
a
c
k
l
s
u
r
p
l
u
sv
a
r
i
a
b
l
ea
s
s
o
c
i
a
t
e
dw
i
t
hei
t
hp
r
i
m
a
lc
o
n
s
t
r
a
i
nt
.l
nT
a
b
l
e
3
.
2
2t
h
ev
a
l
u
e
so
ft
h
er
e
d
u
c
e
dc
o
s
tc
o
e
f
f
i
c
i
e
n
tc
o
r
r
e
s
p
o
n
d
i
n
gt
ot
h
es
u
r
p
l
u
s
v
a
r
i
a
b
l
e
sS
I
S
2
'S3
andS4 01
0
0
a
nd0r
e
s
p
e
c
t
i
v
e
l
y
.T
h
e
s
ea
r
ep
r
e
c
i
s
e
l
y
t
h
eo
p
t
i
m
a
lv
a
l
u
e
so
ft
h
ed
u
a
lv
a
r
i
a
b
l
e
si
nT
a
b
l
e3
.
2
9
.
t
h
ev
a
l
u
eo
fi
t
sr
e
d
u
c
e
dc
o
s
tc
o
e
f
f
i
c
i
n
ti
s
9
.I
ft
h
ei
t
hp
r
i
m
a
lv
a
r
i
a
b
l
ei
sn
o
n
b
a
s
i
c
t
h
ev
a
l
u
eo
ft
h
es
l
a
c
k
l
s
u
r
p
l
u
sv
a
r
i
a
b
l
eo
ft
h
ec
o
r
r
e
s
p
o
n
d
i
n
gdua
1c
o
n
s
t
r
a
i
n.
tI
n
T
a
b
l
e3
.
2
2
X2i
san
o
n
b
a
s
i
cv
a
r
i
a
b
l
ew
i
t
hr
e
d
u
c
c
dc
o
s
tc
o
e
f
f
i
c
i
e
n
t
so
f5
.T
h
i
si
s
t
h
ev
a
l
u
eo
fS2i
nT
a
b
l
e3
.
2
9
.S
i
n
c
eXIandX3a
r
eb
a
s
i
cw
i
t
ht
h
er
c
d
u
c
e
dc
o
s
t
c
o
e
f
f
i
c
i
e
n
tofO
e
c
o
r
s
p
o
n
d
i
n
gs
l
a
c
k
l
s
u
r
p
l
u
sS
Ia
ndS3v
a
r
i
a
b
l
e
sw
i
l
lb
z
e
r
o
a
so
b
s
e
r
v
e
di
nT l
e3
.
2
9
.
Thea
b
o
v
el
i
s
tr
e
f
e
r
r
e
dt
ot
h
ep
r
i
m
a
landd
u
a
lp
r
o
b
l
e
m
si
ns
t
a
n
d
a
r
df
o
r
m
.As
i
m
i
l
a
r
l
i
s
tc
a
nb
co
b
t
a
i
n
e
df
o
rn
o
n
s
t
a
n
d
a
r
df
o
r
m
.T
h
e
r
ew
i
l
lb
ea
p
p
r
o
p
r
i
a
t
em
o
d
i
f
i
c
a
t
i
o
n
sf
o
r

1
4
8

35 A
ODlT
I
O
N
A
LT
O
P
I
C
SI
NL
lN
E
A
RP
R
O
G
R
A
M
M
I
N
G 1
4
9

L
lN
E
A
RP
R
O
G
R
A
M
M
I
N
G

n
e
g
a
t
i
v
ev
a
l
u
e
so
fv
a
r
i
a
b
l
e
s w
e
l
la
se
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
t
s
.Th
el
i
s
t
e
dr
e
f
e
r
e
n
c
e
sc
a
n
b
ec
o
n
s
u
l
t
e
df
o
rf
u
r
t
h
e
ri
n
f
o
r
m
a
l
i
o
n
.

Se
n
s
l
t
l
v
l
t
y1
0C
o
s
tC
o
e
f
f
l
c
l
e
n
t

2
0

2
5

.
----6
1
5
0
-
1

3
.
5
.
2 S
e
n
s
l
t
l
v
i
t
yAnalysls
Thes
o
l
u
t
i
o
n1
0t
h
eLPp
r
o
b
l
e
mi
sd
e
p
e
n
d
e
n
lo
nt
h
ev
a
l
u
e
sf
o
rec
o
e
f
f
i
c
i
e
n
t
s
c
b
a
n
dA(
a
l
s
ot
e
r
m
e
d
p
a
r
a
m
e
t
e
r
s
)
i
n
v
o
l
v
e
di
nt
h
ep
r
o
b
l
e
m
.I
nmanyp
r
a
c
t
i
c
a
ls
i
t
u
a
t
i
o
n
s
t
h
e
s
ep
a
r
n
e
l
e
r
sa
r
eo
n
l
yknowna
p
p
r
o
x
i
m
a
t
e
l
ya
n
dmayc
h
a
n
g
ef
r
o
mi
t
sc
u
r
r
e
n
t
v
a
l
u
ef
o
ra
n
ynumbero
fr
e
a
s
o
n
sp
a
r
t
i
c
u
l
a
r
l
ya
f
l
e
ras
o
l
u
l
i
o
nh
a
sb
e
e
ne
s
t
a
b
l
i
s
h
e
d
.
I
n
s
t
e
a
do
fr
e
c
o
m
p
u
t
i
n
ganews
o
l
u
t
i
o
ni
ti
sp
o
s
s
i
b
l
et
oo
b
t
a
i
naf
r
e
s
ho
n
eb
a
s
e
dont
h
e
e
x
i
s
t
i
n
go
n
ea
n
di
t
sf
i
n
a
lS
i
m
p
l
e
xt
a
b
l
e
.T
h
i
si
s
s
u
ei
ss
i
g
n
i
f
i
c
a
n
ti
np
r
a
c
t
i
c
a
lp
r
o
b
l
e
m
s
ws
o
l
u
t
i
o
nw
i
t
h
o
u
t
w
i
t
ht
h
o
u
s
a
n
d
so
fv
a
r
i
a
b
l
e
sa
n
dc
o
n
s
t
r
a
i
n
t
s
.Thea
d
a
p
t
a
t
i
o
no
fan
e
n
s
i
t
i
v
i
.
a
n
a
l
y
s
i
s
.I
nLPp
r
o
b
l
e
m
s
s
e
n
s
i
t
i
v
i
t
y
r
e
s
o
l
v
i
n
gt
h
ep
r
o
b
l
e
mi
sc
a
l
l
e
ds
a
n
a
l
y
s
i
sr
e
f
e
r
st
od
e
t
e
r
m
i
n
i
n
gt
h
er
a
n
g
eo
fp
a
r
a
m
e
t
e
r
sf
o
rw
h
i
c
ht
h
eo
p
t
i
m
a
ls
o
l
u
t
i
o
n
b
a
s
i
s
e
v
e
nt
h
o
u
g
hv
a
l
u
e
sa
lI
h
es
o
l
u
l
i
o
nmayc
h
a
n
g
e
.
s
t
i
l
lh
a
st
h
esamev
a
r
i
a
b
l
e
si
nt
h
E
x
a
m
p
l
e3
.
1i
su
s
e
df
o
ri
sd
i
s
c
u
s
s
i
o
n
.F
i
g
u
r
e
sa
r
e e
df
o
ri
1
Iu
s
t
r
a
t
i
o
nr
a
t
h
e
rt
h
a
n
.F
i
g
u
r
e
sa
r
en
o
tu
s
e
f
u
lf
o
rmoret
h
a
nt
w
ov
a
r
i
a
b
l
e
s
.G
e
n
e
r
a
l
l
y
m
o
s
l
af
o
r
m
a
lp
r
o
of
in
e
a
rp
r
o
g
r
a
m
m
i
n
gp
r
o
b
l
e
m
sa
l
s
op
e
r
f
o
r
ms
e
n
s
i
t
i
v
i
l
y
c
o
m
p
u
l
e
rc
o
d
e
st
h
a
ts
o
l
v
eI
t w n
e
c
e
s
s
a
r
y1
0i
d
e
n
t
i
f
yt
h
e numbero
f
a
n
a
l
y
s
i
s
.R
e
v
i
s
i
t
i
n
g Example 3
.
1i
c
o
m
p
o
n
e
n
tp
l
n
c
e
m
c
n
tm
a
c
h
i
n
e
so
fl
y
p
eA a
n
dB
.Theo
b
j
e
c
t
i
v
ei
st
om
a
x
i
m
i
z
ee
numbero
fb
o
a
r
t
ob
em
a
n
u
f
a
c
t
u
r
e
d
.C
o
n
s
t
r
a
i
n
tgI
r
e
p
r
e
s
e
n
t
st
h
ea
c
q
u
i
s
i
t
i
o
nd
o
l
l
a
r
s
'
a
i
n
tg
2r
e
p
r
e
s
e
n
t
st
h
ef
l
o
o
rs
p
a
c
ec
o
n
s
t
r
a
i
n
t
.C
o
n
s
t
r
a
i
n
lg
3r
e
p
r
e
s
e
n
t
s
a
v
a
i
l
a
b
l
e
.Cons
t
h
enumbero
fo
p
e
r
a
t
o
r
sa
v
a
i
l
a
b
l
e
. Th
ep
r
o
b
l
e
ms
t
a
t
e
m
e
n
ta
n
dt
h
er
e
s
u
l
t
sa
r
e
r
e
p
r
o
d
u
c
e
db
e
l
o
w(
n
o
t
e X
3
'X
4a
n
dX
sa
r
et
h
cs
l
a
c
kv
a
r
i
a
b
l
e
s
)

-16150

3
0

g s
1

1
6
1
5
0

1
5
2
0

=200-optlmumchanged1
0A
c
1
3
0

1
0

1
5

M
a
c
h
l
n
e
so
fT
y
p
eA
g
u
r
o3
.
9

S
e
n
s
i
t
i
v
i
t
ya
n
a
l
y
s
i
s
c
h
a
n
g
i
n
gc
o
s
lc
i
c
i
e
n
.
t

j
(
X
)
:-99O
xI-9 1
X
2-5250

(
3
.
5
)

C
h
a
n
g
i
n
gt
h
cc
o
s
tc
o
e
f
f
i
c
i
e
n
tc
h
a
n
g
e
st
h
es
l
o
p
eo
ft
h
el
i
n
er
e
p
r
e
s
e
n
t
i
n
ge
o
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
na
tmayl
e
a
dt
oac
h
a
n
g
ei
n s
o
l
u
t
i
o
n
.F
o
re
x
a
m
p
l
e
Eq
u
a
t
i
o
n
5
)c
a
nb
ew
r
i
t
t
e
n
(
3.

s
u
b
j
e
c
It
o
: g
l
(
X
)
:O
.
4
x
l
+
-0
.
6
X
2+X
3=
8
.
5

(
3
.
6
)

x
={990/900)XI([5250+j
]/
9
0
0
)
2

Mi
n
i
m
i
z
e

g
2
(
X
)
:3xI-X2+X
4=
2
5

(
3
.
7
)

g
3
(
X
)
:3xI+6x
70
2+X5=

(
3
.
8
)

x1~O; x2ox3~ 0 x4~ 0 Xs~ 0

(
3
.
9
)

Thes
o
l
u
t
i
o
ni
s

;
x=10.4762X
;=6.4285
;
x=0.4524]=-21437.14
ChangingCostC
o
e
f
f
i
c
l
e
n
tValues(
c
)
:F
i
r
s
tc
o
n
s
i
d
e
rt
h
ee
e
c
to
fc
h
a
n
g
i
n
g
t
h
ec
o
s
tc
o
e
f
f
i
c
i
e
n
l
so
ft
h
ed
c
s
i
g
nv
a
r
i
a
b
l
c
s
.I
fC
I
Ot
h
ec
o
e
f
f
i
c
i
c
n
to
fd
e
s
i
g
nv
a
r
i
a
b
l
eX
I
i
sc
h
a
n
g
e
df
r
o
m-990
w
i
l
lt
h
es
e
to
fb
a
s
i
sv
a
r
i
a
b
l
e
sr
c
m
a
i
nt
h
es
a
m
e
?

I
fC
Iw
e
r
et
ob
emade-1900e
nt
h
es
l
o
p
eo
fel
i
n
ew
i
l
lb
e(
19
0
0
/
9
0
0
)
.Th
e
s
e
n
s
i
t
i
v
i
t
yd
u
et
oC
Ic
a
nb
ee
x
p
e
r
i
e
n
c
e
db
yr
u
n
n
i
n
gS
e
n
s
l
t
l
v
l
t
y
_
c
o
s
t
.
m
.1The
c
h
a
n
g
ew
i
l
hr
e
s
p
e
c
t1
0t
h
eo
r
i
g
i
n
a
lo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
ni
sshowni
na
n
i
m
a
t
i
o
n
.The
o
r
i
g
i
n
a
ls
o
l
u
t
i
o
ni
sm
a
r
k
e
do
nt
h
ef
i
g
u
r
eb
yac
i
r
c
l
e
.Theo
r
i
g
i
n
a
ls
o
l
u
l
i
o
ni
su
n
c
h
a
n
g
e
d
i
t
u
d
eo
ft
h
es
l
o
p
ei
sr
a
i
s
e
d
f
o
ri
n
s
t
a
n
c
eC
I=
1
9
0
0
.Ont
h
eo
t
h
e
rh
a
n
d
whenthema
i
fC
Iw
e
r
e1
0b
emade-200

I
h
ed
o
t
t
e
dl
i
n
ei
n
d
i
c
a
t
e
st
h
enewc
o
s
tl
i
n
ea
n
di
nt
h
i
sc
a
s
e
1
lc
h
a
n
g
et
op
o
i
n
tA
.Thi
si
sshowni
nF
i
g
u
r
e3
.
9 i
sE s
u
l
to
f
t
h
es
o
l
u
t
i
o
nwi
i
Ie
.Thef
l
o
o
rc
o
n
s
t
r
a
i
n
ti
sn
ol
o
n
g
e
rb
i
n
d
i
n
g
.I
tc
u
nb
ee
s
t
a
b
l
i
s
h
e
db
y
r
u
n
n
i
n
gt
h
em-f

cu
l
a
t
i
o
n
s i
n
f
i
n
i
t
e
l
ymanys
o
l
u
t
i
o
n
sa
r
ep
o
s
s
i
b
l
ei
fC
Ih
a
sav
a
l
u
eo
f
s
i
m
p
l
ec
a
-4
5
0
.I
nI
h
i
sc
a
s
ct
h
ec
o
s
tf
u
n
c
t
i
o
ni
sp
a
r
a
l
l
e
lt
ot
h
eb
i
n
d
i
n
gc
o
n
s i
n
t8
3
'Fromt
h
i
s
d
i
s
c
u
s
s
i
o
ni
ti
sa
p
p
a
r
e
n
t 1
0k
e
e
pt
h
el
o
c
a
t
i
o
no
ft
h
eo
r
i
g
i
n
a
ls
o
l
u
t
i
o
nu
n
c
h
a
n
g
e
d

CI m
u
s
tb
eg
r
e
a
l
e
rt
h
a
n4
5
0
.S
u
c
ha
n
u
l
y
s
i
si
sp
o
s
s
i
b
l
ef
o
ra
l
lI
h
eo
t
h
e
rc
o
s
l
c
o
e
f
f
i
c
i
e
n
l
si
n
v
o
l
v
e
di
ni
sp
r
o
b
l
e
m
.I
np
a
r
t
i
c
u
l
a
r
t
h
ea
n
a
l
y
s
i
ss
h
o
u
l
di
n
d
i
c
a
t
ef
o
r
l
F
i
l
e
s1
0b
cd
o
w
n
l
o
n
d
e
df
r
o
mt
1
l
ew
e
bs
i
l
cn
r
ci
n
d
i
c
n
l
e
db
yb
o
l
d
f
a
c
e f
t
y
p
e

1
5
0

5
1
R
E
F
E
R
E
N
C
E
S 1

L
1N
E
A
RP
R
O
G
R
A
M
M
I
N
G

w
h
a
tr
a
n
g
eo
fc
o
s
tc
o
e
f
f
i
c
i
e
n
t
st
h
es
o
l
u
t
i
o
nw
o
u
l
ds
t
i
l
lr
e
m
a
i
nw
h
e
r
ei
ti
s
.Th
i
si
s
r
e
f
e
r
r
e
dt
oa
sd
e
t
e
r
m
i
n
i
n
gs
e
n
s
i
t
i
v
i
t
yt
oc
o
s
tc
o
e
f
f
i
c
i
e
n
t
s
.

ChangeI
nt
h
eResourceL
l
mlts(
bV
e
c
t
o
r
)
:Whent
h
ev
a
l
u
S
o
nt
h
cr
i
g
h
t
h
a
n
d
h
ec
o
n
s
t
r
a
i
n
ti
I
n
e
s
l
p
l
a
n
/b
o
u
n
r
i
e
sa
r
emovedp
a
m
l
l
c
lt
ot
h
e
m
s
e
l
v
e
s
.
s
i
d
ec
h
a
n
g
c
st
l
i
sc
h
a
n
g
e
st
h
ef
e
a
s
i
b
l
er
e
g
i
o
n
.T
h
r
e
f
o
r
e
t
h
eo
p
t
i
m
a
lv
a
I
u
e
smaya
I
s
oc
h
g
ed
u
et
o
f
l
s
i
b
l
:
g
i
o
n
.Oncea
g
a
i
nt
h
ec
h
a
n
g
e
sc
a
nbei
l
l
u
s
t
e
db
yr
u
n
n
i
n
g
t
h
ec
h
a
n
g
e
si
ne
Se
ns
v
i
s
.
m
.F
i
g
u
r
e3
.
1
0i
I
Iu
s
t
e
st
h
ec
h
a
n
g
ei
nt
h
er
i
g
h
t
h
a
n
ds
i
d
ev
a
I
u
eo
f
ef
r
r
s
tn
s
t
r
n
t
.b
.
.mi
t
sv
a
I
u
eo
f8
.
5t
o6
.
5
.Th
i
sc
h
a
n
g
c
a
u
s
t
ob
e
c
o
m
ea
n
a
c
t
i
v
ec
o
n
s
a
i
n
.
tI
nt
h
i
sc
a
s
e

t
h
eo
p
t
i
ma
Is
o
l
u
t
i
o
nh
a
smovedf
r
o
mAt
oB
.N
o
t
et
h
a
ti
n
t
h
eo
r
i
g
i
na
Ip
r
o
b
l
e
mg1
w
a
sn
o
ta
c
t
i
v
e
.I
nenewf
o
r
m
u
l
a
t
i
o
ni
ti
s
.However

i
ns
e
n
s
i
t
i
v
i
t
y
a
n
a
l
y
s
i
s
ep
r
o
b
l
c
mi
st
od
i
s
c
o
v
e
rt
h
cr
a
n
g
eo
fb
.s
o t
h
es
o
l
u
t
i
o
ns
t
i
l
lh
a
sesame
v
a
r
i
a
b
l
e
si
neb
a
s
i
s
.l
fblwe
t
or
e
m
a
i
na
b
o
v
e8
.
0
4
7
5
t
h
e
nt
h
es
o
l
u
t
i
o
nw
o
u
l
ds
t
i
l
lb
e
lw
i
l
ln
o
tbea
na
c
t
i
v
ec
o
n
s
a
i
n
t
atA dg
ChangeI
nt
h
eC
o
e
f
f
l
c
l
e
n
tM
a
t
r
l
xA: A c
h
a
n
g
ei
nt
h
ec
o
e
f
f
i
c
i
e
n
tm
a
t
r
i
xi
s
s
i
m
i
l
a
ri
ne
f
f
e
c
tt
ot
h
a
td
u
et
ot
h
ec
h
a
n
g
ei
nt
h
ec
o
s
tc
o
e
f
f
i
c
i
e
n

tw
h
i
l
ed
i
r
e
c
t
l
y

2
0

S
e
n
s
i
t
i
v
i
t
yt
oc
o
n
s
t
r
a
i
n
tc
o
e
l
f
i
c
i
e
n
t
s(
a
1
1
)

2
0
r
lIil~O

":----

5
@

1
5
2
0
m y

Sen
s
l
t
l
v
l
t
yt
or
e
s
o
u
r
c
eI
I
m
l
t
s

1
0

1
5

M
a
c
h
i
n
e
so
fT
y
p
eA
h
a
n
g
i
n
gc
o
s
lc
o
e
f
f
l
c
l
e
n
.
l
F
l
g
u
r
e3
.
1
1S
e
n
s
i
l
l
v
i
l
ya
n
a
l
y
s
i
sc

1
0

i
m
p
a
c
t
i
n
g p
r
o
b
l
c
mb
yc
h
a
n
g
i
n
gt
h
ef
e
a
s
i
b
l
er
e
g
i
o
n
.T
h
e
s
ec
h
a
n
g
e
sa
l
s
od
e
p
e
n
do
n
' an
o
n
b
a
s
i
cv
a
r
i
u
b
l
c
.
w
h
e
t
h
e
rt
h
ev
a
r
i
a
b
l
ei
nt
h
ec
o
l
u
m
ni
sab
n
s
i
cv
a
r
i
n
b
l
c01
S
e
n
s
l
t
l
v
l
t
y
_
c
o
e
f
f
.
mi
l
l
u
s
t
r
a
t
e
st
h
er
e
s
u
l
to
fc
h
a
n
g
i
n
gt
h
ec
o
e
f
f
i
c
i
e
n
ta
l
l
' An
i
n
c
r
e
a
s
ei
nt
h
ev
a
I
u
emovest
h
eo
p
t
i
ma
Is
o
l
u
t
i
o
nt
oBw
h
i
l
ead
e
c
r
e
a
s
el
e
a
v
e
st
h
e
s
o
l
u
t
i
o
nu
n
c
h
a
n
g
e
d
.I
nb
o
t
hs
i
t
u
a
t
i
o
n
s
X
Ia
n
dx2a
r
es
t
i
l
lp
a
r
to
feb
a
s
i
s
.I
nt
h
ef
i
r
s
t
.b
e
c
o
m
e
sab
i
n
d
i
n
gc
o
n
s
t
r
a
i
nt
.T
h
i
si
si
l
l
u
s
t
r
a
t
e
di
nF
i
g
u
r
e3
.
11
.T
hc
c
a
s
eg
c
o
e
f
f
i
c
i
e
n
ti
si
nt
h
ef
l
f
t
c
o
l
u
m
nw
h
i
c
hc
o
r
r
e
s
p
o
n
d
st
oab
a
s
i
cv
a
r
i
a
b
l
e
.C
h
a
n
g
i
n
g
1
y
i
e
l
ds
i
m
i
l
a
r
c
o
e
f
f
i
c
i
e
n
tv
a
l
u
e
si
nt
h
ef
i
r
s
t column i
no
t
h
e
rc
o
n
s
t
r
a
i
n
t
s wi
er
e
a
d
e
ri
se
n
c
o
u
r
a
g
e
dt
ot
r
yo
t
h
e
rc
h
a
n
g
e
si
nt
h
ec
o
e
f
f
i
c
i
e
n
tu
s
i
n
gc
i
n
f
o
r
m
a
t
i
o
n
.Th
m
f
i
l
e
sa
v
a
i
l
a
b
l
ei
nt
h
ec
o
d
es
e
c
t
i
o
no
ft
h
i
sc
h
a
p
t
e
r
.

o -5

1
0

1
5

b
1
:
6
.
5
-o
p
t
i
m
u
mchanged-a
c
t
i
v
ec
o
n
s
t
r
a
i
n
t
0

a
1
'
:
0
.
6-o
p
t
l
m
u
mc
h
a
n
g
e
d
s

1
5

3
0

...

-.
-

r
o

M
a
c
h
l
n
e
s0
1T
y
p
eA
F
1g
u
r
e3
.
1
0S
e
n
s
i
l
l
v
i
t
ya
n
a
l
y
s
i
s
c
h
a
n
g
l
n
gr
i
g
h
t

h
a
n
ds
i
d
e
.

REFERENCES
1
5

1
.D
a
n
t
z
i
gG
.B
.L
i
n
e
a
rProgrammingandEx
t
e
n
s
i
o

P
r
i
n
c
e
t
o
nU
n
i
v
e
r
s
i
t
yP
r
e
s
s
P
r
in
c
e
t
o
n
.NJ.1
9
6
3
.
.G.L
i
n
e
a
randN
o
n
l
i
n
e
a
rProgrammi
l
Jg
A
d
d
i
s
o
n
W
e
s
l
e
y
.R
e
a
d
i
n
g
.
2
.L
u
e
n
b
e
r
g
e
rD
M
A
.
1
9
8
4
.

152 L
lN
EARPROGRAMMING
3
.W
i
n
s
t
o
nW. L
.
l
n
t
r
o
d
u
c
t
i
o
nt
oM
a
t
h
e
m
a
t
i
c
a
lP
r
o
g
r
a
m
m
i
n
g
.A
p
p
l
i
c
a
t
i
o
n
s and
D
u
x
b
u
r
yP
s
s
B
c
l
m
o
n
t

CA1
9
9
5
.
A
I
g
o
r
i
t
h
m
s
4
.A
r
o
r
a
J
.S
.
l
n
t
r
o
d
u
c
t
i
o
nO
p
t
i
m
a
lD
e
s
i
g
l
l
.McGraw-H

I
i NewY
o
r
k
.1
9
8
9
.
5
.Wil
1
ia
ms
G
.
.L
i
n
e
a
rA
I
g
e
b
r
aw
i
t
hA
p
p
l
i
c
a
t
i
o
n
s
.Wm.C
.BrownP
ub
l
i
s
h
e
r
s
D
u
b
u
q
u
e
.
IA
1
9
91
.
6
.N
o
b
1
c
.
B
.
. dD
a
n
i
e
l
J
.W.
A
p
p
l
i
e
dL
i
ne
a
rA
I
g
e
b
r
a
. e
n
t
i

H
a
l
l
En
g
1
e
w
o
o
dC
l
i
f
f
s
.
N
J
.
1
9
7
7
.

PROBLEMS
r
o
v
i
d
e ag
r
a
p
h
i
c
a
ld
e
f
i
n
i
t
i
o
n
l
s
o
l
u
t
i
o
no
ft
h
c
(
F
o
ra
l
lt
w
o
v
a
r
i
a
b
l
ep
r
o
b
l
e
m
sp
p
r
o
b
l
e
m
.
)

In
e
a
rprogrammingp
r
o
b
l
e
m
:
3
.
1 S
o
l
v
et
h
ef
o
l
l
o
w
i
n
gi
Min
S
u
b
:

f
t
.xl
X:
z
}
:XI+X2
3xI-X2~

XI+2x2~5
XI+X2S 4
XI~O;X2~O

3
.
2 S
o
l
v
ct
h
ef
o
l
l
o
w
i
n
gl
i
n
e
a
rprogrammingp
r
o
b
l
e
m
:

Max

f
(
x)ox
z
)
:XI+

S
u
b
:

3xI-xz~ 3

2x
XI+
2S5
XI+X2S 4
XI2
!0
;X22
!0

3
.
3 S
o
l
v
e
t
hf
o
l
l
o
w
i
n
g
l
i
n
c
a
r
p
r
o
g
r
a
m
m
i
n
g
p
r
o
b
l
c
m
:
Min
S
u
b
:

f
(
x"X
2
)
:XI-X2
3
x
I-X2~ 3
XI+
2x
2S5

PROBLEMS 153
S
u
b
:

3xI-X2S3
XI+2x2~5
XI+x2~4
di
XI2
!0;X2u c
ns
i
g
n

3
.
6 Thel
o
c
a
lb
o
o
k
s
t
o
r
emustd
e
t
e
r
m
i
n
ehowmanyo
fe
a
c
ho
ft
h
ef
o
u
rnewb
o
o
k
s
onp
h
o
t
o
n
i
c
si
tmusto
r
d
e
rt
os
a
t
i
s
f
yt
h
encwi
n
t
e
r
e
s
tg
e
n
e
r
a
t
e
di
nt
h
ed
i
s
c
i
p
l
i
n
e
.
wi
I
Ip
r
o
v
i
d
eap
r
o
f
i
to
f$13
and q
u
i
r
e
s2i
n
c
h
e
so
fs
h
e
l
f
Book1c
o
s
t
s$75
s
p
a
c
c
.Book2c
o
s
t
s$85
wi
1
lp
r
o
v
i
d
eap
r
o
f
i
to
f$10
andr
e
q
u
i
r
e
s3i
n
c
h
e
so
f
s
h
e
l
fs
p
a
c
e
.Book3c
o
$
65
w
i
l
lp
r
o
v
i
d
eap
r
o
f
i
to
f$8
a
n
dr
e
q
u
i
r
1i
n
c
h
o
f
s
h
e
l
fs
p
a
c
e
.Book4c
o
s
t
s$100

w
i
l
lp
r
o
v
i
d
eap
r
o
f
i
to
f$15andr
e
q
u
i
r
e
s4i
n
c
h
e
s
o
fs
h
e
l
fs
p
a
c
e
.F
i
n
dt
h
enumbero
fe
a
c
ht
y
p
et
h
a
tmustbeo
r
d
e
r
e
dt
om i
z
e
tT
o
t
a
ls
h
e
l
fs
p
a
c
ei
s1
0
0i
n
c
h
e
s
.T
o
t
a
lamounta
v
a
i
l
a
b
l
ef
o
ro
r
d
c
r
i
n
gi
s
p
r
o
f
i.
$
5
0
0
0
.I
th
a
sb
e
e
nd
e
c
i
d
e
dt
oo
r
d
e
ra
t1
e
a
s
tat
o
o
f
0o
fBook2andBook4
.

el
o
c
a
lcommunityc
o
l
l
e
g
ei
sp
1
a
n
n
i
n
gt
ogrow b
i
o
t
e
c
h
n
o
1
0
g
yo
f
f
e
r
i
n
g
3
.
7 Th
m
b
i
t
i
o
u
sprogrami
sb
c
i
n
gp
1
a
n
n
c
d
t
h
r
o
u
g
hnewf
c
d
e
r
a
lands
t
a
t
eg
r
a
n
t
s
.Ana
f
o
rr
c
c
r
u
i
t
i
n
ga
tl
e
a
s
t2 s
t
u
d
e fromi
nando
u
to
fs
t
a
t
e
.Theya
r
et
or
e
c
r
u
i
t
a
tl
e
a
s
t40o
u
t
o
f
s
t
a
t
cs
t
u
d
e
n
t
s
.Th
e
ywi
1
1a
t
t
e
m
p
tt
o r
u
i
ta
tl
e
a
s
t30s
t
u
d
e
n
t
s
whoa
r
ei
nt
h
et
o
p20%o
ft
h
e
i
rg
r
a
d
u
a
t
i
n
gh
i
g
hs
c
h
o
o
lc
1a
s
s
.C
u
r
r
e
n
tf
i
g
u
S
b
o
u
t8%o
ft
h
ea
p
p
l
i
c
a
n
t
sfromi
ns
t
a
t
e
and6%o
ft
h
ea
p
p
l
i
c
t
s
i
n
d
i
c
a
t
e a
fromo
u
to
fs
t
a
t
eb
e
l
o
n
gt
oi
spooLTheya
l
s
op
l
a
nt
or
e
c
r
u
i
ta
tl
e
a
s
t40s
t
u
d
e
n
t
s
whoh
a
v
eAPcoursesi
nb
i
o
1
0
g
y
.Th
ed
a
t
as
u
g
g
e
s
tt
h
a
t10%and15%o
f
i
n
s
t
a
t
e
ando
u
t
o
f
s
t
a
l
ea
p
p
l
i
c
a
n
t
s
r
e
s
p
e
c
t
i
v
e
1
yb
c
l
o
n
gt
oi
spoo.
lTh
e
ya
n
t
i
c
i
p
a
l
c
t
h
a
tt
h
ea
d
d
i
t
i
o
n
a
lc
o
s
tp
e
rs
t
u
d
e
n
ti
s$800f
o
rc
a
c
hi
n
s
t
a
t
cs
l
u
d
e
n
ta
n
d$1200
f
o
re
a
c
ho
u
t
o
f
s
t
a
t
cs
t
u
d
e
n
tF
i
n
dt
h
c
i
ra
c
t
u
a
le
n
r
o
l
l
m
e
n
tn
e
c
d
e
dt
om
i
n
i
m
i
z
e
c
o
s
tandt
h
e
i
ra
c
t
u
a
lc
o
st
.

3
.
8 F
i
g
u
r
e
3
.
1
2
r
e
p
s
e
n
t
sa
no
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
e
mt
od
e
l
e
r
m
i
n
et
h
ema
.
x
imumt
O
l
a
1
numbero
f sma
I
1
e
rr
e
c
t
a
n
g
1
e
sa
twi
l
f
i
tw
i
t
h
i
nt
h
cl
a
r
g
e
ro
n
e
.The
d
i
m
e
n
s
i
o
n
sa
r
ei
n
d
i
c
a
t
e
donef
i
g
"C.
nenumbero
ft
h
e1
a
r
g
e
rr
c
t
a
n
g
1
e
s
h
o
u
l
db
ea
t1
e
a
s
t5morct
h
a
nt
h
esma

1e
ro
n
c
s
.Thcr
e
c
t
a
n
g
l
e
sc
a
n
n
o
tb
e
r
o
l
a
t
e
d
.Nof
r
a
c
t
i
o
n
a
lr
c
l
a
n
g
l
e
sa
r
ea
l
1owed.

.
xfromMATI.AB
.
)
(
O
p
t
i
o
n
a
l
:c
o
n
f
i
r
ma
I
ls
o
l
u
t
i
o
n
su
s
i
n
gt
h
eO
p
t
i
m
i
z
a
l
i
o
nToolbo

XI+X2S4

!0;X22
!O
XI2

100

3
.
4 S
o
l
v
et
h
ef
o
l
l
o
w
i
n
gi
Inearprogrammingproblem:

M
S
u
b
:

f
(
x"X
2
)
:XI+X2
3
x
I-X22
!3
XI+
2x
2S5

1
.
5

2
.
5

XI+X2S 4
XI2
!0;X22
!0

3
.
5 S
o
l
v
eef
o
l
l
o
w
i
n
gi
Inearprogrammingproblcm:

'A
'
:
_

1
'
1
.
.

.
.

Rguro3
.
1
2P
r
o
b
l
m
3
.
8
.

70

4
.
1 PROBLEMD
E
F
I
N
I
T
I
O
N

155

n
u
m
e
r
i
c
a
lt
e
c
h
n
i
q
u
e
se
s
s
e
n
t
i
a
l1
0t
h
ed
e
v
e
l
o
p
m
e
n
tofNLP[
1
2
)
.Theb
o
o
k
sf
a
m
i
l
i
a
r
t
ot
h
er
e
n
d
e
rs
h
o
u
l
d do a
d
m
i
r
a
b
l
y
. Thi
sc
h
n
p
t
e
ra
l
s
oi
n
t
r
o
d
u
c
e
st
h
es
y
m
b
o
l
i
c
c
o
m
p
u
t
i
o
l
l(computera
l
g
e
b
r
a
)r
e
s
o
u
r
c
ea
v
a
i
l
n
b
l
ei
nMATLABnamelySymbolic
MathToolbox[
3
)
.

4
.
1 PROBLEMDEFINmON

NONLINEAR
PROGRAMMING

O
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
e
m
swhosem
a
t
h
e
m
a
t
i
c
a
lmodelsa
r
ec
h
a
r
a
c
t
e
r
i
z
e
dbyn
o
n
l
i
n
e
a
r
e
q
u
a
t
i
o
n
sa
r
ec
a
l
l
e
dN
o
n
l
i
n
e
a
rP
r
o
g
r
a
m
m
i
n
g LP)p
r
o
b
l
e
m
s
.I
nCh
a
p
t
r3i
tw
n
o
t
e
da
tt
h
e
s
ep
r
o
b
l
e
m
sa
l
s
of
e
l
li
n
t
ot
h
ec
a
t
e
g
o
r
yofm
a
t
h
e
m
a
t
i
c
a
lp
r
o
g
r
a
m
m
i
n
g
r
em
o
s
t
l
yn
o
n
l
i
n
e
a
r
.I
nChap
r2s
e
v
e
r
a
l
p
r
o
b
l
e
m
s
.Engineeringdesignproblemsa
p
r
o
b
l
e
m
swereexaminedg
r
a
p
h
i
c
a
l
l
ynndi
twase
v
i
d
e
n
t
t
h
a
tc
u
r
v
a
t
u
r
ea
n
dt
h
eg
r
a
d
i
e
n
t
oft
h
ef
u
n
c
t
i
o
n
si
n
v
o
l
v
e
dhadas
i
g
n
i
f
i
c
a
n
ti
n
f
l
u
e
n
c
eont
h
es
o
l
u
t
i
o
n
.I
ns
u
b
s
e
q
u
e
n
t
c
h
a
p
t
e
r
swew
i
l
lc
o
n
t
i
n
u
et
oc
e
n
t
e
red
i
s
c
u
s
s
i
o
no
fo
p
t
i
m
a
l
i
t
yc
o
n
d
i
t
i
o
n
sand
n
u
m
e
r
i
c
a
lt
c
c
h
n
i
q
u
e
sa
r
o
u
n
dt
w
o
v
a
r
i
a
b
l
ep
r
o
b
l
e
m
sb
e
c
a
u
s
eei
d
e
a
scana
l
s
obe
e
x
p
r
e
s
s
e
dg
r
a
p
h
i
c
a
l
l
y
.E
x
t
e
n
s
i
o
nt
omoret
h
a
ntwov
a
r
i
a
b
l
e
si
sq
u
i
t
es
t
r
a
i
g
h
t
f
o
r
w
a
r
d
e
n
t
a
t
i
o
ni
smadet
h
r
o
u
g
ht
h
eu
s
eo
fv
e
c
t
o
ra
1g
e
b
r
a
.
andi
smosts
i
m
p
l
ewhent
h
ep s
MA
TI.A
Bw
i
l
lbeu
t
i
l
i
z
e
df
o
ra
l
lg
r
a
p
h
i
cn
e
e
d
s
.
t
h
e
r
ei
s ab
o
t
t
o
m
u
pp
r
e
s
e
n
t
a
t
i
o
n of m
a
t
e
r
i
a
lf
o
rn
o
n
l
i
n
e
a
r
T
r
a
d
i
t
i
o
n
a
l
l
y
n
c
o
n
s
t
r
a
i
n
e
dproblemsa
r
ed
i
s
c
u
s
s
e
df
i
r
s
tf
o
l
l
o
w
db
yc
o
n
s
t
r
a
i
n
e
d
o
p
t
i
m
i
z
a
t
i
o
n
.U
p
r
o
b
l
e
m
s
.Forc
o
n
s
t
r
a
i
n
e
dp
r
o
b
l
e
m
see
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
e
dproblemi
sd
i
s
c
u
s
s
e
d
f
i
r
st
.A s
i
l
a
rp
r
o
g
r
e
s
s
i
o
ni
so
b
s
e
r
v
e
dw
i
t
h g
a
r
dt
ot
h
enumberofv
a
r
i
a
b
l
e
s
.A
f
o
l
l
o
w
e
dbytwov
a
r
i
a
b
l
e
swhichi
sen
s
i
n
g
l
e
v
a
r
i
a
b
l
eproblemi
si
n
t
r
o
d
u
c
e
d
e
x
t
e
n
d
e
dt
oag
e
n
e
r
n
lproblemi
n
v
o
l
v
i
n
g"v
a
r
i
n
b
l
e
s
.T
h
i
so
r
d
e
ra
l
l
o
w
si
n
c
r
e
m
e
n
t
a
l
i
n
t
r
o
d
u
c
t
i
o
nofnewc
o
n
c
e
p
t
s
b
u
tp
r
i
m
a
r
i
l
ya
l
l
o
w
st
h
ec
r
e
a
t
i
v
eu
s
eofe
x
i
s
t
i
n
gr
u
l
e
s
t
oe
s
t
a
b
l
i
s
hs
o
l
u
t
i
o
n
st
ot
h
ee
x
t
e
n
d
e
dp
r
o
b
l
e
r
n
s
.
Ana
n
a
l
y
t
i
c
a
lf
o
u
n
d
a
t
i
o
ni
se
s
s
e
n
t
i
a
lt
ou
n
d
e
r
s
t
a
n
dande
s
t
a
b
l
i
s
ht
h
ec
o
n
d
i
t
i
o
n
sa
t
t
h
eo
p
t
i
m
a
ls
o
l
u
t
i
o
nw
i
l
lhavet
os
a
t
i
s
f
y
. l
i
si
sn
o
tf
o
rt
h
es
a
k
eo
fm
a
t
h
e
m
a
t
i
c
a
l
c
u
r
i
o
s
i
t
yb
u
ti
s e
s
s
e
n
t
i
n
lcomponentoft
h
en
u
m
e
r
i
c
a
lt
e
c
h
n
i
q
u
e
:n
o
t
a
b
l
yt
h
e
s
t
o
p
p
i
n
gc
r
i
t
e
r
i
n
. Th
en
e
c
e
s
s
a
r
ym
n
t
h
e
m
n
t
i
c
a
ld
e
f
i
n
i
t
i
o
n
s and i
l
l
u
s
t
r
a
t
i
o
n
sa
r
e
i
n
t
r
o
d
u
c
e
di
nt
h
i
sc
h
a
p
t
e
r
.R
e
f
e
r
e
n
c
e
sa
r
n
v
a
i
l
n
b
l
ef
o
rr
e
f
r
e
s
h
i
n
gt
h
ec
a
l
c
u
l
u
sandt
h
e
154

I
nNLPi
ti
sn
o
te
s
s
e
n
t
i
a
la
ta
l
lt
h
ef
u
n
c
t
i
o
n
si
n
v
o
l
v
e
dben
o
n
l
i
n
c
a
r
.I
ti
ss
u
f
f
i
c
i
e
n
ti
f
a
r
emanye
x
a
m
p
l
e
si
ne
n
g
i
n
e
e
r
i
n
gi
nwhicho
n
l
y
j
u
s
tonco
fthemi
sn
o
n
l
i
n
e
a
r
.T
h
e
r
eo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
ni
sn
o
n
l
i
n
e
a
rwh

i
et
h
ec
o
n
s
t
r
a
i
n
t
sa
r
el
i
n
e
a
r
.l
fi
nt
h
i
sc
a
s
et
h
e
o
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
ni
saq
u
a
d
r
a
t
i
cf
u
n
c
t
i
o
n
t
h
e
s
ep
r
o
b
l
e
m
sa
r
et
e
r
m
e
dl
i
l
l
e
a
rq
u
a
d
r
a
l
i
c
o
r mostp
r
e
l
yone
x
p
e
r
i
e
n
c
et
o
p
r
o
b
l
e
m
s(LQP).Optimizationproblemsf
i
d
e
n
t
i
f
yt
h
em
a
t
h
e
m
a
t
i
c
a
lmodelc
o
m
p
r
i
s
i
n
go
ft
h
ed
e
s
i
g
nv
a
r
i
a
b
l
e
s
o
b
j
e
c
t
i
v
e
andt
h
e
c
o
n
s
t
r
a
i
n
l
s
. A knowledge o
fe
n
g
i
n
e
e
r
i
n
go
rt
h
ea
p
p
r
o
p
r
i
a
t
ed
i
s
c
i
p
l
i
n
ei
sa
l
s
o
e
s
s
e
n
t
i
a
l
t
oc
s
t
a
b
l
i
s
ham
a
t
l
l
e
m
a
t
i
c
a
lmode
.
lP
r
i
m
a
r
i
l
y
t
h
i
si
n
v
o
l
v
e
sd
e
t
e
r
m
i
n
i
n
gt
h
e
f
u
n
c
t
i
o
n
a
lr
e
l
a
t
i
o
n
s
h
i
p
samongt
h
cd
e
s
i
g
nv
a
r
i
a
b
l
e
s
.Thcr
e
m
a
i
n
i
n
gt
a
s
kt
h
e
ni
s1
0
e
s
t
a
b
l
i
s
ht
h
es
o
l
u
t
i
o
n
.
Howd
o
e
sonee
s
t
a
b
l
i
s
ht
h
es
o
l
u
t
i
o
nt
ot
h
en
o
n
l
i
n
e
a
ro
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
e
m
?
I
nm
a
t
h
e
m
a
t
i
c
s(
n
f
t
e
rn
l
la
tt
h
i
ss
t
a
g
cL
h
e
r
ei
sam
a
t
h
e
m
a
t
i
c
a
lmodel f
o
rt
h
e
p
r
o
b
l
e
m
)t
h
es
o
l
u
t
i
o
l
li
so
b
t
a
i
n
e
dbys
a
t
i
s
f
y
i
n
gt
h
en
e
c
e
s
s
aandSI
{
c
i
e
n
tc
o
n
d
i
t
i
o
n
s
e
c
e
s
s
a
r
yc
o
n
d
i
t
i
o
n
sa
r
t
h
o
s
er
e
l
a
t
i
o
n
st
h
a
ta
r
e
l
a
t
e
dt
ot
h
ec
l
a
s
sofproble~. Then
C d
i
d
a
t
ef
o
rt
h
eoptimums
o
l
u
t
i
o
nm
l
l
s
ts
a
t
i
.
.Ifitd
o
e
s
e
na
n
dt
h
i
si
si
m
p
o
r
t
a
n
t

i
tmaybea
no
p
t
i
m
a
ls
o
l
u
t
i
o
n
.Toq
u
a
l
i
f
yad
e
s
i
g
nv
e
c
t
o
rXP(
Xr
e
p
r
e
s
e
n
t
st
h
ed
e
s
i
g
n
v
e
c
t
o
r
)a
sa
n optimumi
tmusts
n
t
i
s
f
ya
d
d
i
t
i
o
n
n
lr
e
l
a
t
i
o
n
sc
n
l
l
e
dt
h
es
u
j
J
i
c
i
e
n
t
c
o
n
d
i
t
i
o
n
s
.T
h
e
r
e
f
o
r
e optimums
o
l
u
t
i
o
nmusts
a
t
i
s
f
ybo
n
e
c
e
s
s
a
r
ya
n
ds
u
f
f
i
c
i
e
n
t
c
o
n
d
i
t
i
o
n
s
.T
h
i
sc
h
a
p
t
e
re
s
t
a
b
l
i
s
h
e
st
h
e
s
ec
o
n
d
i
t
i
o
n
sf
o
rt
h
eo
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
e
m
.
Example4
.
1i
se
s
t
a
b
l
i
s
h
e
dn
e
x
tandi
su
s
e
di
ns
e
v
e
r
a
lwaysi
nt
h
er
e
m
a
i
n
d
e
ro
fE
c
h
a
p
t
e
r1
0d
e
v
c
l
o
pt
h
ec
o
n
d
i
t
i
o
n
sm
e
n
t
i
o
n
e
da
b
o
v
e
. Once 1hec
o
n
d
i
l
i
o
n
sa
r
e
t
h
en
u
m
e
r
i
c
a
lt
e
c
h
n
i
q
u
e
si
no
p
t
i
m
i
z
a
t
i
o
nwi
Ui
n
c
o
r
p
o
r
a
t
ethemt
oe
s
t
a
b
l
i
s
h
a
v
a
i
l
a
b
l
e
t
h
es
o
l
u
t
i
o
n
.

nExample4
.
1
4
.
1
.
1 ProblemFormulatlo
Theproblemi
s s
t
r
i
c
t
e
dt
otwov
a
r
i
a
b
l
e
st
odrawg
r
a
p
h
i
c
a
ls
u
p
p
o
r
tf
o
rsomeo
fE
d
i
s
c
u
s
s
i
o
n
s
.T
h
e
r
ea
r
etwoc
o
n
s
t
r
a
i
n
t
swhichd
u
r
i
n
gt
h
ed
e
v
e
l
o
p
m
e
n
to
ft
h
i
sc
h
a
p
t
e
r

mays
w
i
t
c
hb
e
t
w
c
c
ne
q
u
a
l
i
t
yandi
n
e
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
t
st
oi

lu
s
t
r
a
t
e l
a
t
e
df
e
a
t
u
r
e
s
.

Problem: Findt
h
cr
e
c
t
a
n
g
l
eo
ft
h
cl
a
r
g
e
s
ta
r
(
i
n p
o
s
i
t
i
v
eq
u
a
d
r
a
n
t
)a
t
c
a
nbe
a
n
s
c
r
i
b
e
dw i
nag
i
v
e
ne
l
l
i
p
s
eands
a
t
i
s
f
yap
r
e
s
c
r
i
b
e
dl
i
n
e
a
rc
o
n
s
t
r
a
i
nt
.

Fromcp
r
o
b
l
e
ms
p
e
c
i
f
i
c
a
t
i
o
nt
h
ee
l
l
i
p
s
ew
i
l
lp
r
o
v
i
d
eani
n
c
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
t
i
n
e
a
rr
e
l
a
t
i
o
namongt
h
cv
a
r
i
a
b
l
c
sw
i
l
lp
r
o
v
i
d
e c
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
tf
o
r
w
h
i
l
eel
t
b
i
sc
x
a
m
p
l
e
.

156

NONL
lNEARPROGRAMMING

4
.
1

M
i
n
i
r
n
i
z
e

PROBLEMDEFINITION

f
{
X
)
l
n

(
4
.
5
)

(
4
.
6
)

S
u
b
j
e
c
tt
o
: [
h
(
X
)
]=
0
2
.
5

x
z

157

[
g
(
X
)
]
m$
;0

(
4
.
7
)

0
w$
X1
;X$
;Xup

(
4
.
8
)

F
o
rt
h
es
p
e
c
i
f
i
cp
r
o
b
l
e
mb
e
i
n
gd
i
s
c
u
s
s
e
d
andrefeningt
oF
i
g
u
r
e4
.
1t
h
ed
e
s
i
g
n
'
d
i
n
a
t
ev
a
l
u
e
sx
l dX2 a
la
l
l
o
wt
h
cc
o
m
p
u
t
a
t
i
o
no
ft
h
e
v
a
r
i
a
b
l
e
sa
r
et
h
ecool
lIa
ra
r
e
a
.Th
eo
p
t
i
r
n
i
z
a
t
i
o
nproblemi
s
r
e
c
t

Min
i
r
n
i
z
e

.
=
4)+(~) 1$;0
. (
x
T
/

S
u
b
j
e
c
t
t
o
: h
l
(
x xz
}
:2
0
x
I+1
5X2- 30 0
81(X X2):

//

/f

Jff

//

0
.
5

//

0
.
5

1
.
5

j{XI X~: X
IX2

O
$
;
x
l
$
;
3
;
2
.
5

X
1

F
l
g
u
r
e4
.
1 C
o
n
s
l
r
a
i
n
l
so
fExa
m
p
l
e4
.
1
.

(
4
.
1
0
)
(
4
.
1
1
)

(
4
.
1
2
)

Thes
i
d
ec
o
n
s
t
r
a
i
n
t
si
n(
4
.
1
2
)c
a
na
l
s
ob
ep
o
s
t
u
l
a
t
e
da
so
n
e
s
i
d
e
dandc
a
nbew
r
i
t
t
e
n
a
s
X1~O;

MathematlcalModel: F
i
g
u
r
e4
.
1c
a
p
t
u
r
c
st
h
ee
s
o
ft
h
ep l
c
m
.Codef
i
l
c
s
c
s
s
a
r
yf
o
rMA
T
I
.
A
Bt
oc
r
e
a
t
et
h
e
Ex4L1.mlandcreate_elllpse.ma
r
ef
i
l
c
sn
h
e
r
ea
r
etwov
a
r
i
a
b
l
e
sXI
andX2l
e
r
ea
r
es
t
a
n
d
a
r
dm
a
t
h
e
m
a
t
i
c
a
le
x
p
r
e
s
s
i
o
n
s
f
i
g
u
r
e
.
'
l
l
i
p
s
e
andt
h
es
t
r
a
i
g
h
tl
i
n
ei
snoproblcma
f
t
e
rC
h
a
p
t
e
r3
.
f
o
ree

O$;x2$;3

(
4
.
9
)

x2~0

(
4.
13
)

4
.
1.
2 DlscussionofConstraints
Using t
h
er
e
l
a
t
i
o
n
s(
4
.
9
)(
4
.
1
2
)s
e
v
e
r
a
la
d
d
i
t
i
o
n
a
lc
1a
s
s
c
so
fp
r
o
b
l
e
m
sc
a
nb
e
e
1a
t
i
o
n
s
.Th
e
ya
r
eexaminedb
e
l
o
w
.
d
e
s
c
r
i
b
e
dbyi
n
c
l
u
d
i
n
go
n
l
yas
u
b
s
e
to
ft
h
cr

StandardFormat:The s
t
a
n
d
a
r
df
o
r
m
a
t oft
h
e NLP i
sr
e
p
r
o
d
u
c
e
dh
e
r
ef
o
r
convemence:

UnconstralnedProblem:Thcrea
r
enof
u
n
c
t
i
o
n
a
lc
o
n
s
t
r
a
i
n
t
sa
l
t
h
o
u
g
ht
h
es
i
d
e

c
o
n
s
t
r
a
i
n
t
sa
r
cn
e
c
e
s
s
a
r
yt
okeept
h
es
o
l
u
t
i
o
nf
i
n
i
t
e
.Fort
h
i
scxample

Mi
n
i
r
n
i
z
e f
(
x
t
tX2'
Xn)

(
4
.1
)

M
i
n
i
r
n
i
z
e

S
u
b
j
e
c
t h
k
(
x h

.
x

)=0
k=1

(
4
.
2
)

)$
(XI
X2

..

Xn
;0
j=
1

2
.
.

(
4.
3
)

x
l$;

(
4.
4
)

$
;x
l
'

Xj

;=1

2
.
.
.
n

I
nv
e
c
t
o
rn
o
t
a
t
i
o
n
F
i
I
1
0b
ed
o
w
n
l
o
a
d
e
df
r
o
mI
h
ew
e
bs
i
l
c ci
n
d
i
c
l
l
l
e
db
yb
o
l
d
f
n
c
e r
i
f
l
y
p
e
.
I

j
{
XI

~): x1~

(
4
.
9
)

I
nt
h
i
sp
r
o
b
l
e
mi
f d
e
s
i
g
nv
a
r
i
a
b
l
c
sa
r
eunboundeda
tcu
p
p
e
rl
i
m
i

tt
h
e
ne
s
o
l
u
t
i
o
nwouldbea
tt
h
el
a
r
g
e
s
tp
o
s
i
t
i
v
ev
a
l
u
o
fX
IandX2At
w
o
s
i
d
e
dl
i
m
i
tf
o
rt
h
e
o
n
s
i
b
i
l
i
t
yo
f
d
c
s
i
g
nv
a
r
i
a
b
l
e
si
su
s
u
a
l
l
yagoodi
d
e
a
.Thed
e
s
i
g
n
e
rd
o
e
sh
a
v
et
h
ep
d
e
f
i
n
i
n
gana
c
c
e
p
t
a
b
l
ed
e
s
i
g
ns
p
a
c
e
.

E
q
u
a
l
i
t
yConstralnedProblem1
: Thef
u
n
c
t
i
o
n
a
lc
o
n
s
t
r
a
i
n
t
si
nt
h
i
sp
r
o
b
l
e
m
o
n
l
ye
q
u
a
l
i
t
i
e
s
.Withr
e
f
e
r
e
n
c
e1
0Example4
.
1cf
o
l
l
o
w
i
n
gp
r
o
b
l
e
mc
a
nbes
c
lup
n
e
q
u
a
l
i
t
yt
oanc
q
u
a
l
i
t
y
)
:
(
a
f
t
e
rc
h
a
n
g
i
n
gci

158

NONL
1N
EAAPAOGAAMMING
Minimize

4
.
2

f
t
.
x
.
X
2
)
:x
.x
2

S
u
b
j
e
c
tt
o
: 1
1
)
(
x
.
.X
2
)
:20x
.+1
5X
2-30 0
(
x
.
h2
X
2
)
:(~/4) +(4) 1=0
O~XI

:
:
;
;
3
;

0~x2::;;3

I
n
t
u
i
t
i
v
e
l
ys
u
c
haproblemmayn
o
tb o
p
t
i
m
i
z
e
ds
i
n
c
et
h
etwoc
o
n
s
t
r
a
i
n
t
sby
t
h
e
m
s
e
l
v
e
ss
h
o
u
l
de
s
t
a
b
l
i
s
ht
h
ev
a
l
u
e
sf
o
rt
h
eIwod
e
s
i
g
nv
a
r
i
a
b
l
e
s
.Th
ea
r
g
u
m
e
n
l
s
u
s
e
di
nLPf
o
ra
c
c
e
p
l
a
b
l
ep
r
o
b
l
e
md
e
f
i
n
i
l
i
o
na
r
ea
l
s
ov
a
l
i
dh
.T
h
e
r
ei
sa
l
w
a
y
st
h
e
p
o
s
s
i
b
i
l
i
t
yo
fm
u
l
t
i
p
l
es
o
l
u
t
i
o
n
s
w
h
i
c
hi
sas
t
r
o
n
gf
e
a
l
u
r
eo
fn
o
n
l
i
n
a
rp
r
o
b
l
e
m
s
.I
n
s
u
c
ha
ne
v
e
n
tt
h
es
e
to
fv
a
r
i
a
b
l
e
st
h
a
ty
i
e
l
delow
s
lv
a
l
u
ef
o
rt
h
eo
b
j
e
c
t
i
v
ew
i
l
lb
e
t
h
eo
p
t
i
m
a
ls
o
l
u
t
i
o
n
.NOle s
u
c
has
o
l
u
t
i
o
ni
so
b
t
a
i
n
e
dbys
c
a
n
n
i
n
gt
h
ea
c
c
e
p
t
a
b
l
e
s
o
l
u
t
i
o
n
sr
a
t
h
e
r a
nt
h
r
o
u
g
ha
p
p
l
i
c
a
t
i
o
no
fanyr
i
g
o
r
o
u
sc
o
n
d
i
t
i
o
n
s
.

E
q
u
a
l
l
t
yConstralnedProblem2
:I
f
t
h
ep
r
o
b
l
e
mwe

1
0i
nc
Iudeo
n
l
yoneo
fl
h
e

M
A
T
H
E
M
A
T
I
C
A
L
C
O
N
C
E
P
T
S

159

s
o
m
e
c
o
n
s
t
r
a
i
n
t
s
a
r
e
a
c
t
i
v
e(
e
q
u
a
l
i
t
y
)a
tt
h
eo
p
t
i
m
a
ls
o
l
u
t
i
o
na
so
t
h
e
r
w
i
s
ei
twould
bea
nu
n
c
o
n
s
t
r
a
i
n
e
ds
o
l
u
t
i
o
n
.

4
.
2 MATHEMATICALCONCEPTS
L
i
k
eLP
somemathematicald
e
f
i
n
i
t
i
o
n
sa
r
en
s
a
r
yb
e
f
o
r
et
h
en
s
a
r
ya
n
d
s
u
f
f
i
c
i
e
n
tc
o
n
d
i
t
i
o
n
sf
o
rt
h
eNLPc
a
nbee
s
t
a
b
l
i
s
h
e
d
.D
e
f
i
n
i
t
i
o
n
sa
r
en
e
e
d
e
df
o
rb
o
t
h
ea
n
a
l
y
t
i
c
a
ld
i
s
c
u
s
s
i
o
na
sw
e
l
la
sn
u
m
e
r
i
c
a
lt
e
c
h
n
i
q
u
e
s
. MATLAB p
r
o
v
i
d
e
sa
S
y
m
b
o
l
i
cMathToolbox whichp
e
r
m
i
t
ss
y
m
b
o
l
i
cc
o
m
p
u
t
a
t
i
o
ni
n
t
e
g
r
a
l
e
di
nc
n
u
m
e
r
i
c
a
le
n
v
i
r
o
n
m
e
n
to
fMATLAB.Thi
sa
l
l
o
w
sI
h
eu
s
e
r10e
x
p
l
o
p
r
o
b
l
e
m
si
n
c
a
l
c
u
l
u
sl
i
n
e
a
ra
l
g
e
b
r
as
o
l
u
t
i
o
n
so
fs
y
s
t
e
mo
fe
q
u
a
t
i
o
n
sando
t
h
e
ra
r
e
a
s
.I
nf
a
c
t

u
s
i
n
gs
y
m
b
o
l
i
cc
o
m
p
u
t
a
t
i
o
ns
t
u
d
e
n
t
sc
a
ne
a
s
i
l
yr
e
c
o
v
e
rep
r
q
u
i
s
i
t
ei
n
f
o
r
m
a
t
i
o
n

1'he
f
o
rt
h
ec
o
u
r
s
e
.As
h
o
r
th
a
n
d
s
o
ne
x
e
r
c
i
s
et
os
y
m
b
o
l
i
cc
o
m
p
u
t
a
t
i
o
ni
sp
r
o
v
i
d
e
d
.
p
r
i
m
a
r
yd
e
f
i
n
i
t
i
o
n
swen
e
e
da
r
ed
e
r
i
v
a
t
i
v
e
s
p
a
r
t
i
a
ld
e
r
i
v
a
t
i
v
e
s
m
a
t
r
i
c
e
s
d
e
r
i
v
a
t
i
v
e
s
a
n
ds
o
l
u
t
i
o
n
so
fn
o
n
l
i
n
e
a
re
q
u
a
t
i
o
n
s
.
o
fm
a
t
r
i
c
e
s

f
o
rexample
c
o
n
s
t
r
a
i
n
t
s
Minimize

f
t
.
x
l
X
2
)
:X
IX
2

S
u
b
j
e
c
l
l
o
: h
(
x
"
X
2
)
:(~/4) +(~) )=0
2

0
:
:
;
;X
I:
;3
;

0
:
:
;
;X
2
:
:
;
;3

T
h
i
si
sav
a
l
i
do
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
e
m
.As
i
m
i
l
a
rp
r
o
b
l
e
mc
a
nb
ed
e
f
i
n
e
dw
i
l
ht
h
ef
i
r
s
t
c
o
n
s
t
r
a
i
n
tbyi
t
s
e
l
f
.

I
n
e
q
u
a
l
l
t
yConstrainedProblem: I
nt
h
i
sc
a
s
et
h
ec
o
n
s
t
r
a
i
n
t
sa
r
ca
l
li
n
e
q
u
a
l
i
t
i
e
s
.
Av
a
r
i
a
t
i
o
nonE
x
a
r
n
p
l
e4
.
1wouldbe(
t
h
ee
q
u
a
l
i
t
yc
o
n
s
a
i
n
ti
st
r
a
n
s
f
o
r
m
e
dt
oa
n
e
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
t
)
t
x
.
.x
2
l
:-XIXz
Minimize f

S
u
b
j
c
c
tt
o
: g.(x"X
2
)
:20X
I+1
5X
2-3
0
:
:
;
;0
g
2
(
X
"
X
Z
)
:(
x
t
/
4
)+(4) 1
:
:
;
;
0

O
:
:
;
;
X
I:
:
;
;
3
;

0
:
:
;
;
x
2
:
:
;
;
3

L
ikei
t
sc
o
u
n
t
e
ti
nl
i
n
e
a
rprogramming i
si
sav
a
l
i
do
p
l
i
m
i
z
a
t
i
o
np
r
o
b
l
e
m
.

4
.
2
.
1 SymbollcComputatlonUslngMATLA
Oneo
ft
h
eb
e
s
lwayst
og
e
lfami
1i
a
rw
i
t
hs
y
m
b
o
l
i
cc
o
m
p
u
t
a
t
i
o
ni
st
ot
a
k
et
h
eq
u
i
c
k
h
eS
y
m
b
o
l
i
cMathToolboxa
v
a
i
l
a
b
l
ei
nt
h
eMATLABDemos
o
n
l
i
n
ei
n
t
r
o
d
u
c
t
i
o
n10t
d
i
a
l
o
gbox[
4
]
.Th
ef
o
l
l
o
w
i
n
gs
e
q
u
e
n
c
el
o
c
a
t
e
st
h
el
u
t
o
r
i
a
l
:
>> demos

Symbolic Math - > Introduction


y
m
b
o
l
i
co
p
e
r
a
t
i
o
n
si
nMAT
1
.
ABi
st
h
c
Thec
o
m
p
u
t
a
l
i
o
n
a
le
n
g
i
n
ee
x
e
c
u
l
i
n
g s
db
yW
a
t
e
r
l
o
oMaple
I
n
c
.l
fE a
d
e
ri
s
kemelo
fMaplem
a
r
k
e
l
e
da
n
dsuppo
a
l
r
e
a
d
yf
a
m
i
l
i
a
rw
i
t
hMaple
e
nMA
T
1
.
A
Bp
r
o
v
i
d
e
sahookt
h
r
o
u
g
hwhichMaple
commandsc
a
nb
ee
x
u
t
e
di
nMATLAB.Th
es
y
m
b
o
l
i
cc
o
m
p
u
t
a
t
i
o
ni
nMATLABi
s
p
e
r
f
o
r
m
e
du
s
i
n
gas
y
m
b
o
l
i
co
b
j
e
c
to
rs
y
m
.T
h
i
si
sa
n
o
t
h
e
rd
a
t
at
y
p
ei
Iket
h
enumber
ands
t
r
i
n
gd
a
t
at
y
p
e
su
s
e
di
n r
l
i
e
re
x
e
r
c
i
s
e
s
.Th
eS
y
m
b
o
l
i
cMathToolboxu
s
e
ss
ym
o
b
j
e
c
t
st
o p
r
e
s
e
n
ts
y
m
b
o
l
i
cv
a
r
i
a
b
l
e
s
e
x
p
r
e
s
s
i
o
n
s
andm
a
t
r
i
c
e
s
.
I
nt
h
ee
x
e
r
c
i
s
et
h
a
tf
o
l
l
o
w
s
af
u
n
c
t
i
o
no
fonev
a
r
i
a
b
l
e
andtwof
u
n
c
t
i
o
n
so
fIwo
1
)a
t
1
u
s
e
df
o
ri
I
Iu
s
t
r
a
l
i
o
n
.Drawingont
h
e
v
a
r
i
a
b
l
e
s(
c
o
n
s
t
r
a
i
n
t
sfromExample4.
a
u
o
r
'sc
1assroomex
n
c
ei
sp
r
e
l
i
m
i
n
a
r
yd
i
s
c
u
s
s
i
o
ni
si
nt
e
r
m
so
fv
a
r
i
a
b
l
e
sX
andyf
o
ri
m
p
r
o
v
e
dc
o
m
p
r
e
h
e
n
s
i
o
n
.I
nl
a
t
e
rs
e
c
t
i
o
n
s
s
u
b
s
c
r
i
p
t
sonxa
r
eu
s
e
dt
od
e
f
i
n
e
m
u
l
t
i
p
l
ev
a
r
i
a
b
l
e
ss
ot
h
a
tet
r
a
n
s
i
t
i
o
nt
ot
h
eg
e
n
e
r
a
lproblemc
a
nb
ef
a
c
i
l
i
t
a
t
e
du
s
i
n
g
v
e
c
t
o
rd
e
s
c
r
i
p
t
i
o
n
.Thef
u
n
c
t
i
o
n
si
nt
h
c
s
ce
x
e
a
i
s
e
sa
r
e

E
q
u
a
l
l
yv
a
l
i
dwouldb
eaproblemt
h
a
ti
n
c
Iudedj
u
s
toneo
fl
h
ec
o
n
s
t
r
a
i
n
t
so
ra
n
y
numberofin
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
t
s
.
I
ti
se
s
s
e
n
l
i
a
lt
ou
n
d
e
r
s
t
a
n
db
o
t
ht
h
en
a
t
u
r
eandt
h
enumbero
fc
o
n
s
t
r
a
i
n
t
sa
s
w
e
l
la
showt
h
e
ya
f
f
e
c
tt
h
ep
r
o
b
l
e
m
.I
ng
e
n
e
r
a
l
e
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
t
sa
r
ee
a
s
y10
h
a
n
d
l
e
m
a
t
h
e
m
a
t
I
C
a
l
l
y
d
i
f
f
i
c
u
l
t
t
o
s
a
t
i
s
f
yn
u
m
e
r
i
c
a
l
l
y
.
a
n
d
m
o
r
e
r
e
s
t
r
i
c
t
i
v
e
o
n
t
h
e
s
e
a
r
c
h f
o
rt
h
es
o
l
u
t
i
o
n
. I
n
e
q
u
a
l
i
t
y c
o
n
s
t
r
a
i
n
t
s a
r
ed
i
f
f
i
c
u
l
tt
o r
e
s
o
l
v
e
m
a
t
h
e
m
a
t
i
c
a
l
l
yanda
r
emoref
l
e
x
i
b
l
ew
i
t
hr
e
s
p
e
c
tt
ot
h
es
e
a
r
c
hf
o
rt
h
eo
p
t
i
m
a
l
s
o
l
u
t
i
o
na
st
h
e
yd
e
f
i
n
eal
a
r
g
e
rf
e
a
s
i
b
l
er
e
g
i
o
n
.Aw
e
l
l
p
o
s
e
dproblemrequir~s t
h
a
t

ftx)=12+(x-I)2(x-2)(x-3)
g
.
(
x

y
)
:20x+1
5y-30

l-1

g
2
(
X

y
)
:x
2
/
4+

I
T
O
Or
c
l
c
a
s
co
f1
0
0n
e
wSwdentEdillolli
n
c
l
u
d
c
sI
h
cs
y
m
b
o
l
i
cl
o
o
l
b
O
l
u
l
o
n
gw
i
l
hI
h
eb
a
s
i
cM
A
T
L
A
D
sortw t

1
6
0

4
.
2 MATHEMATICAlCONCEPTS

NONL
lNEARPROGRAMMING

l
j
l

"
J
'

syms y f g1 g2 g definitionof multiple objects


whos % types of variables in the workspace
Bytes
Class
Name
Size
126
sym object
1X1
f
126
sym object
g
1x1
128
sym object
1X1
gl
128
sym obje'ct
1x1
92
1X1
126
sym object
x
y
1X1
126
snobject

9'
" [gl;g2) 9 column vector based on gl
g2
9
[20*x+15*y-30]
1/4*x+y-1]

9 can be the constraint vector in optimization


problems
thepartial derivatives of 9 with respect to design
v
ariables is called the Jacobian matrix
theproperties of this matrix are important for
n
umerical techniques

xy . [
x y]; rowvector of variables
xy) calculating the Jacobian
J'
" jacobian(g

Grand total is 14 elements using 760 bytes


x1)*(x-1)*(x2)*(x3
)
f
.
. 12 + (

definingx as a single symbolic object

S
E
.

sym('x')

"

x
x

'E=

fS5
1
nl
da

l
ef
o
l
l
o
w
i
n
gMATLABs
e
s
s
i
o
nwasc
a
p

r
e
da
sad
i
a
r
yf
i
l
ea
n
de
d
i
t
e
di
nat
e
x
t
f
o
eMATLAB p
r
o
m
p
td
o
e
sn
o
ta
p
p
e
.l
eb
o
l
d
f
a
c
ew
o
r
d
sa
r
e
e
d
i
t
o
r
.The
commandsa
ter
e
a
d
e
rw
i
l
lt
y
p
ea
tt
h
ecommandl
i
n
e
.

constructing f
J

20 1
5
]
[1/4 1]

12+(x-1)"2*(x2)*(x-3)
diff(}

first derivative

aos

ezp10t(f)

2*(x-1)*(x-2)*(x-3)+(x-1)^2*(x-3)+(x-1)"2*(x-2)

0
4))
ezp10t(f[

note the chain rule for derivatives


% note the independent variable is assumed to be x

x2
)
diff(
the second d
erivative wrt x
ans
2*(x-2)*(x3)+4*(x-1)*(x-3)+4*(x-1)*(x2)+2*(x-1)"'2
}
a

'nHE

X2

aL=

(x
fg*

fn4
1a2

the third derivative wrt x

qd

nu

-wnuw

--

Ed

ATY

X5
*
0+

=*

2x

t-14

Enu

gg

.ZO:~

define 9

g2 can only have partial derivatives


gl
independent variables have to be identified
}

t-

aEMnu

g
'8t=

id an 2

partial derivative

a plot of f for -2 pi x 2 pi (default)

plot between 0 <= x <= 4

df diff(f);
h01d 00
0
41) plotting function and derivative
ezp10t(df[
c
ombine with MATLAB graphics - draw a line
1ioe([0 4
][
00
)
'C010r'
r
')

3
5

g2 '
" 0.2S*x + Y 1
; define g2

1
6
1

9
[20*x+15*y3
0
]
1/4*x+y-1]
to e
valuate 9 at x
1Y
subs(g
{x
y}
{1
2.5})
ans
27.5000
1.7500

2.5

A
d
d
i
t
i
o
n
a
lsymbo
i
1
cc
o
m
p
u
t
a
t
i
o
n
sw
i
l
lbei
n
t
r
o
d
u
c
e
dt
h
r
o
u
g
hc
o
d
ea
sa
p
p
r
o
p
r
i
a
t
e
.
e
s
u
l
to
fb
o
t
hn
u
m
e
r
i
ca
n
ds
y
m
b
o
l
i
cc
o
m
p
u
t
a
t
i
o
n
sc
a
nb
ee
a
s
i
l
y
N
o
t
et
h
a
ter
c
o
m
b
i
n
e
da
10ngw
i
t
hg
r
a
p
h
i
c
st
op
r
o
v
i
d
eap
o
w
e
r
f
u
lc
o
m
p
u
t
i
n
ge
n
v
i
r
o
n
m
e
n
.
t

162

N
O
NL
lN
E
A
AP
A
O
G
A
A
M
M
I
N
G

4
.
2

4
.
2
.
2 B
asicMathematicalConcepts

M
A
T
H
E
M
A
T
I
C
A
LC
O
N
C
E
P
T
S 163

T
a
b
l
e4.
1 C
a
l
c
u
l
a
t
i
o
no
rDer
i
v
a
U
v
e

Theb
a
s
i
cm
a
t
h
e
m
a
t
i
c
a
le
1
e
m
e
n
t
si
nt
h
ed
i
s
c
u
s
s
i
o
no
fNLPa
r
ed
e
r
i
v
a
t
i
v
e
sp
a
r
t
i
d
e
r
i
v
a
t
i
v
e
sv
e
c
t
o
r
sm
a
t
r
i
c
e
sJ
a
c
o
b
i
a
na
n
dH
e
s
s
i
a
n
.Weh
a
v
cu
s
e
dt
h
eS
y
m
b
o
l
i
c
MathT
o
o
l
b
o
xi
nt
h
ep
r
e
v
i
o
u
ss
c
c
t
i
o
nt
oc
a
l
c
u
l
a
t
esomeo
ft
h
e
s
eq
u
a
n
t
i
t
i
e
sw
i
t
h
o
u
t
d
c
f
i
n
i
n
gt
h
e
m
.Th
e
s
et
o
p
i
c
sw
i
l
lh
a
v
eb
e
e
ne
x
t
e
n
s
i
v
e
l
yc
o
v
e
r
e
di
nt
h
ef
o
u
n
d
a
t
i
o
n
c
o
u
r
s
e
so
nm
a
t
h
e
m
a
t
i
c
si
nm
o
s
td
i
s
c
i
p
l
i
n
e
s
.Ab
r
i
e
fr
e
v
i
e
wi
so
f
f
e
r
e
di
ni
ss
e
c
t
i
o
n
u
s
i
n
gM
A
T
L
A
B
.l
i
so
p
p
o
r
t
u
n
i
t
yw
i
l
la
l
s
ob
eu
t
i
l
i
z
e
dt
oi
n
c
r
e
a
s
ef
a
m
i
l
i
a
r
i
t
yw
i t
h
e
i
tw
i
t
heMATLABcommandst
h
a
tw
e
r
e
S
y
m
b
o
l
i
cMathT
o
o
l
b
o
xa
n
di
n
c
o
r
p
o
r
a
t
a
r
l
i
e
rc
h
a
p
t
e
r
.
u
s
e
di
n e
Functlono'OneV
a
r
i
a
b
l
e
:f
t
.
x
)i
d
e
n
t
i
f
i
e
saf
u
n
c
t
i
o
no
fo
n
ev
a
r
i
a
b
l
e
:

dx

3
3
3
3

0
.
1
0
.
0
1
0
.
0
0
1

x+dx

d
f
l
d
x

4
3
.
1
3
.
0
1
3
.
0
0
1

1
8
0.
48
5
1
0
.
0
4
0
8
O
.4

D
e
r
i
v
a
t
i
v
e

1
8
4
.
8
5
1
4
.
0
8
0
5
4
.
0
0
8

4
4
4
4

f
(
x
) 1
2+(x-li(x-2
)(x-3)

o
ft
h
ec
h
a
n
g
ei
nt
h
ef
u
n
c
t
i
o
nv
a
l
u
e
4
f
10ec
h
a
n
g
ei
n d
i
s
p
l
a
c
e
m
e
n
l
a
x
.F
o
rE
e
x
a
m
p
J
e
t
h
ed
e
r
i
v
a
l
i
v
ea
tx=3w
i
t
hax=0
.
0
0
1i
so
b
t
a
i
n
e
da
s

i
su
s
e
da
sas
p
e
c
i
f
i
ce
x
a
m
p
l
eo
fs
u
c
haf
u
n
c
t
i
o
n
.Thed
e
r
i
v
a
l
i
v
eo
fef
u
n
c
t
i
o
na
te
sw
r
i
t
t
e
na
s
l
o
c
a
t
i
o
nxi

E
lI_
f
.
:
(
3+0
.
0
0
1
)1
(
3
2
-.11nnSl

(
4
.
1
4
)

- 1 kx+ax).
f
(
x
)
dx-ilm
Ax
=
l
l
f
I
1
L
ax
o
ax
a
x
>
o
a
x
xi
st
h
ep
o
i
n
ta
b
o
u
tw
h
i
c
h d
e
r
i
v
a
t
i
v
ei
scompud.ri
st
h
cd
i
s
t
a
m

et
ua
n
e
i
g
h
b
o
r
i
n
gp
o
i
n
twhos
l
o
c
a
t
i
o
nt
h
e
r
e
f
o
r
ew
i
l
lbex+a
x
.Th
ev
a
l
u
eo
ft
h
ed
e
r
iv
a
t
iv
e
i
so
b
t
a
i
n
e
da
sal
i
m
i
to
ft
h
cr
a
t
i
oo
fed
i
f
f
e
r
e
n
i
nt
h
ev
a
l
u
eo
ft
h
ef
u
n
c
t
i
o
na
te
t
w
op
o
i
n
t
s( ed
i
s
s
e
p
a
r
a
t
i
n
gt
h
et
w
op
o
i
n
t
s() t
h
i
ss
e
p
t
i
o
ni
s
ec
o
m
p
u
t
a
t
i
o
no
ft
h
cd
e
r
i
v
a
t
i
v
ef
o
rt
h
es
p
e
c
i
f
i
ce
x
n
p
l
ei
su
s
u
a
l
l
y
r
e
d
u
c
e
d

z
e
r
o
.Th
du
s
i
n
gt
h
ep
r
o
d
u
c
tr
u
l
e
.R
e
s
u
l
t
sf
r
o
mt
h
ee
x
e i
s
eo
ns
y
m
b
o
l
i
cc
o
m
p
u
t
a
t
i
o
n
o
b
t
a
i
n
p
r
o
v
i
d
et
h
cr
e
s
u
J
t

d
x
l.rc3

0
.
0
0
1

(
4
.
1
5
)

Thed
e
r
i
v
a
t
i
v
cf
o
rl
h
es
i
n
g
J
e
v
a
r
i
a
b
l
ef
u
n
c
t
i
o
na
la
n
yv
a
l
u
cxi
sa
J
s
oc
aJ
le
dl
h
es
l
o
p
e
o
rt
h
eg
r
a
d
i
e
t
l
to
fl
h
cf
u
n
c
t
i
o
n a
lp
o
i
n.
tI
fal
i
n
ei
sd
r
a
w
nt
a
n
g
e
n
l10t
h
ef
u
n
c
t
i
o
n
a
lt
h
ev
a
l
u
ext
h
et
a
n
g
e
n
to
fE g
l
et
h
a
ti
sJ
i
n
em
a
k
e
sw
i t
h
ex
a
x
i
sw
i
l
lh
a
v
e
n
ev
a
l
u
ca
st
h
ed
e
r
i
v
a
t
i
v
e
.I
fi
sa
n
g
l
ei
s9e
n
t
h
es

=tan9
F
i
g
u
r
e4
.
2i
l
l
u
s
t
r
a
t
e
set
a
n
g
e
n
c
yp
r
o
p
e
r
t
yo
ft
h
ed
e
r
i
v
a
t
i
v
e
.

= 3)+(x-1
) ) +

1
)^ 2
)

T
a
b
l
e4
.
1i
l
l
u
s
t
r
a
t
e
sel
i
m
i
t
i
n
gp
r
o
c
e
s
sf
o
rt
h
cd
e
r
i
v
a
t
i
v
eo
ft
h
ef
u
n
c
t
i
o
nc
h
o
s
e
na
s
ec
x
a
m
p
J
e
.Thed
e
r
i
v
a
t
i
v
ei
sb
e
i
n
ge
v
a
l
u
a
t
e
da
tp
o
i
n
t3
.Ast
h
ed
i
s
p
J
a
c
e
m
e
n
t
sa
r
e
r
e
d
u
c
e
d v
a
l
u
eo
ft
h
er
a
t
i
oa
p
p
r
o
a
c
h
c
s v
a
l
u
eo
f4w
h
i
c
hi
st
h
ea
C
l
u
a
lv
a
l
u
eo
f
t
h
ed
e
r
i
v
a
t
i
v
e
.Wh
e
naxi
sl
a
r
g
e(
w
i
t
hav
a
l
u
eo
f1
)cr
a
t
i
oi
s1
8
.T
h
i
si
ss
i
g
n
i
f
i
c
a
n
t
l
y
d
i
f
f
e
r
e
n
tf
r
o
mt
h
ev
a
l
u
eo
f4
.W
i
t
haxo
f0
.
1t
h
ev
a
l
u
ei
sa
r
o
u
n
d4.851
.W
i
t
hf
u
r
e
r
r
e
d
u
c
t
i
o
nt
oO
. 1t
h
er
a
t
i
oh
a
sav
a
l
u
eo
f4
.
0
0
8
.F
romt
h
e
s
ec
o
m
p
u
t
a
l
i
o
n
si
tC
a
nb
c
e
x
p
e
c
t
e
d a
saxa
p
p
r
o
a
c
h
e
s0t
h
er
a
t
i
ow
i
l
lr
e
a
c
ht
h
ee
x
a
c
tv
a
l
u
eo
f4
.
0
.

MATLABC
ode:F
i
g
u
r
e4
.
2w
a
sc
r
e
a
l
e
db
yt
h
ec
o
d
et
a
n
g
e
n
t
.
m
.I
nt
h
es
a
m
ed
i
r
e
c
t
o
r
y

ec
o
d
ed
e
r
l
v
a
t
i
v
e
.
mp
r
o
v
i
d
e
st
h
ei
l
l
u
s
t
r
a
t
i
o
no
ft
a
n
g
e
n
c
yt
h
r
o
u
g
haf
i

r
e
a
n
i
m
a
t
i
o
ni
nM
A
T
L
A
B
.I
nef
i
g
u
r
ewindowb
o
t
ht
h
ef
u
n
c
t
i
o
n(
d
r
a
w
na
st
h
ec
u
r
v
e
)a
s
f
a
n
daxa
r
ed
r
a
w
n
.Ast
h
en
e
i
g
h
b
o
r
i
n
gp
o
i
n
t
sa
r
cc
l
o
s
e
r

w
e
l
la
st
h
el
i
n
er
e
p
r
e
s
e
n
t
i
n
g4
i
ti
se
v
i
d
e
omt
h
ef
i
g
u
r
et
h
a
tt
h
ec
u
r
v
ec
a
nb
ea
p
p
r
o
x
i
m
a
t
e
db
yas
t
r
a
i
g
h
tl
i
n
e
.
e
r
i
v
a
t
i
v
ei
st
a
n
g
e
n
tt
ot
h
ec
u
r
v
ea
s e
ya
r
ei
n
d
i
s
t
i
n
g
u
i
s
h
a
b
l
e
.
Whenax=0.001ed
HherD
e
r
i
v
a
t
i
v
e
s
:Th
ed
e
r
i
v
a
t
i
v
eo
ft
h
ed
c
r
i
v
a
t
i
v
ei
st
h
ec
a
l
l
e
desecond
o
r
m
a
l
l
yi
ti
sd
c
f
i
n
e
d
d
e
r
i
v
a
l
i
v
e
.F

NumericalD
e
r
i
v
a
t
i
v
eCompu
'
i
o
n
: Manyn
u
m
c
r
i
c
a
lt
e
c
h
n
i
q
u
c
si
nNLP(1 u
i
ec
o
m
p
u
t
a
t
i
o
no
fd
e
r
i
v
a
t
i
v
e
s
.Mosts
o
f
t
w
a
r
ei
m
p
l
e
m
e
n
t
i
n
gt
h
e
s
et
e
c
h
n
i
q
u
e
sd
on
o
t

u
s
es
y
m
b
o
l
i
cc
o
m
p
u
t
a
t
i
o
n
.A
u
t
o
m
a
t
i
o
na
n
d s
eo
f
u
s
e q
u
i
r
ea
te
s
ed
e
r
i
v
a
t
i
v
e
s
er
e
s
u
l
t
si
nT
a
b
l
e4
.
1j
u
s
t
i
f
yen
u
m
e
r
i
c
a
lc
o
m
p
u
t
a
t
i
o
n
b
ec
o
m
p
u
t
e
dn
u
m
e
r
i
c
a
l
l
y
.l
o
fad
e
r
i
v
a
t
i
v
et
h
r
o
u
g
hat
e
c
h
n
i
q
u
ec
a
l
l
e
dt
h
e
f
i
r
s
l
f
o
r
w
a
r
dd

r
e
n
c
e[
5
]
.U
s
i
n
gav
e
r
y
s
m
a
l
lp
e
u
r
b
a
t
i
o
naxl
h
d
e
r
i
v
a
t
i
v
ea
tap
o
i
n
li
sn
u
m
e
r
i
c
a
l
l
yc
a
l
c
u
l
a
t
e
da
st
h
er
a
t
i
o

!
!
.
.
(
-li"L
!
;
r
n .
1
.
4yalx
t
J
i
l
-d
xl
d
x
)

1
0 ax

S
i
m
i
l
a
r
l
yt
h
el
h
i
r
dd
e
r
i
v
a
l
i
v
ei
sd
e
f
i
n
e
da
s

ax

(
4.
16
)

164

4
.
2 MATHEMATICALCONCEPTS

NONL
lNEARPROGRAMMING

t
a
n
g
e
n
t-s
l
o
p
e-d
e
r
l
v
a
t
i
v
ea
tx
=3

d
f
l _I;~~(x +ax

y
)-f
(
x

y
)
d
x
l
^

/
:
u
1
(
.
<
.)1)

24

22

165
(
4.
19
)

Thca
b
o
v
er
e
l
a
t
i
o
ne
x
p
r
e
s
s
e
st
h
ep
a
r
t
i
a
ld
e
r
i
v
a
t
i
v
eo
ft
h
ef
u
n
c
t
i
o
n
f
w
i
t
hr
e
s
p
e
c
tt
oE
v
a
r
i
a
b
l
ex
.G
r
a
p
h
i
c
a
l
l
yt
h
i
ss
u
g
g
c
s
t
st
h
a
tt
h
ctwop
o
i
n
t
sa
r
ed
i
s
p
l
a
c
e
dh
o
r
i
z
o
n
t
a
l
l
y
(
a
s
s
u
m
i
n
gt
h
ex
a
x
i
si
sb
o
r
i
z
o
nt
a
1
)
.I
nt
h
eabovee
x
p
r
e
s
s
i
o
n(
a
n
dp
r
i
o
ro
n
e
st
o
o
)
t
h
e
s
u
b
s
c
r
i
p
ta
f
t
e
rt
h
ev
e
r
t
i
c
a
l
l
i
n
ee
s
t
a
b
l
i
s
b
e
st
b
ep
o
i
n
ta
u
t
w
h
i
c
hep
a
r
t
i
a
ld
e
r
i
v
a
t
i
v
ew
i
t
h
p
e
c
tt
oxi
sb
e
i
n
ge
v
a
l
u
a
t
c
d
.(
x

y
) s
e
a
n
y
/
a
l
lp
o
i
n
t
s
.As
i
m
i
l
a
re
x
p
r
e
s
s
i
o
nc
b
e
w
r
i
t
t
e
nf
o
rt
h
ep
a
r
a
ld
e
r
i
v
a
t
i
v
eo
f
f
w
i
t
h s
p
e
c
tt
oy
.F
o
rt
h
i
se
x
a
m
p
l
e
:

l
o
1
2
+
(
x
t
)

(
x
l
)

(
x
2
1

(
x
3
)
2
01-"
1
8

[-~.f

:"=2y
d
x-2
'
d
y

= =

y 1
:
F
o
rt
h
ep
o
i
n
t
x 2

d
f d
f
dx-" d
y

1
2

f
(
2
I
)=1
; =1 =2

1
0

U
n
t
i
lt
h
i
sp
o
i
n
twc h
a
v
e madeu
s
eo
fs
y
m
b
o
l
sr
e
p
r
c
s
e
n
t
i
n
gc
h
a
n
g
e
si
nv
a
l
u
e
s
v
a
r
i
a
b
l
e
s
)w
i
t
h
o
u
taf
o
r
r
n
a
ld
e
f
i
n
i
t
i
o
n
.I
ni
sbook
(
f
u
n
c
t
i
o
n
s

a
o

0
.
5

1
.
5

2
.5

3
.
5

F
l
g
u
r
e4
.
2l
Iu
s
l
r
a
l
l
o
n0
1t
h
el
a
n
g
e
n
.
t

=)

(
4
.
1
7
)

s(): p
r
e
s
e
n
t
sf
i
n
i
t
e
l
s
i
g
n
i
f
i
c
a
n
tc
h
a
n
g
e
si
nt
h
eq
u
a
n
t
i
t
y()
d(
)o
():r
e
p
r
e
s
e
n
t
sd
ie
r
c
n
t
ia
l
/
i
n
f
i
n
i
t
e
s
i
m
a
lc
h
a
n
g
e
si
n()

Changesi
nf
u
n
c
t
i
o
n
so
c
c
u
rduet
oc
h
a
n
g
e
si
nt
h
ev
a
r
i
a
b
l
e
s
.Fromc
a
l
c
u
l
u
s[
1
]t
h
c
d
i r
e
n
t
i
a
lc
h
a
n
g
ei
nj
{
x

y
)(
duet
ot
h
ed
i
f
f
e
r
e
n
t
i
a
lc
h
a
n
g
ei
n v
a
r
i
a
b
l
e
sx(dx)
andy(dy)i
se
x
p
r
e
s
s
e
da
s

d
jd
.d
fdv
df=d
:
J
x+
:
J
x_
.
.d
y
J
L"

"
J

T
h
i
sc
a
ngoonp
r
o
v
i
d
e
def
u
n
c
t
i
o
nh
a
ss
u
f
f
i
c
i
e
n
th
i
g
h
e
r
o
r
d
e
rd
e
p
e
n
d
e
n
c
eonE
i
n
d
e
p
e
n
d
e
n
tv
a
r
i
a
b
l
ex
.

F
u
n
c
t
i
o
no
fTwoV
B
r
i
B
b
l
e
s
:Thel
w
o
v
a
r
i
a
b
l
ef
u
n
c
t
i
o
n
j
{
x
:
x
214+y2-1

(
4
.
2
0
)

Forc
o
n
v
c
n
i
e
n
c
eands
i
m
p
l
i
c
i
t
yt
h
es
u
b
s
c
r
i
p
tr
e
p
r
e
s
e
n
t
i
n
gep
o
i
n
twhcret
h
e
e
x
p
r
c
s
s
i
o
ni
se
v
a
l
u
a
t
e
di
sn
o
ti
n
d
i
c
a
t
e
d
.Thed
e
f
i
n
i
t
i
o
no
ft
h
ep
a
r
t
i
a
ld
e
r
i
v
a
t
i
v
ei
s
a
p
p
a
r
e
n
ti
nt
h
ea
b
o
v
ee
x
p
r
e
s
s
i
o
na
sh
o
l
d
i
n
gya
tac
o
n
s
t
a
n
tv
a
l
u
ei
m
p
l
i
e
sa
t=0.
A
n
o
t
h
e
ri
n
t
e
r
p
r
e
t
a
t
i
o
nf
o
rt
h
ep
a
r
t
i
a
ld
e
r
i
v
a
t
i
v
ec
a
nb
eo
b
s
e
r
v
e
dfromt
h
ca
b
o
v
e
d
e
f
i
n
i
t
i
o
n
:c
l
t
a
n
g
eo
ft
l
t
ef
u
n
c
t
i
o
np
e
ru
n
i
tc
i
t
a
n
g
ei
nt
l
t
ev
a
r
i
a
b
l
e
s
.

(
4
.
1
8
)

I
lu
s
t
r
a
t
et
h
er
e
q
u
i
r
e
dm
a
t
h
c
m
a
t
i
c
a
lc
o
n
c
e
p
t
s
.Thef
i
r
s
ti
m
p
o
r
t
a
n
tf
c
a
t
u
r
e
i
sc
h
o
s
e
nt
oi
o
ftwoo
rmorev
a
r
i
a
b
l
e
si
s t
h
ed
e
r
i
v
a
t
i
v
e
sd
e
f
i
n
e
df
o
ras
i
n
g
l
ev
a
r
i
a
b
l
e(
o
r
d
i
n
a
r
y
d
e
r
i
v
a
t
i
v
e
s
)don
o
t
a
p
p
l
y
.
ee
q
u
i
v
a
l
e
n
tc
o
n
c
e
p
th
e
r
ei
st
h
ep
a
r
t
i
a
ld
e
r
i
v
a
t
i
v
e
.P
a
r
t
i
a
l
d
e
r
i
v
a
t
i
v
e
sa
r
edcfmedf
o
r
c
a
c
hi
n
d
e
p
e
n
d
c
n
tv
a
r
i
a
b
l
e
.l
ep
a
r
t
i
a
ld
e
r
i
v
a
t
i
v
ei
sd
e
n
o
t
e
d
.The
s
ed
c
r
i
v
a
t
i
v
e
sa
r
eo
b
t
a
i
n
e
di
nt
h
esamewaya
st
h
eo
r
d
i
n
a
r
y
byt
h
esymbold
d
e
r
i
v
a
t
i
v
ee
x
c
e
p
tt
h
co
t
h
e
rv
a
r
i
a
b
l
e
sa
r
eh
e
l
da
tac
o
n
s
t
a
n
tv
a
l
u
e
.I
nt
h
ec
x n
p
l
e
when
c
o
m
p
u
l
i
n
gep
t
i
a
ld
e
r
i
v
a
t
i
v
eo
fxt
h
cv
a
l
u
eo
fyi
sn
o
ta
l
l
o
w
e
dt
oc
h
a
n
g
e
.

G
r
a
d
i
e
n
t0
1t
h
eF
u
n
c
t
i
o
n
:I
nt
h
ef
u
n
c
t
i
o
no
fas
i
n
g
l
ev
a
r
i
a
b
l
et
h
ed
e
r
i
v
a
l
i
v
cwas
a
s
s
o
c
i
a
t
e
dwi
t
h
es
l
o
p
. I
ntwoo
rmorev
a
r
i
a
b
l
e
st
h
es
l
o
p
ei
se
q
u
i
v
a
l
e
n
t1
0e
g
r
a
d
i
e
n
.
tTh
eg
r
a
d
i
e
n
ti
sav
e
c
t
o
r

anda
ta
n
yp
o
i
n
tr
e
p
s
e
n
t
st
h
ed
i
r
e
c
t
i
o
ni
nw
h
i
c
h
t
h
e
f
u
n
c
t
i
o
nw
i
l
li
n
c
r
e
a
s
em
o
s
tr
a
p
i
d

E
x
a
r
n
i
n
i
n
gt
h
ec
o
n
v
e
n
t
i
o
n
a
lo
b
j
e
c
t
i
v
eo
fNLP
m
i
n
i
m
i
z
a
t
i
o
no
fo
b
j
e
c
t
i
v
em
c
t
i
o
n
s
t
h
eg
r
a
d
i
e
n
th
a
san
a
t
u
r
a
lp
a
r
tt
op
l
a
yi
nt
h
e
d
e
v
e
l
o
p
m
e
n
to
fmethodst
os
o
l
v
et
h
ep
r
o
b
l
e
m
.Theg
r
a
d
i
e
n
ti
scomposedo
ft
h
ep
a
r
t
i
a
l
d
c
r
i
v
a
t
i
v
e
so
r
g
a
n
i
z
e
da
sav
c
c
l
o
r
.V
e
c
t
o
r
si
n i
sbooka
r
ecolumnv
e
c
t
o
r
su
n
l
e
s
so
t
h
c
r
n
o
t
e
d
.Theg
r
a
d
i
e
n
th
a
sas
t
a
n
d
a
r
dm
a
t
h
e
m
a
t
i
c
a
lsymbo.
ll
ti
sd
c
f
i
n
e
da
s

4
.
2M
A
T
H
E
M
A
T
I
C
A
L
C
O
N
C
E
P
T
S 167

166 NONL
lN
EARPROGRAMMING

Vf

1
=x
2
/4+I-l
3
Dc
o
n
l
o
u
rf
o
r(
f

(
4
.
2
1
)

3
2

si
n
f
o
r
m
a
t
i
o
nu
s
i
n
gg
r
a
p
h
i
c
s
.Th
e
Att
h
i
ss
t
a
g
ei
ti
sa
p
p
r
o
p
r
i
a
t
et
oc
o
n
s
o
l
i
d
a
t
ei
g
r
a
p
h
ic
a
Ids
c
r
i
p
t
i
o
no
fee
x
n
p
l
ed
e
f
i
n
e
di
nE
q
u
a
t
i
o
n(
4.
18
)h
a
st
obet
h
r
e
e

ano
e
rf
o
ryandt
h
et
h
i
r
df
o
rj
{
x
y
)
.C
h
a
p
t
e
r
d
i
m
e
n
s
i
o
n
a
la
swen
e
e
donea
x
i
sf
o
rx
2i
n
t
r
o
d
u
c
e
dt
h
r
e
e
d
i
m
e
n
s
i
o
n
a
lp
l
o
t
t
i
n
gu
s
i
n
gMATLAB.Amoreu
s
e
f
u
ld
e
s
c
r
i
p
t
i
o
no
f
t
h
ep
r
o
b
l
mi
st
ou
s
ec
o
n
t
o
u
rp
l
o
t
s
.C
o
n
t
o
u
rp
l
o
t
sa
r
edrawnf
o
rs
p
e
c
i
f
i
cv
a
l
u
e
so
ft
h
e
f
u
n
c
t
i
o
n
.I
nt
h
ei
I
Iu
s
t
r
a
t
i
o
nt
h
ev
a
l
u
e
sf
o
rt
h
ec
o
n
t
o
u
r
so
f
f
a
r
e01

2
d3
.I
nF
i
g
u
r
e
4
.
3twok
i
n
d
so
fc
o
n
t
o
u
rp
l
o
t
sa
r
es
h
o
w
n
.I
nt
h
et
o
ph
a
l
fat
h
r
e
e
d
i
m
e
n
s
i
o
n
a
l(
3
D
)
c
o
n
t
o
u
rp
l
o
ti
sshown

w
h
i
l
eon l
o
w
e
r
h
a
l
f samei
n
f
o
r
m
a
t
i
o
ni
sp
r
e
s
e
n
a
t
w
o
d
i
m
e
n
s
i
o
n
a
l(
2
D
)c
o
n
t
o
u
rp
l
o
t(
t
h
i
si
sc
o
n
s
i
d
e
dt
h
s
t
a
n
d
a
r
dc
o
n
t
o
u
rp
l
o
t
)
.Thc
2Dc
o
n
t
o
u
rp
l
o
ti
smoreu
s
e
f
u
lf
o
rg
r
a
p
h
ic
a
Is
u
p
p
o
r
to
fsomeo
ft
h
ei
d
e
a
si
nNLP.Th
e
dw
i
t
hi
ner
e
m
a
i
n
d
e
ro
ft
h
cb
o
o
k
.
3Dc
o
n
t
o
u
rw
i
l
lbed
i
s
p
e
n
s

0
3

2
.
5

2
D
c
o
n
l
o
u
r
3

2
.
5

MATLABC
ode:Flg4_3.m. Thea
n
n
o
t
a
t
i
o
ni
nF
i
g
u
r
e4
.
3(
i
n
c
l
u
d
i
n
gt
h
et
a
n
g
e
n
tl
i
n
c
o
u
g
ht
h
ep
l
o
t i
t
i
n
gcommandsa
v
a
i
l
a
b
l
eone
andt
h
eg
r
a
d
i
e
n
t ow)wasdone r
menub
s
ri
nt
h
ef
i
g
u
r
ewindow(MATLADv
e
r
s
i
o
n5
.
2a
n
dl
a
t
e
r
)
.Am01
"p
O
l
"t
i
o
nu
f
cp
l
o
t
sa
g
e
n
e
r
a
t
e
dt
h
r
o
u
g
ht
h
es
t
a
t
c
m
e
n
t
si
nt
h
em-f
i
Iei
n
d
i
c
a
t
e
da
b
o
v
e
.Thec
o
d
c
mixesn
u
m
e
r
i
cand s
y
m
b
o
l
i
cc
o
m
p
u
t
a
t
i
o
n
.I
ta
l
l
o
w
sm
u
l
t
i
p
l
ep
l
o
t
sandt
a
r
g
e
t
s
g
r
a
p
h
i
ccommandst
os
p
e
c
i
f
i
cp
l
o
t
s
.

g
l
l
l
d
l
e
n
l0
1S

- longenl815
2

F
l
g
u
r

4
.
3G
r
a
d
i
e
n
la
n
dl
a
n
g
e
n
!I
I
n
a
!ap
o
i
n
.
t

D
i
s
c
u
s
s
i
o
no(F
i
g
u
r
e4
.
3
:I
nF
i
g
u
r
e4
.
3p
o
i
n
tPi
sonc
o
n
t
o
u
rf=O
.P
o
i
n
tQ i
son
t
h
econtourf=2
.P
o
i
n
tSi
sont
h
econtourf=1
.P
o
i
n
tRh
a
sesameyv
a
l
u
ca
sp
o
i
n
t
P d esamexva
Ju
ca
sp
o
i
n
tQ.Forev
a
l
u
e
si
nt
h
ef
i
g
u
r
et
h
ec
o
n
t
o
u
rv
a
l
u
ei
s0
.
7
5
(
t
h
i
sv
a
l
u
es
h
o
u
l
db
ed
i
s
p
l
a
y
e
di
nMATLABwindowwhcnt
h
ecod
F
lg4_3.mi
sr
u
n
)
.

From f
i
g
u
r
et
h
ev
a
l
u
eoffw
i
I
lchangea
l
o
n
gg
r
a
d
i
e
n
td
i
r
e
c
t
i
o
n
.I
fd
f
r
e
p
r
e
s
e
n
t
sa
d
i
f
f
e
r
e
n
t
ia
Jmovea
J
ong g
e
n
tl
i
n
e
en(
dX a
r
em
e
a
s
u
r
e
da
J
ong g
e
n
tl
i
n
e
)

df=~dX+ =0
eJ
y

l
i
n
ePQi
sameasure4
f
l
i
n
ePRr
e
p
r
e
s
e
n
t
sc
h
a
n
g
e
si
ndfwhcnd
.
yi
s0

E
q
u
a
t
i
o
n(
4
.
2
3
)s
h
o
u
l
db
cz
e
r
ob
e
c
a
u
s
emovingt
a
n
g
c
n
t
i
a
l
l
y(
as
m
a
l
la
m
o
u
n
t
)t
h
e
v
a
Ju
o
f
f
i
sn
o
tc
h
a
n
g
e
d
.

lineRQ p
r
e
s
e
n
t
sc
h
a
n
g
e
si
ndfwhend
.
xi
s0
M
a
t
h
e
m
a
t
i
c
a
l
l
y

fcanonlybeestimatedbyaddingthechangcsalongthelinesPR
andRQsincecd
e
f
i
n
i
t
i
o
no
ft
h
ep
a
r
t
i
a
ld
e
r
i
v
a
t
i
v
e
so
n
l
yp
c
r
m
i
t
sc
a
l
c
u
l
a
t
i
n
gc
h
a
n
g
c
s
a
l
o
n
gt
h
ec
o
o
r
d
i
n
a
t
ed
i
r
e
c
t
i
o
n
s
.
h
ed
o
t
t
e
dl
i
n
ei
st
h
et
a
n
g
e
n
t
.Theg
r
a
d
i
e
n
ta
tt
h
csamcp
o
i
n
ti
sn
o
r
m
a
l
Atp
o
i
n
tSt
(
p
c
r
p
c
n
d
i
c
u
l
a
r
)t
oct
I
.
n
g
e
n
t di
sd
i
r
t
e
dt
o
w
a
r
di
n
c
r
e
a
s
i
n
gt
h
eva
Ju
eo
fe
f
u
n
c
t
i
o
n(
i
n
d
i
c
a
t
e
dbyea
r
r
o
w
)
.Byd
e
f
i
n
i
t
i
o
n
i
fd
f
r
e
p
r
e
s
e
n
t
sad
i
f
f
e
r
e
n
t
ia
Jmove
a
Jongt
h
eg
r
a
d
i
e
n
ta
tanyp
o
i
n
t
e
n(dxP m
e
a
s
u
r
e
da
l
o
n
gt
h
eg
r
a
d
i
e
n
tv
e
c
t
o
r
)

+
dy

df=

(
4
.
2
3
)

(
4
.
2
2
)

Th
eJ
a
c
o
b
i
a
n[
J
]d
e
f
i
n
e
sau
s
e
f
u
lwayt
oo
r
g
a
n
i
z
et
h
eg
r
a
d
i
e
n
t
so
fs
e
v
c
r
a
J
f
u
n
c
t
i
o
n
s
.U
s
i
n
gt
h
r
e
cv
a
r
i
a
b
l
e
sandtwof
u
n
c!
i
o
n
sj
{
x
y
z
)a
n
dg(x
y
z
)t
h
ed
e
f
i
n
i
t
i
o
n
o
ft
h
cJ
a
c
o
b
i
a
ni
s

~acobian:

I
e
J
f
[
J
J=
1~x

~ ~~ 1

I~|

(
4
.
2
4
)

168

4
.
2 MATHEMATIC
AlCONCEPTS

NONL
lNEARPROGRAMMING

1
6
9

I
nEq
u
a
t
i
o
n(
4
.
2
4
)t
h
eg
r
a
d
i
e
n
t
so
f
t
h
ef
u
n
c
t
i
o
na
p
p
i
nesamer
o
w
.Thef
i
r
s
trow

Eq
u
a
t
i
o
n
s(
4
.
2
7
)and(
4
.
2
8
)w
i
l
lapp
rq
u
i
t
eo
f
t
e
ni
ns
u
c
c
e
e
d
i
n
gc
h
a
p
t
e
r
s
.A f
e
w

r
a
d
i
e
n
to
f
fwhilethesecondrowi
st
h
eg
r
a
d
i
e
n
tofg
.l
ft
h
etwof
u
n
c
t
i
o
n
sa
r
e
i
s eg
d
fdg]
Ti
nef
u
n
c
t
i
o
n
s
d
u
e
c
o
U
e
c
t
e
di
n
t
oacolumnv
e
c
t
o
r
.t
h
ed
i
f
f
e
r
e
n
t
i
a
lc
h
a
n
g
e
s[
t
oed
i
f
f
e
r
e
n
t
i
a
lc
h
a
n
g
ei
nt
h
cv
a
r
i
a
b
l
e
s[
d
xdyd
z
]
c
a
nb
ee
x
p
r
e
s
s
e
da
sam
a
t
r
i
x
m
u
l
t
i
p
u
c
a
t
i
o
nu
s
i
n
gt
h
eJ
a
c
o
b
i
a
n

m
i
n
u
t
e
so
ff
a
m
i
l
i
a
r
i
z
a
t
i
o
nw
i
l
lp
r
o
v
i
d
es
u
s
t
a
i
n
e
dcompreh
n
s
i
o
nl
a
t
e
r
.

+))+

whichi
ss
i
m
i
l
a
rt
oE
q
u
a
t
i
o
n(
4
.
2
0
)
.

()

I~

H
e
s
s
i
a
n
:TheH
e
s
s
i
a
nm
a
t
r
i
x[
H
]i
st
h
esamea
st
h
em
a
t
r
i
xo
fs
e
c
o
n
dd
e
r
i
v
a
t
i
v
e
so
f
t
.
.
x
y
)
af
u
n
c
t
i
o
nofs
e
v
e
r
a
lv
a
r
i
a
b
l
e
s
.Forf

Ida
[

SlngleV
a
r
l
a
b
l
e
:Th
eT
a
y
l
o
rs
e
r
i
e
si
sau
s
e
f
u
lmechanismt
oa
p
p
r
o
x
i
m
a
t
eev
a
l
u
e
o
ft
h
ef
u
n
c
t
i
o
nf
t
.
.
x
)a
t p
o
i
n
t(
x
)i
fef
u
n
c
t
i
o
ni
sc
o
m
p
l
e
t
e
l
yknowna
tp
o
i
n
t
p+
xp le e
xp
s
i
o
ni
s(
f
o
rf
i
n
i
t
cn)

rill--L

]4

lEE E

rBEE-EEEL

=
-J

(
4
.
2
5
)

4
.
2
.
3 Taylor'sTheorern/Serles

=
1
I
i
JL
[ i
J
L
[I

(
4
.
2
6
)

Th
eH
e
s
s
i m
a
t
r
i
xi
ssymm
i
c
.F
o
r exampled
e
f
i
n
e
dbyE
q
ui
o
n(
4
.
1
8
)
.
h

n
v &

fAU

(
4
.
2
7
)
Th
eH
e
s
s
i
a
ni
s

d
'
i
d
x1
i
J
x
2
~ a
'
i
=
1dX
2a
x1 ~

i
J
x
n
d
x

f
l
l

i
J
x
1
a
x
n

(
4
.
2
8
)

d
.
c

(
4
.
3
0
)

l
ft
h
ef
i
r
s
tt
e
r
mi
sb
r
o
u
g
h
tt
oel
e
f
t ee
q
u
a
t
i
o
n
.d
i
s
c
a
r
d
i
n
gt
h
ee
r
r
o
rt
e
r
mc
a
nb
e
w
r
i
t
t
e
na
s

FunctlonofnV
a
r
l
a
b
l
e
s
: Forf
t
.
.X
)
.where]=
[
X
I
X
2
.... x

.
r
r
.eg
r
a
d
i
e
n
ti
s

Thes
e
r
i
e
si
sw
i
d
e
l
yu
s
e
di
nmostd
i
s
c
i
p
l
i
n
e
st
oe
s
t
a
b
l
i
s
hc
o
n
t
i
n
u
o
u
sm
o
d
e
l
s
.I
ti
sE
m
a
i
n
s
t
a
yo
fmanyn
u
m
e
r
i
c
a
lt
e
c
h
n
i
q
u
e
si
nc
Iu
d
i
n
gt
h
o
s
ei
no
p
t
i
m
i
z
a
t
i
o
n
.E
q
u
a
t
i
o
n
(
4
.
2
9
)i
su
s
u
a
l
l
yt
r
u
n
c
a
t
e
dthef
i
r
s
ttwoo
r t
e
r
m
swi eu
n
d
e
r
s
t
a
n
d
i
n
g
a
p
p
r
o
x
i
m
a
t
i
o
nw
i
l
ls
u
f
f
e
rsomee
r
r
o
rwhoseo
r
d
e
rd
e
p
e
n
d
son t
e
r
m ti
sb
e
i
n
g
t
r
u
n
c
a
t
e
d
:

+
{AY.
1
.
1
.
[(Xp+A
t
)f(XJ&Jh)+21dlJ
)
2+O( )
3

I
d
y
d
x

I~ (2+

M
=
f

)
f

)
)

)
=

p+
d
x
L()+
2
1d
X
1

)
(
4
.
31

-Ab-xh

I
nEquation(
4
.
3
1
)
t
h
ef
i
r
s
tt
e
r
moner
i
g
h
ti
sc
a
l
l
e
dt
h
e
f
i
r
s
t
o
r
d
e
r
/
l
i
n
e
a
rv
a
r
i
a
t
i
o
n
w
h
i
l
et
h
es
e
c
o
n
dt
e
r
mi
ses
e
c
o
n
d
o
r
d
e
r
/
q
u
a
d
r
a
t
i
cv
a
r
i
a
t
i
o
n
.
F
i
g
u
r
e4.
4d
e
m
o
n
s
t
r
a
t
e
st
h
ea
p
p
r
o
x
i
m
a
t
i
o
n
su
s
i
n
gT
a
y
l
o
r
'ss
e
r
i
e
so
fv
a
r
i
o
u
so
r
d
e
r
s
5w
i
t
hr
e
s
p
e
c
tt
o o
r
i
g
i
n
a
lf
u
n
c
t
i
o
n(
r
e
d
)
.Th
ep
r
i
n
c
i
p
a
li
d
e
ai
st
od
e
a
lw
i

a
tp
o
i
n
t2.
t
h
ea
p
p
r
o
x
i
m
a
t
i
n
gc
u
r
v
ewhichh
a
sknownp
r
o
p
e
e
si
n
s
t
e
a
do
ft
h
eo
r
i
g
i
n
a
lc
u
r
v
e
.
l
Cc
o
n
s
t
a
n
tT
a
y
l
o
rs
e
r
i
e
si
sw
o
e
f
u
l
l
yi
n
a
d
e
q
u
a
t
e
.l
el
i
n
e
a
re
x
p
a
n
s
i
o
ni
so
n
l
y
m
a
r
g
i
n
a
l
l
yb
e
t
t
e
r
.Theq
u
a
d
r
a
t
i
ce
x
p
a
n
s
i
o
na
p
p
r
o
x
i
m
a
t
e
st
h
er
i
g
h
ts
i
d
eo
ft
h
ef
u
n
c
t
i
o
n
t
ot
h
el
e
f
ts
i
d
ewhichi
si
ns
i
g
n
i
f
i
c
a
n
te

o
r
. l
ef
i
f
t
h
o
r
d
e
re
x
p
a
n
s
i
o
ni
s
comp
d
e
f
i
n
i
t
e
l
ya
c
c
e
p
t
a
b
l
ei
nt
h
er
a
n
g
es
h
o
w
n
.Tos
e
ehowt
h
eT
a
y
l
o
rs
e
r
i
e
si
su
s
e
d
c
o
n
s
i
d
e
rt
h
eq
u
a
d
r
a
t
i
cc
u
r
v
ea
b
o
u
tcp
o
i
n
tx=2
.
5
.Att
h
i
sp
o
i
n
t v
a
l
u
eo
ft
h
e
f
u
n
c
t
i
o
nf
t
.
.
2
.
5
) 1
1.
43
7
5
t
h
ev
a
l
u
eo
ft
hf
i
r
s
td
e
r
i
v
a
t
i
v
e
f
'
(
2
.
5
)0
.
7
5
d
t
h
c
v
a
l
u
c
o
ft
h
es
e
c
o
n
dd
c
r
i
v
a
t
i
v
ei
sf
"
(
2
.
5
) 4
.UsingE
q
u
a
t
i
o
n(
4
.
3
1
)

f
t
.
.2.
5+ =1

5+(
-0)+

(
4
.
3
2
)

F
o
rd
i
f
f
e
r
e
n
tv
a
l
u
e
so
fA
t
.bo p
o
s
i
t
i
v
candn
c
g
a
t
i
v
e
t
h
ev
a
l
u
eo
f
j
(
x
)c
a
nb
eo
b
t
a
i
n
e
d
.
l
ep
l
o
to
fE
q
u
a
t
i
o
n(
4.
32
)s
h
o
u
l
db
et
h
esamea
s o
n
el
a
b
e
l
e
dq
u
a
d
r
a
t
i
ci
nF
i
g
u
r
e4.
4
.

1
7
0

NONL
lNEARPROGRAMMING

4
.
3 GRAPHICALSOLUTIONS

A
p
p
r
o
x
l
m
a
t
!
o
nu
s
!
r
t
gT
a
y
l
o
rs
e
r
l
e
s

1
7
1

f(xp+dx

+)=f(xp')++
I~YI

22

iMyPyp)][~]
-0

2
0

1
1
t

Fornv
a
b
l
e
s

w
i
t
ht
h
ec
u
r
r
e
n
tp
o
i
n
ta
n
dd
.
Xt
h
ed
i
s
p
l
a
c
e
m
e
n
tv
e
c
t
o
r

f
p+
I
l
X)=f(X
+jmTm
p)+

18

(
4
.
3
4
)

(
4.
35
)

f
i
f
t
hor
d
e
ra~d o
r
l
g
!
n
a
l

4
.
3 GRAPHICALSOLUTIONS
E

l
eg
r
a
p
h
i
c
a
1s
o
l
u
t
i
o
n
sa
r
ep
r
e
s
e
n
t
e
df
o
rt
h
r
e
ek
i
n
d
so
fp
r
o
b
l
e
m
s
:u
n
c
o
n
s
a
i
n
e

de
q
u
a
l
i
t
y

c
o
n
s
t
r
a
i
n
e
d

a
n
di
n
e
q
u
a
l
i
t
yc
o
n
s
i
n
e
d

b
a
s
e
do
nt
h
ef
u
n
c
t
i
o
n
si
nE
x
a
m
p
l
e4
.
1
.

4
.
3
.
1 UnconstralnedProblem
12

M
i
n
i
m
i
z
e J
t
XI
x
2
)
:XIX2
D

1
.
5

0
.
5

2
.
5

3
.
5

F
i
g
u
r
e
4
.
4T
a
y
l
o
rs
e
r
i
e
sa
p
p
r
o
x
i
m
a
l
i
o
n
.
MATLABC
ode:F
i
g
u
r
e4
.
4i
sc
o
m
p
l
e
t
e
l
yc
r
e
a
t
e
du
s
i
n
gt
h
ec
o
d
eFlg4_
4
.
m
.1
tu
s
e
st
h
e
M
A
T
L
A
B
p
r
o
v
i
d
e
ds
y
m
b
o
l
i
c
t
a
y
l
o
c
"f
u
n
c
t
i
o
nt
og
e
n
e
r
a
t
et
h
ev
a
r
i
o
u
se
x
p
a
n
s
i
o
n
s
a
b
o
u
tp
o
i
n
t2
.
5
.

TwoorMoreV
a
r
i
a
b
l
e
s
:Thes
e
r
i
e
s o
n
l
ye
x
p
a
n
d
e
dt
o qu
'
a
t
i
ct
e
r
m
s
.The
t
r
u
n
c
a
t
i
o
ne
r
r
o
ri
si
g
n
o
r
e
d
.
'
h
et
w
o
v
a
r
i
a
b
l
ef
u
n
c
t
i
o
ne
x
p
a
n
s
i
o
ni
sshowni
nd
e
t
a
i
1a
n
d
a
l
s
oo
r
g
a
n
i
z
e
di
nt
e
r
m
so
fv
e
c
t
o
r
sa
n
dm
a
t
r
i
c
e
s
.Thef
i
r
s
t
o
r
d
e
re
x
p
a
n
s
i
o
ni
se
x
p
r
e
s
s
e
d
i
nt
e
r
m
so
ft
h
eg
r
a
d
i
e
n
.
tThes
e
c
o
n
d
o
r
d
e
re
x
p
a
n
s
i
o
nw
i
l
lb
ee
x
p
r
e
s
s
e
di
nt
e
r
m
so
ft
b
e
H
e
s
s
i
a
nma
i
x

I
M
P
I
+
I

(w-Y+
I
t
pY
/

+
(
1}

f
(
x
'
J Y
/
J+
I
l
y
) J
t
x
'
J
'y
)+1

2
1)~I (x

2 q

.
I
"
p

e
)
2
f

(
4.
33
)

If d
i
s
p
l
a
c
e
m
e
n
t
sa
r
eo
r
g
a
n
i
z
e
da
sac
o
l
u
m
nv
e
c
t
o
r[]
T
t
h
ee
x
p
a
n
s
i
o
ni
n
(
4
.
3
3
)c
a
nb
ee
x
p
r
e
s
s
e
di
nac
o
n
d
e
n
s
e
dm
a
n
n
e
ra
s

o~xl ~ 3
;

0~x2 ~ 3

(
4
.
3
6
)
(
4.
37
)

Thes
i
d c
o
n
s
t
r
a
i
n
t
ss
r
v
et
ol
i
m
i
tt
h
ed
e
s
i
g
ns
p
a
c
e
.C
o
n
s
i
d
e
r
i
n
gt
h
eo
b
j
e
c
t
i
v
e
f
u
n
c
t
i
o
ni
n(
4
.
3
6
)i
ti
sc
l
e
a
rt
h
a
tt
h
eminimumv
a
l
u
eo
ffwi
1
lb
er
a
l
i
z
e
di
ft
h
e
v
a
r
i
a
b
l
e
sa
r
ea
tt
h
emaximum(
X
i
=3
=3
)
.Thev
a
l
u
eo
ft
h
ef
u
n
c
t
i
o
ni
s9
.
F
i
g
u
r
e4
.
5i
l
1u
s
t
r
a
t
e
st
h
ep
r
o
b
l
e
m
.Thes
o
l
u
t
i
o
ni
sa
tt
h
eb
o
u
n
d
a
r
yo
ft
h
ed
e
s
i
g
n
s
p
a
c
e
.Fig4_5.mi
st
h
eMATLABc
o
d
ea
tw
i
l
lp
r
o
d
u
c
eF
i
g
u
r
e4
.
5
.Themaximum
v
a
l
u
eo
ft
h
ef
u
n
c
t
i
o
ni
s0(
b
a
s
e
dont
h
es
i
d
ec
o
n
s
t
r
a
i
n
t
s
)
.I
nF
i
g
u
r
e4.
5t
h
et
a
n
g
e
n
t
a
n
dt
h
eg
r
a
d
i
e
n
tt
ot
h
eo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
na
tt
h
es
o
l
u
t
i
o
na
r
edrawni
nu
s
i
n
g
MATLABp
l
o
t
e
d
i
tc
ommands.l
Di
sp
a
r
t
i
c
u
l
a
re
x
a
m
p
l
et
h
es
i
d
ec
o
n
s
t
r
a
i
n
t
sa
r
e
n
e
c
e
s
s
a
r
yt
od
e
t
e
r
m
i
n
et
h
es
o
l
u
t
i
o
n
.I
ft
h
ed
e
s
i
g
ns
p
a
c
ew
e
r
ei
n
c
r
e
a
s
e
dt
h
es
o
l
u
t
i
o
n
w
o
u
l
dc
o
r
r
e
s
p
o
n
d
i
n
g
l
yc
h
a
n
g
e
.
a
m
p
l
ec
h
o
s
e
nt
oi
l
l
u
s
t
r
a
t
eeu
n
c
o
n
s
t
r
a
i
n
e
dp
r
o
b
l
e
md
e
f
i
n
e
db
y(
4
.
3
6
)a
n
d
Thex
(
4.
3
7
)i
sn
o
tu
s
u
a
l
l
ye
m
p
l
o
y
e
dt
od
e
v
e
l
o
pt
h
eo
p
t
i
m
a
l
i
t
yc
o
n
d
i
t
i
o
n
sf
o
rg
e
n
e
r
a
l
u
n
c
o
n
s
t
r
a
i
n
e
dp
r
o
b
l
e
m
s
.Twor
e
q
u
i
r
e
m
e
n
t
sa
r
ee
x
p
e
c
t
e
dt
ob
es
a
t
i
s
t
i
e
db
ys
u
c
h
p
r
o
b
l
e
m
s
:(
1
)t
h
es
o
l
u
t
i
o
nm
u
s
tb
ei
nt
h
ei
n
t
e
r
i
o
ro
ft
h
ed
e
s
i
g
ns
p
a
c
ea
n
d(
2
)t
h
e
r
em
u
s
t
b au
n
i
q
u
es
o
l
u
t
i
o
no
rep
r
o
b
l
e
mi
s unimoda
.
lN
e
v
e
r
t
h
e
l
e
s
s
i
np
r
a
c
t
i
c
a
l
a
p
p
l
i
c
a
t
i
o
n
st
h
e
s
ec
o
n
d
i
t
i
o
n
smayn
o
te
x
i
s
.Mosts
t
o
f
t
w
a
r
ea
r
ed
e
v
e
l
o
p
e
dt
oa
p
p
l
yt
h
e
o
p
t
i
m
a
l
i
t
yc
o
n
d
i
t
i
o
n
sa
n
dt
h
e
yr
a
r
e
l
yv
e
r
i
f
ye
s
er
e
q
u
i
r
e
m
e
n
t
sa
r
eb
e
i
n
gme
.I
t
ti
s
1
l
yt
h
ed
e
s
i
g
n
e
r
'
sr
e
s
p
o
n
s
i
b
i
l
i
t
yt
oe
n
s
u
r
et
h
e
s
er
e
q
u
i
r
e
m
e
n
t
sa
r
eme
.Tod
t
e
v
e
l
o
p
u
s
ua
eo
p
t
i
m
a
l
i
t
yc
o
n
d
i
t
i
o
n
sa
na
It
e
m
a
t
eu
n
c
o
n
s
t
r
a
i
n
e
dp
r
o
b
l
e
mi
sp
r
e
s
e
n
t
e
d
:
M
i
n
i
m
i
z
e J
t
XI
x
2
)
:(XI-1
)
2+(X2- 1
)
2-XIX2

(
4
.
3
8
)

7
2

NONL
lNEARPROGRAMMING

4
.
3 GRAPHICALSOLUnONS

173

3
.
5

2
.
5

1
.
5

0
.
5

0
.
5

1
.
5

2
x

2
.
5

3
.
5

x
t

'

F
l
g
u
r
e4.
5 U
n
c
o
n
s
l
r
a
i
n
e
ds
o
l
u
t
i
o
n
.

o$XI$ 3
;

O$X2$ 3

g
u
r
e4.6

4
.
3
.
2 EqualltyConstrainedProblem
F
o
rat
w
o
v
a
r
i
a
b
l
cp
r
o
b
l
e
mwec
a
no
n
l
yu
t
i
l
i
z
eonec
o
n
s
t
r
a
i
n
tf
o
ram
e
a
n
i
n
g
f
u
l
o
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
e
m
:
Imi
z
e
Min

j
{
X
I
x
2
)
:'
X
IX2

S
u
b
j
e
c
t

(
X
I
hl
x
2
)
: /
4+~ -1

X
!

OSX1$3;

(
4
.
3
7
)

.
6
.
mp
r
o
v
i
d
e
st
h
ec
o
d
e
.
t
h
a
twi
I
I
F
i
g
u
r
e4
.
6d
i
s
p
l
a
y
st
h
ep
r
o
b
l
e
ma
n
di
t
ss
o
l
u
t
i
o
n
.Flg4_
g
e
n
e
r
a
l
e most o
fF
i
g
u
r
e4
.
6
.I
ti
sc
J
e
a
r a
tt
h
es
o
l
u
t
i
o
na
p
p
e
a
r
st
ob
eo
t
x
i=2a
n
dx
i=2
.Theo
p
t
i
m
a
lv
a
l
u
eo
feo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
ni
s2
.I
nt
h
ec
o
n
t
o
u
r
p
l
o
t
eo
ptimumi
sap
o
i
n
tmakingi
td
i
f
f
i
c
u
Jt
t
odrawt
h
eg
r
a
d
i
e
n
t
.Ifi
si
st
h
ec
a
s
e

e
ni
nt
h
r
e
ed
i
m
c
n
s
i
o
n
sa
tt
h
eoptimumcg
r
a
d
i
e
n
twi
l
I
iei
nap
l
a
n
et
a
n
g
e
n
tt
ot
h
e
sp
l
a
n
es
h
o
u
l
dn
o
tc
h
a
n
g
et
h
ev
a
l
u
eo
ft
h
eo
b
j
e
c
t
i
v
c
f
u
n
c
t
i
o
ns
u
r
f
a
c
e
.Movingi
n i
i
so
b
s
e
r
v
a
t
i
o
ni
su
s
c
dt
od
e
v
e
l
o
pt
h
en
e
s
s
a
r
yc
o
n
d
i
t
i
o
n
sl
a
t
e
r
.
f
u
n
c
t
i
o
n
.Th

(
4.
36
)

(
4
.
3
9
)

U
n
c
o
n
s
t
r
a
i
n
e
dp
r
o
b
l
e
m
:i
n
t
e
r
i
o
r 1
l
I
l
l
o
n
.
OSx
2S3

(
4
.
3
7
)

F
i
u"e 4
.
7(
r
o
u
g
hFIg4

7.m) i
1
Iu
s
t
r
a
t
e
st
h
ep
r
o
b
l
e
m
.Thed
a
s
h
e
dl
i
n
ei
se

nsain.
tS
i
n
c
ei
ti
sa
nequa
Ji
t
yc
o
n
s
t
r
a
i
n
t

t
h
es
o
l
u
t
i
o
nmustb
eap
o
i
n
tont
h
ed
a
s
h
e
d

1a
p
p
e
a
r
st
oj
u
s
tg
r
ec
o
n
s
t
r
a
i
n
ta
n
dt
h
e
r
e
f
o
r
ei
st
h
e
l
i
n
e
.Thec
o
n
t
o
u
ro
ff=
Ju
eo
ft
h
ef
u
n
c
t
i
o
nw
i
t
h
o
u
tv
i
o
l
a
t
i
n
gt
h
ec
o
n
s
t
r
a
i
n
.
tI
nF
i
g
u
r
e4
.
7
minimump
o
s
s
i
b
l
eva
eg
r
a
d
i
e
n
to
ft
h
eo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
na
swe
I
la
s eg
r
a
d
i
e
n
to
f ec
o
n
s
t
r
a
i
n
t
sa
tt
h
e

i
l
Iu
s
t
r
a
t
e
d
.I
ta
p
p
e
a
r
sa
ta
ti
sp
o
i
n
tt
h
e
s
eg
r
a
d
i
e
n
l
sa
r
ep
a
r
a
l
l
e
lc
v
c
n
s
o
l
u
t
i
o
na
r
t
h
o
u
g
he
ya
r
ed
i
r
e
c
t
e
do
p
p
o
s
i
t
et
oc
a
c
ho
t
h
e
r
.Byd
e
f
i
n
i
t
i
o
nt
h
eg
r
a
d
i
e
n
ti
si
nt
h
e
d
i
r
e
c
t
i
o
no
ft
h
cmostr
a
p
i
di
n
c
r
e
a
s
eo
ft
h
ef
u
n
c
t
i
o
na
tt
h
es
e
l
e
c
t
e
dp
o
i
n.
tT
h
i
si
sn
o
ta
c
o
i
n
c
i
d
e
n
c
.T
hi
sf
a
c
ti
su
s
e
dt
oe
s
t
a
b
l
i
s
ht
h
en
e
c
e
s
s
a
r
yc
o
n
d
i
t
i
o
n
sf
o
rt
h
ep
r
o
b
l
e
m
.

I
nedProblem
4
.
3
.
3 I
n
e
q
u
a
l
l
t
yConst

Th
enumbero
fi
n
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
t
si
sn
o
td
e
p
e
n
d
e
n
tont
h
enumbero
fv
a
r
i
a
b
l
e
s
.F
o
r
i
1
I
u
s
n
bo
c
o
n
s
i
n
tf
u
n
c
t
i
o
n
so
fExample4.
1a
r
f
o
r
m
u
l
a
t
e
da
si
n
e
q
u
a
l
i
t
y
a
m
t
s
.
c
o
n
s
Minimize

j
{
x) X2):-XIX2

(
4.
36
)

174

4
.
4 ANAlYTICAlCONDITIONS

NONL
lNEARPROGRAMMING

3
.
5

3
.
5

175

2
.
5

1
.5

1
.
5

0
.
5

0
.
5

0
.
5

1
.
5

2
.
5

3
.
5

1
.
5

0
.
5

F
l
g
u
r
e4.7 E
q
u
a
l
i
t
yn
s
t
r
a
l
n
e
dp
r
o
b
l
e
m
.

x
2
)
:2
0xI+15x2-30S0
S
u
b
j
e
c
t
t
o
: gl(XI

x
I
ISO

g2(Xl
>X
2
)
: /4+

2
.
5

OSX2S3

F
l
g
u
r
e4.8 I
n
e
q
u
a
l
i
t
yc
o
n
s
t
r
a
l
n
e
dp
r
o
b
l
e
m
.

M
i
n
i
r
n
i
z
e

j
{
X
x
2
)
:x
lx2

S
u
b
j
e
c
t1
0
:

hl
x
2
)
:2
0X
I+1
5X2-30=
0
(
X
I

(
4
.
4
0
)
(
4.
41
)

(
x
I
/
4
)+(~) IS0

g
l
(
X
I
x
:

OSXIS3;

xt

(
4
.
9
)
(
4
.
1
0
)
(
4
.
1
1
)

(
4
.
3
7
)

F
i
g
u
r
e
4
.
8i
l
Iu
s
t
r
a
t
e
st
h
e
g
r
a
p
h
i
c
a
l
s
o
l
u
t
i
o
n
.T
h
e
s
o
l
u
t
i
o
ni
s
a
tt
h
ei
n
t
e
r
s
e
c
t
i
o
no
f
t
h
e
d
o
t
t
e
dl
i
n
e
s
.Thecodei
sav
a
i
l
a
b
l
ei
nFlg4_8.m. l
eg
r
a
d
i
e
n
t
sa
r
edrawnu
s
i
n
gt
h
e
p
l
o
t
e
d
i
tf
u
n
c
t
i
o
n
sont
h
emenubari
nt
h
ewindow.A
l
s
of
o
n
ts
t
y
l
eandf
o
n
ts
i
z
eh
a
v
e
l
o
t
e
d
i
tcommands.Theo
p
t
i
m
a
ls
o
l
u
t
i
o
nmustl
i
eono
rt
oE
b
e
c
na
d
j
u
s
t
e
du
s
i
n
gt
h
ep
l
e
f
to
ft
h
ed
a
s
h
e
dl
i
n
e
.S
i
m
i
l
a
r
l
yi
t must i
I
ebelow o
ron t
h
ed
a
s
h
e
dc
u
r
v
e
.
1s
od
e
c
r
e
a
s
et
h
eo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
nv
a
l
u
ea
smucha
s
S
i
m
u
l
t
a
n
e
o
u
s
l
yi
ts
h
o
u
l
da
p
o
s
s
i
b
l
e
.

oSXIS3;

0SX2S3

(
4
.
1
2
)

Th
eg
r
a
p
h
i
c
a
ls
o
l
u
t
i
o
no
fS
e
c
t
i
o
n4.
3
.
3s
u
g
g
e
s
t
st
h
a
tt
h
ei
n
e
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
tw
i
l
lb
e
a
c
t
i
v
ea
tt
h
es
o
l
u
t
i
o
n
.If i
st
h
ec
a
s
e
t
h
e
nF
i
g
u
r
e4
.
8w
i
l
lo
n
c
ea
g
a
i
nr
e
p
r
e
s
e
n
te
o
.Thi
sw
i
l
lb
ev
c
r
i
f
i
e
di
nt
h
cn
c
x
ts
c
cl
Io
n
.
g
r
a
p
h
i
c
a
ls
o
l
u
t
i
o
nt
ot
h
i
ss
e
c
t
i
o
ns

4
.
4 ANAL
YTICALCONDITIONS
4
.
3
.
4 E
q
u
a
l
i
t
yandl
n
e
q
u
a
l
l
t
yConstralnts
Example4
.
1i
sd
e
v
e
l
o
p
e
dw
i
t
hb
o
t
ht
y
p
e
sonc
o
n
s
t
r
a
i
n
t
sp
r
e
s
e
n.
tI
li
s p
r
o
d
u
c
e
dh
e
r
e
:

Ana
l
y
t
i
c
a
lc
o
n
d
i
t
i
o
n
sr
c
f
e
rt
o n e
s
s
a
r
yands
u
f
f
i
c
i
e
n
tc
o
n
d
i
t
i
o
n
s tw
i
l
lp
e
r
m
i
t
o
n
d
i
t
i
o
n
s
t
h
er
e
c
o
g
n
i
t
i
o
no
ft
h
es
o
l
u
t
i
o
nt
ot
h
eo
p
t
i
m
a
ld
e
s
i
g
np
r
o
b
l
c
m
.Thcc
d
e
v
e
l
o
p
e
dh
c
r
e empower t
h
en
u
m
e
r
i
c
a
lt
e
c
h
n
i
q
u
e
st
of
o
l
l
o
wl
a
t
e
r
. They a
r
e

176

NONL
lNEARPROGRAMMING

4
.
4 ANAlYTICALCONDIT10NS

i
n
t
r
o
d
u
c
e
di
n sames
e
q
u
e
n
c
ea
si
nep
r
e
v
i
o
u
ss
e
c
t
i
o
n
.l
n
s
t
e
a
doff
o
n
n
a
l
h
ec
o
n
d
i
t
i
o
n
sa
r
ee
s
t
a
b
l
i
s
h
e
dl
e
s
sf
o
n
n
a
l
l
y from t
h
e
m
a
t
h
e
m
a
t
i
c
a
ld
e
t
a
i
l
st
g
e
o
m
e
t
r
i
c
a
ld
e
s
c
r
i
p
t
i
o
no
f p
r
o
b
l
e
ma
n
d
l
o
rr
o
u
g
hi
n

i
t
i
v
er
e
a
s
o
n
i
n
g
.F
o
n
n
a
l
d
c
v
e
l
o
p
m
e
n
to
ft
h
ca
n
a
l
y
t
i
c
a
lc
o
n
d
i
t
i
o
n
sc
a
nb
ef
o
u
n
di
nR
e
f
e
r
c
n
c
e
s6
8
.F
o
r
e
s
t
a
b
l
i
s
h
i
n
gt
h
ec
o
n
d
i
t
i
o
n
si
ti
se
x
p
e
c
t
e
dt
h
a
tt
h
es
o
l
u
t
i
o
ni
si
nt
h
ei
n
t
e
r
i
o
ro
fe
f
e
a
s
i
b
l
r
e
g
i
o
n
a
n
dt
h
e
r
ei
so
n
l
yoneminimum.

X
2
)
:(
X
I- 1
)2+(X2- 1
)
2-XIX2

OSxIS3;

(
4
.
3
8
)
(
4
.
3
7
)

OSX2S3

F
i
g
u
r
c4
.
6p
r
o
v
i
d
e
dac
o
n
t
o
u
rp
l
o
to
ft
h
ep
r
o
b
l
e
m
.F
i
g
u
r
e4
.
9p
r
o
v
i
d
e
sa
ti
sameshp
l
o
t
.Thecommandst
oc
r
e
a
t
e
t
h
r
e
e
d
i
m
e
n
s
i
o
n
a
lp
l
o
to
ft
h
esamep
r
o
b
l
e
m
.I
F
i
g
u
r
e4
.
9a
r
ea
v
a
i
l
a
b
l
ei
nFlg4_9.m.Thef
i
g
u
r
en
e
e
d
st
ob
er
o
t
a
t
e
di
n
t
e
m
c
t
i
v
c
l
yt
o
a
p
p
e
a
ra
si
l
l
u
s
t
e
d
.At
a
n
g
e
n
tp
l
a
n
ei
sdmwn minimumf
o
re
m
p
h
a
s
i
s
.F
i
g
u
r
e
4
.
9wi
l
1beu
s
dt
oi
d
e
n
t
i
f
yt
h
ep
r
o
p
e
e
so
f
t
h
ef
u
n
c
t
i
o
nJ
(
x
l
x
z
)a
tt
h
eminimum.
T)
Theminimumi
si
d
e
n
t
i
t
i
e
dbyas
u
p
e
c
r
i
p
ta
s
t
e
r
i
s
k(X
o
rl
x
i
X
2
J
t
u
d
y
i
n
g
.S
F
i
g
u
r
e4
.
9
t
h
ef
u
n
c
t
i
o
nh
a
saminimumv
a
l
u
eo
f-2
w
h
i
l
ex
j 2andx
22
.I
fe
v
a
l
u
e
so
fX1a
n
d
l
o
rX2we
t
oc
h
a
n
g
ee
v
e
nbyas
l
i
g
h
tamountfrom o
p
t
i
m
a
lv
a
l
u
e
i
na
n
yd
i
r
e
c
t
i
o
n
t
h
ev
a
l
u
eo
f
t
h
ef
u
n
c
t
i
o
nwi
l
1c
e
r
t
a
i
n
l
yi
n
c
r
e
a
s
es
i
n
c
eX
'i
st
h
el
o
w
e
s
t
s
e
n
t
i
n
g e f
u
n
c
t
i
o
nf R
e
p
r
c
s
e
n
t
i
n
g E
p
o
i
n
to
ft
h
ec
o
n
c
a
v
es
u
r
f
a
c
er
e
p
d
i
s
p
l
a
c
e
m
c
n
tf
r
o
mt
h
eoptimumv
a
l
u
e
so
ft
h
ev
a
r
i
a
b
l
e
sa
sd
X
andt
h
cc
h
a
n
g
ci
nt
h
e
f
fromd
i
r
e
c
to
b
s
e
r
v
a
t
i
o
ni
ti
sc
l
e
a
rt
h
a
tt
h
e
f
u
n
c
t
i
o
nv
a
l
u
cf
r
o
mt
h
eoptimuma
sl
o
p
t
i
m
a
ls
o
l
u
t
i
o
nmustb
eap
o
i
n
ta
ts
a
t
i
s
f
i
e
s

f
o
ral
1dX

a
p
p
r
o
x
i
m
a
t
e
dbyap
l
a
n
et
a
n
g
e
n
tt
ot
h
ef
u
n
c
t
i
o
na
tt
h
es
o
l
u
t
i
o
n(showni
nF
i
r
e
4
.
9
)
(
c
o
n
s
i
d
c
ri
s af
i
r
s
t
o
r
d
r
T
a
y
l
o
rs
e
r
i
e
se
x
p
a
n
s
i
o
n
)
.Movingt
oa
n
yp
o
i
n
ti
nt
h
e
p
l
a
n
efromt
h
eoptimum(
s
e
eF
i
g
u
r
e4
.
9
)w
i
l
ln
o
tc
h
a
n
g
et
h
ev
a
l
u
co
ft
h
ef
u
n
c
t
i
o
n
E f
o
r
edf O
.Movingawayf
r
o
moptimumi
m
p
l
i
e
s dx1a
n
ddx
r
en
o
tz
e
r
o
.
2a
R
e
w
r
i
t
i
n
gEq
u
a
t
i
o
n(
4
.
2
0
)i
nt
c
n
n
so
fX1a
n
dX2

]
[
:
J
=
O

df=

1+

-2
4

X
2

F
1g
u
r
e4
.
9 Thr
ee
d
i
m
e
n
s
i
o
n
a
lp
l
o
t0
1F
i
g
u
r
e4
.
6
.

l
1p
o
i
n
t
si
nep
l
a
n
e
dx11
:
-0a
n
ddx
:
-O
.Th
e
r
e
f
o
r
e
S
i
n
c
et
h
i
ss
h
o
u
l
dh
o
l
df
o
ra
21
df-fl

dX - v

-fl
dX2- v

o
rt
h
eg
r
a
d
i
e
n
tof
fa
tt
h
eoptimumm
u
s
tbez
e
r
o
. a
ti
s

F
l
r
s
t
O
r
d
e
rC
o
n
d
l
t
l
o
n
s
: Thesamei
d
c
ac
a
nb
ca
p
p
l
i
e
di
nt
h
el
i
m
i
t
t
h
a
ti
sf
o
r
i
n
f
i
n
i
t
e
s
i
m
a
ld
i
s
p
l
a
c
e
m
e
n
t dx1 a
n
d dx
a
b
o
u
t
X
T
h
e
f
u
n
c
t
i
o
n
i
t
s
e
l
f
c
a
n
b
e
2

df=

H FHX

x
.

4
f
>0

Thep
r
o
b
l
e
mu
s
e
df
o
ri
l
l
u
s
t
r
a
t
i
o
ni
s

J
(

1
0

4
.
4
.
1 UnconstrainedProblem

M
i
n
i
m
i
z
e

1
n

=0

(
4.
42
)

Vf(x~x;)

=0

(
4
.
4
3
)

Eq
u
a
t
i
o
n(
4
.
4
3
)e
x
p
r
e
s
s
e
st
h
en
e
c
e
s
s
a
r
yc
o
n
d
i
t
i
o
n

o
rf
i
r
s
t

o
r
d
e
rc
o
n
d
i
t
i
o
n
s(F
OC)
f
o
ru
n
c
o
n
s
t
r
a
i
n
e
do
p
t
i
m
i
z
a
t
i
o
n
.I
ti
st
e
r
m
e
dt
h
ef
r
r
s
t

o
r
d
e
rc
o
n
d
i
t
i
o
nb
e
c
a
u
s
eE
q
u
a
t
i
o
n
r
a
d
i
e
n
to
rt
h
ef
i
r
s
td
e
r
i
v
a
t
i
v
e
.E
q
u
a
t
i
o
n(
4.
4
3
)i
su
s
e
dt
oi
d
e
n
t
i
f
yt
h
c
(
4
.
4
3
)u
s
c
seg
p
o
s
s
i
b
l
es
o
l
u
t
i
o
n
st
ot
h
eo
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
e
m
.If f
u
n
c
t
i
o
nw
e
r
et
of
l
i
po
v
e
rs
oa
t
t
h
esamef
u
n
c
t
i
o
nwc
t
omaximizet
h
ev
a
l
u
eo
fef
u
n
c
t
i
o
nt
h
es
o
l
u
t
i
o
nf
o
rt
h
e
v
a
r
i
a
b
l
sw
i
l
lb
ea
tt
h
esamev
a
l
u
eo
ft
h
ed
e
s
i
g
nv
a
r
i
a
b
l
c
.I
ti
sc
l
e
a
rt
h
a
tEq
u
a
t
i
o
n
.
4
3
)appliesthemaximump
r
o
b
l
e
ma
l
s
o
.E
q
u
a
t
i
o
n(
4.
4
3
)b
yi
t
s
e
l
fw
i
l
ln
o
t
(
4
n
e
c
e
s
s
a
r
y
d
e
t
e
r
m
i
n
et
h
eminimumv
a
l
u
eo
ft
h
ef
u
n
c
t
i
o
n
.A
d
d
i
t
i
o
n
a
lc
o
n
s
i
d
e
r
a
t
i
o
n
sa
t
oe
n
s
u
r
ea
tt
h
es
o
l
u
t
i
o
ne
s
t
a
b
l
i
s
h
e
dby f
i
r
s
t
o
r
d
e
rc
o
n
d
i
t
i
o
n
si
so
p
t
i
m
a
li
ni
s
c
a
s
caminimum.F
o
rag
e
n
e
r
a
lu
n
c
o
n
s
t
r
a
i
n
e
dp
r
o
b
l
c
m
t
h
en
e
c
e
s
s
a
r
yc
o
n
d
i
t
i
o
n
so
f
Eq
u
a
t
i
o
n(
4.
4
3
)c
a
nb
es
t
a
t
e
da
s

4.
4

178 NONL
lN
EARPROGRAMMING

Vf(
X
')
=0

af=jaTH

(
4.
46
)

SDC:Fori
sp
r
o
b
l
e
m[
s
e
eE
q
u
a
t
i
o
n(
4
.
2
6
)
]

blr

T
h
i
si
ss
i
r
n
i
l
a
rt
oE
q
u
a
t
i
o
n(
4.
35
)e
x
c
e
p
tee
x
p
a
n
s
i
o
ni
sa
b
o
u
tcs
o
l
u
t
i
o
n
.6
f
mustb
eg
r
c
a
t
e
rt
h
a
nz
e
r
o
.Employingt
h
en
e
c
e
s
s
a
r
yc
o
n
d
i
t
i
o
n
s(
4.
4
4
)
t
h
ef
i
r
s
tt
e
r
m
4
5i
sz
e
r
o
.T
h
i
sI
c
a
v
c
sef
o
l
l
o
w
i
n
gi
n
e
q
u
a
l
i
t
y
ont
h
er
i
g
h
t
h
a
n
ds
i
d
eo
fE
q
u
a
t
i
o
n4.

4
7
b
)r
e
p
r
e
s
e
n
tap
a
i
ro
f
l
i
n
e
a
r
e
q
u
a
t
i
o
n
sw
h
i
c
hc
a
nbes
o
l
v
e
d
E
Q
u
a
t
i
o
n
s(
4
.
4
7
a
)a
n
d(
4.
I2=2.Thev
a
l
u
eoft
h
et
u
n
c
t
i
o
nfis-2.Sof
a
ro
n
l
yt
h
en
e
c
e
s
s
a
r
y
as-X1=2
c
o
n
d
i
t
i
o
n
sh
a
v
eb
e
e
ns
a
t
i
s
f
i
e
d
.Thev
a
l
u
e
sa
b
o
v
ec
a
na
l
s
or
e
f
e
rt
oap
o
i
n
twheret
h
e
n
c
t
i
o
ni
samaximum

o
rwheret
h
e
r
ei
sas
a
d
d
l
ep
o
i
n
.
t

(
4
.
4
5
)

(
4.
4
7
b
)

=2
-1)-x1=

Second-DrderC
o
n
d
i
t
i
o
n
s
:The s
e
c
o
n
d
o
r
d
e
rc
o
n
d
i
t
i
o
n
s (SOC) a
r
eu
su
a
l
1
y
u
f
f
i
c
i
e
n
tc
o
n
d
i
t
i
o
n
s
.I
tc
a
nb
ei
n
f
e
'
e
d a
tt
h
e
s
ec
o
n
d
i
t
i
o
n
sw
i
l
li
n
v
o
l
v
e
r
e
g
a
r
d
e
da
ss
s
e
c
o
n
dd
e
r
i
v
a
t
i
v
e
so
ft
h
f
u
n
c
t
i
o
n
.TheSOCi
so
f
t
e
no
b
t
a
i
n
e
dt
h
r
o
u
g
ht
h
eT
a
y
l
o
r
e
x
p
a
n
s
i
o
no
ft
h
ef
u
n
c
t
i
o
nt
os
c
c
o
n
do
r
d
e
r
.I
fX i
s s
o
l
u
t
i
o
n
a
n
dd
.
Xr
e
p
r
e
s
e
n
t
se
changeof v
a
r
i
a
b
l
c
sfrom o
p
t
i
m
a
lv
a
l
u
ewhichw
i
l
ly
i
e
l
dac
h
a
n
g
e
f
i
dt
h
e
n
+6X))=(X
+ t a h r

(
4
.
4
4
)

whichi
sp
r
i
m
a
r
i
l
yav
e
c
t
o
re
x
p
r
e
s
s
i
o
no
ft
h
er
e
l
a
t
i
o
ni
nE
q
u
a
t
i
o
n(
4.
4
3
)
.E
q
u
a
t
i
o
n
(
4
.
4
4
)i
su
s
e
dt
oe
s
t
a
b
l
i
s
ht
h
ev
a
l
u
eo
ft
h
ed
e
s
i
g
nv
a
r
i
a
b
l
e
sX'b
o
t
ha
n
a
l
y
t
i
c
a
l
l
ya
n
d
n
u
m
e
r
i
c
a
l
l
y
.

A
N
A
L
Y
T
I
C
A
LC
O
N
D
I
T
I
O
N
S 1
7
9

I
si
tp
o
s
i
t
i
v
ed
e
f
i
n
i
t
e
?

.
X
(
i
) Notp
o
s
s
i
b
l
e1
0t
e
s
ta
l
ld
(
i
i
)TocaJcula~ the e
i
g
e
n
v
a

u
e
so
fH

1=
~I=\
=
-1 2-
l

'
h
ee
i
g
e
n
v
a
l
u
e
sa
r
=

=3andt
h
em
a
t
r
i
xi
sp
o
s
i
t
i
v
ed
e
f
i
n
i
t
e
.

whereH(X

)i
st
h
eH
e
s
s
i
a
nm
a
t
r
i
x(
t
h
ema
i
xo
fs
e
c
o
n
dd
e
r
i
v
a
t
i
v
e
s
)o
ft
h
ef
u
n
c
t
i
o
n

fa
tt
h
cp
o
s
s
i
b
l
o
ptimumv
a
l
u
eXF
o
rer
e
l
a
t
i
o
n
si
nE
q
u
a
t
i
o
n(
4
.
4
6
)t
oh
o
l
d
t
h
e

(
u
i
)Toc
a
l
c
u

a
t
ed
e
t
e
r
r
n
i
n
a
n
t
so
fa
l
lo
r
d
e
r
a
li
n
c
Ju
det
h
emaind
i
a
g
o
n
a
la
n
d
i
n
c
J
udet
h
ee
l
e
m
e
n
ti
n ef
r
r
s
trowandf
i
r
s
tcolumn

1
2
1
>
0

qJ

nu

(
i
i
)Thee
i
g
e
n
v
a
l
u
e
so
fH(X')a
r
ea
l
lp
o
s
i
t
i
v
c
.

'a

T
H(X')
6
X
d
.
X
.
6
X>O
.
(
i
) F
o
ra
l
lp
o
s
s
i
b
l
e.

mustbep
o
s
i
t
i
v
ed
e
f
i
n
i
t
e
.T
h
e
r
ea
r
et
h
r
e
ewayst
os
t
a
b
l
i
s
ht
h
eH i
s
m
a
t
r
i
xH(X
p
o
s
i
t
i
v
ed
e
f
i
n
i
t
e
.

(
i
u
)Thed
e
t
e
r
m
i
n
a
n
t
so
f]l
o
w
e
ro
r
d
e
r
s(
s
u
b
m
a
t
r
i
c
e
s
)o
fH(
X
'
)t
h
a
ti
n
c
l
u
d
et
h

maind
i
a
g
o
n
a
la
r
ea
I
lp
o
s
i
t
i
v
e
.

Them
a
t
r
i
xi
sp
o
s
i
t
i
v
ed
e
f
i
n
i
t
e
.

e
e

o
n
l
y(
i
i
)a
nd(
i
i
i
)c
a
nb
ep
r
a
c
t
i
c
a
l
l
ya
p
p
l
i
e
d
whichi
si
l
l
u
s
t
r
a
t
e
db
e
l
o
w
.
Ofther

Sec_
4_
4-1.mp
r
o
v
i
d
e
sec
o
n
f
i
r
m
a
t
i
o
no
ft
h
en
u
m
e
r
i
c
a
lv
a
l
u
e
sf
o
rt
h
i
se
x
a
m
p
l
e
u
s
i
n
gS
y
m
b
o
l
i
cToolboxa
n
db
a
s
i
cMATLABcommands.

Example
4
.
4
.
2 EqualityConstrainedProblem
M
i
n
i
r
n
i
z
e

j
(

x
2
)
:(
X
I-1
)
2+(
X
2-1
)
2-X
IX
2
XI

(
4.
38
)

Th
cp
r
o
b
l
e
m
:
O~XI ~3;

O S; x2~3

(
4.
37
)

FDC

=2(xl-2)-x2=O

(
4.
4
7
a
)

M
i
n
i
r
n
i
z
e

f
(
X
I
x
2
)
:X
IX
2

(
4
.
36
)

Suect

hl(xl x2):xt/4+~-1

(
4.
39
)

O~xl~3;

(
4.
37
)

O!:>x2~3

180 NONL
lN
E
A
RPROGRAMMING

4
.
4

F
ig
u
r
e4
.
7i
l
l
u
s
t
r
a
t
e
deg
r
a
p
h
icaJs
o
l
u
t
i
o
n
.I
tw n
o
t
i
c
e
da
ta
tt
h
es
o
l
u
t
i
o
nt
h
e
g
r
a
d
i
e
n
to
f o
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
n dt
h
cg
r
a
d
i
e
n
to
ft
h
ec
o
n
s
t
r
a
i
n
twe
p
a
r
a
l
l
e
la
n
d
o
p
p
o
s
i
t
e
l
yd
i
r
e
c
t
e
d
.
E
x
a
m
i
n
i
n
go
t
h
e
rf
e
a
s
i
b
l
ep
o
i
n
t
si
nFistIm4.7(ont
h
ed
a
s
h
e
d
c
u
r
v
e
)i
tc
a
nb
ea
s
c
e
r
t
n
e
dt
h
a
tt
h
es
p
e
c
i
a
lg
e
o
m
e
t
r
i
c
a
lr
e
l
a
t
i
o
n
s
h
i
pi
so
n
l
yp
o
s
s
i
b
l
c
a
tt
h
es
o
l
u
t
i
o
n
.Att
h
cs
o
l
u
t
i
o
nap
r
o
p
o
r
t
i
o
n
a
lr
e
l
a
t
i
o
n
s
h
i
pc
x
i
s
t
sb
e
t
w
e
e
nt
h
eg
r
a
d
i
e
n
t
s
a
tt
h
es
o
l
u
t
i
o
n
.U
s
i
n
gt
h
cc
o
n
s
t
a
n
to
fp
r
o
p
o
r
t
i
o
n
a
l
i
t
yA
I(
ap
o
s
i
t
i
v
ev
a
l
u
e
)t
h
c
r
e
l
a
t
i
o
n
s
h
i
pb
e
t
w
e
e
nt
h
eg
r
a
d
i
c
n
t
sc
a
nbee
x
p
r
e
s
s
e
da
s

=IVhl or Vf+ =0

(
4
.
4
8
)

E
q
u
a
t
i
o
n(
4
.
4
8
)i
su
s
u
a
l
l
yo
b
t
a
i
n
e
di
namoref
o
r
m
a
lwayu
s
i
n
gt
h
cmethodo
f
L
a
g
r
a
n
g
em
u
l
t
i
p
l
i
e
r
s
.

t
h
ep
r
o
b
l
e
mi
st
r
a
n
s
f
o
r
m
e
dbyi
n
t
r
o
d
u
c
i
n
g
MethodofLagrange: I
nl
h
i
smethod
a
naugmentedf
u
n
c
t
i
o
n
c
a
l
l
e
dt
h
eLa
g
r
a
n
g
i
a
n
.a
st
h
eo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
ns
u
b
j
e
c
tt
ot
h
e
same u
a
l
i
t
yc
o
n
s
n
t
s
.TheLa
g
r
a
n
g
i
a
ni
sd
e
f
i
n
e
da
st
h
esumo
ft
h
eo
r
i
g
i
n
a
l
i
n
e
a
rc
o
m
b
i
n
a
t
i
o
no
ft
h
ec
o
n
s
t
r
a
i
n
t
s
. l
ec
o
c
f
f
i
c
i
c
n
t
so
fl
h
i
s
o
b
j
e
c
t
i
v
cf
u
n
c
t
i
o
na
n
dal
l
i
n
c
a
rc
o
m
b
i
n
a
t
i
o
n known La
g
r
a
n
g
em
u
l
t
l
i
e
r
s
.Withr
c
f
c
r
e
n
c
e1
0t
h
c
ss
c
c
t
i
o
n
e
x
a
m
p
l
ci
ni
Minimize

F
(
x
2
A
I
)=
f
(
x
l
x
u+A
Ihl
(
X
I
x
2
)

(
4.
4
9
)

A
N
A
L
Y
T
I
C
A
LC
O
N
O
I
T
I
O
N
S 1
8
1

Eq
u
a
t
i
o
n
s(
4.
51
)e
x
p
r
e
s
seFOCorn
S
yc
o
n
d
i
t
i
o
n
sf
o
r q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
e
d
p
r
o
b
l
e
mi
ntwov
a
r
i
a
b
l
e
s
.Th
cl
a
s
te
q
u
a
t
i
o
ni
nt
h
ca
b
o
v
es
e
ti
st
h
eC
O
n
s
i
n
te
q
u
a
t
i
o
n
.
i
se
n
s
u
r
e
st
h
es
o
l
u
t
i
o
ni
sf
e
a
s
i
b
l
e
.Th
ef
i
r
s
ttwoe
q
u
a
t
i
o
n
sc
a
nb
ea
s
s
e
m
b
l
e
di
n

v
c
c
t
o
rf
o
r
mt
oy
i
c
l
dt
h
esamei
n
f
o
r
m
a
t
i
o
ne
x
p
r
e
s
s
e
dbyE
q
u
a
t
i
o
n(
4.
48
)w
h
i
c
hwas
o
b
n
e
dg
r
a
p
h
i
c
a
l
l
y
.l
eI
e
f
t
h
a
n
dexp
s
s
i
o
n
si
nef
i
r
s
tlwoe
q
u
a
t
i
o
n
sa
b
o
v
ca
r
e
t
h
eg
r
a
d
i
e
n
to
ft
h
eL
a
g
r
a
n
g
i
a
nf
u
n
c
t
i
o
n
:
VF= +IVh.=0
A
p
p
l
y
i
n
gEq
u
a
t
i
o
n
s(
4
.
5
1
)t
ot
h
cexampleo
fi
ss
e
c
t
i
o
n
:
aFIX
1
? =-x.+ =0
oX1

aF
=
x
.

x
.=o
-"
"
2+
'2
"
'
"
"

ol
2

"
.
:

_
_

(
4
.
5
2
)

VA2

h144-l=o

bec
o
m
p
l
e
t
ep
r
o
b
l
e
mi
sd
e
v
e
l
o
p
e
da
s

Equ
a
t
i
o
n
s(
4
.
5
2
) p
r
e
s
e
n
t e
q
u
a
t
i
o
n
si
n v
a
r
i
a
b
l
e
swhichs
h
o
u
l
dd
e
t
e
r
m
i
n
e
Minimize

F
(
X
I
x
2
A
I
)=-XIX2+
1
(
.
/4+4-1)

(
4
.
5
0
)

S
u
b
j
e
c
lt
o
: hl
(
X
I
x
2
)
:4/4+4-1

(
4.
39
)

O
:
:
;
;
X
I:
:
;
;
3
;

(
4
.
3
7
)

0
:
:
;
;
x
2
:
:
;
;
3

I
ses
o
l
u
t
i
o
nt
ot
h
e l
s
f
o
r
m
e
dp
r
o
b
l
c
m[
E
qu
a
l
i
o
n
s(
4
.
5
0
)
(
4
.
3
9
)
(
4
.
3
7
)
]t
h
esame
a
st
h
es
o
l
u
t
i
o
nt
ot
h
eo
r
i
g
i
n
a
lp
r
o
b
l
e
m[
E
q
u
a
t
i
o
n
s(436)

(
4
.
3
9
)

(
4
3
7
)
]
?
I
f
t
h
e
d
e
s
i
g
n
i
si
f
s
i
b
l
e

enmostd
e
f
i
n
i
t
e
l
yy
e
s
.F
o
rf
e
a
s
i
b
l
ed
e
s
i
g
n
s
h
l
(
x
"X2)=0
d
ot
i
v
ef
u
n
c
t
i
o
n
si
nEq
u
a
t
i
o
n
s(
4
.
3
6
)and(
4
.
5
0
)a
r
et
h
es
a
m
e
.I
fd
e
s
i
g
ni
sn
o
l
t
h
c
nbyd
e
f
i
n
i
t
i
o
nt
h
e
r
ei
snos
o
l
u
t
i
o
na
n
y
w
a
y
.
f
e
a
s
i
b
l
e
1
)a
s u
n
c
o
n
s
t
r
a
i
n
e
d
m
c
t
i
o
l
li
n
beFOCa
r
eo
b
t
a
i
n
e
dbyc
o
n
s
i
d
e
r
i
n
gF(x"X2
ev
a
r
i
a
b
l
c
sX
I
X2

l
'Th
i
sp
r
o
v
i
d
e
sr
e
er
e
l
a
t
i
o
n
st
os
o
l
v
ef
o
rx
j

x
i

dF i
)
f ah
=+.-!.=o
)
xl '.'. ax
ax i

aF

af." ahl

=0
= +
a
A
2 ax2" '
X
.a
2
aF

=h.=O

(
4
.
5
1
)

t
h
ev
a
l
u
e
sf
o
rx
j
.x
i

j
. Note E
q
u
a
t
i
o
n
s(
4
.
5
2
)o
n
l
yd
e
f
i
n
et
h
en
e
c
e
s
s
a
r
y
c
o
n
d
i
t
i
o
n
s
whichmeanst
h
es
o
l
u
t
i
o
nc
o
u
l
dbeamaximuma
l
s
o
.
Th
c
r
ei
sonep
r
o
b
l
e
mw
i
t
hr
e
s
p
e
c
tt
ot
h
es
e
to
fE
q
u
a
t
i
o
n
s(
4
.
5
2
)
.Th
e
q
u
a
t
i
o
n
sa
r
e
.
tMostp
r
e
r
e
q
u
i
s
i
t
ec
o
u
r
s
e
sonn
u
m
c
r
i
c
a
lm
e
t
h
o
d
sdon
o
ta
t
t
e
m
p
tt
o
an
o
n
l
i
n
c
a
rs
e
s
o
l
v
an
o
n
l
i
n
e
a
rs
y
s
t
e
mo
fe
q
u
a
t
i
o
n
s
.U
s
u
a
l
l
y
.l
h
c
yo
n
l
yh
a
n
d
l
eas
i
n
g
l
eonet
h
r
o
u
g
h
Newton-Raphsono
rt
h
eb
i
s
c
c
t
i
o
nm
e
t
h
o
d
.I
nf
a
c
t
NLPo
rd
e
s
i
g
no
p
t
i
m
i
z
a
t
i
o
ni
s
a
t
i
o
n
s
a
l
b
e
i
to
f
p
r
i
m
a
r
i
l
ya
b
o
u
tt
c
c
h
n
i
q
u
e
sf
o
rs
o
l
v
i
n
gas
y
s
t
e
mo
fn
o
n
l
i
n
e
a
r u
i
smeansE
q
u
a
t
i
o
n
s(
4
.
5
2
)c
a
n
n
o
tb
es
o
l
v
c
du
n
t
i
lwea
d
v
a
n
c
cr
t
h
e
r
s
p

i
f
i
cf
o
r
m
s
.Th
sc
o
u
r
s
e
.F
o
r
t
u
n
a
t
e
l
y
h
a
v
i
n
gat
o
o
lI
i
k
eMA
A
Bo
b
v
i
a
t
e
st
h
i
sd
i
f
f
i
c
u
l
t
y
a
n
d
i
ni
n
gj
u
s
t
i
f
i
c
a
t
i
o
no
fu
s
i
n
gan
u
m
e
r
i
c
a
l
/
s
y
m
b
o
l
i
ct
o
o
lf
o
rs
u
p
p
o
r
t
i
n
g
p
r
o
v
i
d
e
sas
d
e
v
e
l
o
p
m
e
n
to
ft
h
ec
o
u
r
s
e
.2

MA
A
BCode: I
nMATLAB.t
h
e
r
ea
r
etwowayso
fs
o
l
v
i
n
gE
q
u
a
t
i
o
n
s(
4
.
5
2
)
u
s
i
n
g
s
y
m
b
o
l
i
cs
u
p
p
o
r
tf
u
n
c
t
i
o
n
so
ru
s
i
n
gt
h
cn
u
m
e
r
i
c
a
ls
u
p
p
o
r
tf
u
n
c
t
i
o
n
s
.Bothp
r
o
c
e
d
u
r
c
s
es
y
m
b
o
l
i
cf
u
n
c
t
i
o
ni
s
h
a
v
cl
i
m
i
t
a
t
i
o
n
swhena
p
p
l
i
c
dt
oh
i
g
h
l
yn
o
n
l
i
n
e
a
rf
u
n
c
t
i
o
n
s
.Th
s
o
l
v
e
a
n
d
n
u
m
e
r
i
c
a
lf
u
n
c
t
i
o
ni
s
f
s
o
l
v
e
.Then
u
m
e
r
i
c
a
lt
c
c
h
n
i
q
u
ei
sa
ni
t
e
r
a
t
i
v
eo
n
e
a
n
dr
e
q
u
i
r
e
syout
oc
h
o
o
s
ea
ni
n
i
t
i
a
lg
u
c
s
st
os
t
a
r
tt
h
ep
r
o
c
e
d
u
r
e
.Q
u
i
t
eo
f
t
e
ns
e
v
c
r
a
l
d
i
f
f
e
r
e
n
tg
u
e
s
s
e
smayb
c q
u
i
r
e
dt
of
i
n
dcs
o
l
u
t
i
o
n
.S
o
l
u
t
i
o
n
so
fs
y
s
t
c
m
so
f
e
q
u
a
t
i
o
n
sa
r
ef
u
n
d
a
m
e
n
t
a
lt
ot
h
er
e
s
to
feb
o
o
k
.Thef
o
l
l
o
w
i
n
gc
o
d
ei
si
n
c
l
u
d
e
d
ah
a
n
d
so
ne
x
e
r
c
i
s
e
.I
ti
sa
va
i
I
a
b
l
ea
sSec4_
4_2.m.I
tr
e
q
u
i
r
e
seqns4_
4_2.mt
o
l
v
ecommand.Thcco
o
fef
i
l
c
sa
r
el
i
s
t
c
db
c
l
o
w
e
x
e
c
u
t
e
o
1
1
1
C
n
r
co
l
h
e
r1
0
0
1
8b
e
s
l
d
c
sM
A
T
L
A
Ds
u
c
h M
U
l
h
c
n
d I
d c
c
l
.

1
8
2 N
O
N
l
l
N
E
A
RP
R
O
G
R
A
M
M
I
N
G

4
.
4A
N
A
L
Y
T
I
C
A
LC
O
N
D
I
T
I
O
N
S 1
8
3

S 4_4_2.m
N
ecessary/Sufficient conditions for
E
qualityconstrained problem

d
efine initial values
xinit=[l 1 0.5]'; initial guess for x1x21am1

the equations to be solved are available in


eqns4_4_2.m

Optimizationwith MATLABSection 4.4.2


% Dr. P.Venkataraman

xfina1 ~ fso1ve('eqns4_4_2'
xinit);

Minimize f(X1
x2)

-x1

fprintf('Thenumericalsolution (x1*
x2*
lam1*):n')
disp(xfina1);

-------

eqn
s
4
_
4_2.m
function ret eqns4_4_2(x)
x is a vector
x
(l) x1x(2)
x2x(3)
lam1
ret=[(-x(2) + 0.5*x(1)*x(3))
(-x(l) + 2*x(2) x(3))
(0.25*x(1)*x(1) + x(2)*x(2) -1)];

symbolic procedure

--"--'

define symbolic variables


format compact .
syms x1 x2 lam1 h1 F
d
efine F
);
F -x1*x2 + lam1*(X1*X1/4 + x2*x2 1
h1 x1*x1/4 +x2*x2 1
;

OutputI
nM
A
T
L
A
BCommandWlndow

he g
radient of F
grad1 ~ diff(F
X1);
grad2 = diff(F
X2);

optimalvalues
sa
isfaction of necessary conditions
grad2
h1'X1
x2
lam1');
[lams1 xs1 xs2]
sOlve(grad1

the solution is returned as a vector of


the three unknowns in case of multiple solutions
l
ams1 is the solution vector for lam1 etc.
;mORmT:the results are sorted habeti ly

fprintf is used to print a string in the


commandwindow
disp is used to print values of matrix

f -xs1.*xs2;
x2
lam

*
f*)n')
fprintf('The solution (x1*
disp(double([xs1 xs2 lams1 f]))

----------

% Numerical procedure

x2*
lam1* f*):
The solution (x1*
1.4142
0.7071
1.0000 -1.0000
-1.4142 -0.7071
1.0000 -1.0000
-1.0000
1.0000
-1.4142
0.7071
1.4142
-0.7071
-1.0000
1.0000
Optimization terminated successful1y:
Relative function va1ue changing by 1ess than
OPTIONS.TolFun
The numerica1 solution (x1*
x2*
lam

*
):
1.4141
0.7071
0.9997
Th
es
y
m
b
o
l
i
cc
o
m
p
u
t
a
t
i
o
ng
e
n
e
r
a
t
e
sf
o
u
rs
o
l
u
t
i
o
n
s
.Onlyt
h
ef
i
r
s
to
n
ei
sv
a
l
i
df
o
rt
h
i
s
p
r
o
b
l
e
m
.T
h
i
si
sd
e
c
i
d
e
db
yes
i
d
ec
o
n
s
t
r
a
i
n
t
se
x
p
r
e
s
s
e
dbyEq
u
a
t
i
o
n(
4.
37
)
.T
h
i
si
s

i
I
Iu
s
t
r
a
t
i
o
no
ft
h
eap
r
i
o
r
imannerbyw
h
i
c
ht
h
es
i
d
ec
o
n
s
t
r
a
i
n
t
sa
f
f
e
c
tt
h
ep
r
o
b
l
e
m
.
Ont
h
eo
t
h
e
rh
a
n
d
.en
u
m
e
r
i
c
a
lt
e
c
h
n
i
q
u
e
sp
r
o
v
i
d
eo
n
l
yo
n
es
o
l
u
t
i
o
n1
0t
h
ep
r
o
b
l
e
m
.
lIsi
saf
u
n
c
t
i
o
no
ft
h
ei
n
i
t
i
a
lg
u
e
s
s
.G
e
n
e
r
a
l
l
y
.n
u
m
e
r
i
c
a
lt
h
n
i
q
u
e
sw
i
l
ld
e
l
i
v
e
r
s
o
l
u
t
i
o
n
sc
l
o
s
e
s
tt
ol
h
ep
o
i
n
tt
h
e
ys
t
a
r
t
e
df
r
o
m
.Thes
o
l
u
t
i
o
ni
s

---

solution to non-linear system using fsolve


see help fsolve

the unknowns have to be defined as a vector


the functions have to be set up in an m-file

x;=1
.
4
1
4
1
;x
;
=
0
.
7
0
7
1
;=1
.0

'
h
es
o
h
o
n
sf
o
rX
I dX2C

bev
e
r
i
f
i
e
df
r
o
mt
h
eg
r
a
p
h
i
c
a
ls
o
l
u
t
i
o
ni
nF
i
g
u
r
e4
.
7
.
Th
es
o
l
u
t
i
o
n
sc
a
na
l
s
ob
ev
e
r
i
f
i
e
dt
h
r
o
u
g
ht
h
eh
a
n
d
h
e
l
dc
a
l
c
u
l
a
t
o
r
.Th
ev
a
l
u
ea
n
ds
i
g
n
a
r
eu
s
u
a
l
l
yi
m
m
a
t
e
r
i
a
lf
o
re
s
t
a
b
l
i
s
h
i
n
gt
h
eo
p
t
i
m
u
m
.F
o
raw
e
l
l
p
o
s
e
dp
r
o
b
l
e
m
f
o
r
i
ts
h
o
u
l
db
ep
o
s
i
t
i
v
es
i
n
c
ci
n
c
a
s
i
n
gt
h
ec
o
n
s
t
r
a
i
n
tv
a
l
u
ci
su
s
e
f
u
lo
n
l
yi
fwca
r
et
r
y
i
n
g
t
oi
d
e
n
t
i
f
yal
o
w
e
rv
a
l
u
ef
o
reminimum.Th
i
sf
o
r
c
e
sha
n
dft
omovci
no
p
p
o
s
i
t
e

4.
4

184 NONUNEARPROGRAMMING
d
i
r
e
c
t
i
o
n
s(
g
r
a
d
i
e
n
t
s
)
.I
n
c
r
e
a
s
i
n
gc
o
n
s
t
r
a
i
n
tv
a
l
u
ec
a
nbea
s
s
o
c
i
a
t
e
dw
i
t
he
n
l
a
r
g
i
n
gt
h
e
f
e
a
s
i
b
l
edomainw
h
i
c
hmayy
i
e
l
dab
e
t
t
e
rd
e
s
i
g
n
.

LagrangeMult


l
l
e
r
s
:TheL
a
g
r
a
n
g
em
u
l
t
i
p
l
i
e
rmethodi
sa
nl
e
g
a
n
tf
o
r
m
u
l
a
t
i
o
n
noverview
i
tseemss
t
r
a
n
g
eatwe
t
oo
b
t
a
i
nt
h
es
o
l
u
t
i
o
nt
oac
o
n
s
t
r
a
i
n
e
dp
r
o
b
l
e
m
.I
h
a
v
et
oi
n
t
r
o
d
u
c
ea
na
d
d
i
t
i
o
n
a
lunknown(

1
)t
os
o
l
v
et
h
ec
o
n
s
t
r
a
i
n
e
dp
r
o
b
l
e
m
.T
h
i
s
v
i
o
l
a
t
e
sec
o
n
v
e
n
t
i
o
n
a
lr
u
l
ef
o
rNLP t
h
ef
e
w
e
rt
h
ev
a
r
i
a
b
l
e
s
t
h
eb
e
t
t
e
rt
h
e
c
h
a
n
c
e
s ofo
b
t
a
i
n
i
n
g es
o
l
u
t
i
o
n
. As i
n
d
i
c
a
t
e
di
nt
h
ed
i
s
c
u
s
s
i
o
ne
a
r
l
i
e
r
E
n
s
f
o
r
m
a
t
i
o
no
fac
o
n
s
t
r
a
i
n
e
dp
r
o
b
l
e
mi
n
t
oa
nu
n
c
o
n
s
t
r
a
i
n
e
d
L
a
g
r
a
n
g
i
a
na
l
l
o
w
st
h
ea
l
s
oh
a
sap
h
y
s
i
c
a
ls
i
g
n
i
f
i
c
a
n
c
e
.Att
h
es
o
l
u
t
i
o
ni
t
p
r
o
b
l
e
m
.TheL
a
g
r
a
n
g
em
u
l
t
i
p
l
ira
a
t
i
oo
f c
h
a
n
g
ei
nt
h
eo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
nt
ot
h
ec
h
a
n
g
ei
n
e
x
p
r
e
s
s
e
ser
c
o
n
s
t
r
a
i
n
tv
a
l
u
e
.Toi
l
l
u
s
t
r
a
t
ei
sc
o
n
s
i
d
e
r
:

(
4
.
8
)

1
0E
q
u
a
t
i
o
n(
4
.
5
3
)
t
h
r
e
ee
q
u
i
v
a
l
e
n
tr
e
p
r
e
s
e
n
t
a
t
i
o
n
sf
o
rt
h
eL
a
g
r
a
n
g
i
a
na
r
es
h
o
w
n
.The
FOCare

h"
a
J
E
i axizkh'

dF

(
4
.
5
4
)

=+) A
.
"~'.k =0; i=
12

(
4
)

S
u
b
j
e
c
t

: (
X
)
]=o

(
4
.
8
)

E
q
u
a
t
i
o
n
s(
4
.
5
4
)a
n
d(
4)p
r
o
v
i
d
et
h
e1
1+1
r
e
l
a
t
i
o
n
st
od
e
t
e
r
m
i
n
e n+1
unknowns

dF

(
4
.
5
4
)i
sa
l
s
oe
x
p
r
e
s
s
e
da
s

n
u

-hH

n
v
r

V
'
n

J
V

EE

Att
h
es
o
l
u
t
i
o
nt
h
eFOCdeemst
h
a
tdF=0(
w
h
i
c
hc
a
na
l
s
ob
es
e
e
ni
nt
h
ea
b
o
v
e
d
e
t
a
i
l
e
de
x
p
a
n
s
i
o
n
)
.Hence

(
4
.
5
5
)

6
1
1

Theabovedependencedosn
o
ta
f
f
e
c
tt
h
ee
s
t
a
b
l
i
s
h
m
e
n
to
ft
h
eo
p
t
i
m
a
ld
e
s
i
g
n
.I
t
d
o
e
s hav
a
ni
m
p
o
r
t
a
n
tr
o
l
ei
nt
h
ed
i
s
c
u
s
s
i
o
no
fd
e
s
i
g
ns
e
n
s
i
t
i
v
i
t
yi
nNLP
p
r
o
b
l
e
m
s
.

GeneralEqualltyC
o
n
s
t
r
a
l
n
e
c
lProblem: Rememberingn-1>0
M
i
n
i
m
i
z
e f(X)
J
'
I

(
4
.
5
)

(
4
.
6
)

Subject[
l
I
(
X
)
]=0
0
wSXSXup
X1

Second
O
r
d
e
rC
o
n
d
i
t
i
o
n
s
:At s
o
l
u
t
i
o
nd
e
t
e
r
m
i
n
e
dby FOCt
h
ef
u
n
c
t
i
o
n
X.Changesi
nX a
r
en
o
ta
r
b
i
t
r
a
r
y
.Theyh
a
v
et
o
s
h
o
u
l
di
n
c
r
e
a
s
ef
o
rc
h
a
n
g
e
si
n6
s
a
t
i
s
f
yel
i
n
e
a
r
i
z
e
de
q
u
a
l
i
t
yc
o
n
s
a
i
n
ta
tt
h
es
o
l
u
t
i
o
n
.1
ti
st
a
k
e
nf
o
rg
r
da
t
t
h
e
s
u
f
f
i
c
i
e
n
tc
o
n
d
i
t
i
o
n
sMeu
s
u
a
l
l
ya
p
p
l
i
e
di
nas
m
a
l
ln
e
i
g
h
b
o
r
h
o
o
do
ft
h
eo
p
t
i
m
u
m
.
changesa
r
ec
o
n
t
e
m
p
l
a
t
e
d
o
n
l
yw
i
t
hr
e
S
P
E
C
tt
oX andn
o
tw
i
t
hr
e
s
p
e
c
tt
ot
h
E
A
l
s
o
L
a
g
r
a
n
g
e
m
u
l
t
i
p
l
i
e
r
T
K
L
a
g
r
a
n
g
e
m
e
t
h
o
d
i
sobnc
a
l
l
e
d
t
h
em
e
t
h
o
d
q
f
u
n
d
e
t
e
r
m
i
n
e
d
c
o
e
f
f
i
c
i
e
n
t
.
i
n
d
i
c
a
t
i
n
gi
ti
so
o
tav
a
r
i
a
b
l
e
.1
nt
h
e a
l

yc

d
e
r

iv
a

o
noft

heFOCfor
:
p e
m

m
l L
a
g
r
a
n
g
i

a
I
w
asc

on
u
n

on
s
i
B
0

r
r

ngf

nt

h
e
p
r
e
v
i
o
u
s
s

0n(

c
o
n
s
t
r
a

i
n
e
dmi

n
1

a
t
i
o

n
) SOCc
b
eexpec
dt
os
a
t
i
s
f
yt
h
e
f
o
l
l
o
w
i
n
gr
e
l
a
t
i
o
n
s
:

(
4
.
8
)
=F(

X*+

F'
(
X
.

)
)
=VF

(
'
(
X
'
)

Theaugmentedp
r
o
b
l
e
mw
i
t
ht
h
eL
a
g
r
a
n
g
i
a
n
:

Minim

0
wSXSXup
X1

x
.
E
q
u
a
t
i
o
n(
4
.
8
)
i
su
s
e
da
f
t
e
rt
h
es
o
l
u
t
i
o
ni
s
o
b
t
a
i
n
e
d
i
fn
e
c
e
s
s
a
l
y
.
E
q
u
a
t
i
o
n

dF= ~dx+
:
:
a
A
.

dX _.. .+
d
X
2-'2 d
-

1- d
h

(
4
.
6
)

0
w~X~Xup
X1

1
dF=df+
Idl
dF.

S
u
b
j
e
c
tt
o
: [
h
(
X
)
]=0

F=f+
I
h
l

dF.

A
N
A
L
Y
T
I
C
A
lC
O
N
D
I
T
I
O
N
S 185

L(X)

T
z

'

'

)
)
]

VhTAX=O

F(X
')=f(X)+

k=l

I
nt
h
ea
b
o
v
e[V2F(X
l
i
st
h
eH
e
s
s
i
a
no
ft
h
eL
a
g
r
a
n
g
i
a
n

w
i
t
hr
e
s
p
e
c
tt
ot
h
ed
e
S
I
g
n

=f(X)+A'h
;
=f(
+lTh

d[v(

(
4
.
5
3
)

v
a
r
i
a
b
l
e
so
n
l
ye
v
a
l
u
a
t
d
a
tt
h
es
o
l
u
t
i
o
n
.
A
l
s
o

t
h
eF0CmquimthatVF(X)=0.With
r
e
f
e
r
e
n
c
et
otwov
a
r
i
a
b
l
e
sa
n
donec
o
n
s
t
r
a
i
n
t
:

4.
4

186 NONL
lN
E
A
RPROGRAMMING

)
+


)
+

t
:
.F=

M=i

A
N
A
L
Y
T
I
C
A
LC
O
N
D
I
T
I
O
N
S

8
7

T
r
a
n
s
f
o
r
m
a
t
l
o
ntosnE
q
u
a
l
i
t
yC
o
n
s
t
r
a
i
n
e
dP
r
o
b
l
e
m

+2

Minimize
(
4
.
5
7
)

z
i
:20xI+15~ -30+z
i=0

S
u
b
j
e
c
t
l
o
: g
l
(
X
I
'
X
z
}+

+2tL

t
:
.
.
xl
t
:
.
.
x
2

(
4
.
3
6
)

j
(
X
I
X
2
)
:X
IX
2

g
2
(
XX
2
)+ :xi/4+x~-1 +z~=0
dh/ox
2

(
4.
59
)

(
4
.
6
0
)

(
4
.
5
8
)
OSX2S3

OSXIS3;

dh1/dx1

(
4
.
3
7
)

S
u
b
s
t
i
t
u
t
eEq
u
a
t
i
o
n(
4.
58
)i
nEq
u
a
t
i
o
n(
4
.
5
7
)a
n
dt
h
eSOCr
e
q
u
i
r
e
st
h
a
tt
h
ee
x
p
r
e
s
s
i
o
n
e
r
i
v
a
t
i
v
e
si
nEq
u
a
t
i
o
n
s(
4
.
5
7
) d(
4
.
5
8
)
i
n bramustbep
o
s
i
t
i
v
e
. d
e
v
a
l
u
a
t
e
da
tt
h
emimum.
A
p
p
l
y
i
n
gt
h
es
e
c
o
n
d
o
r
d
e
rc
o
n
d
i
t
i
o
nt
ot
h
eexampleo
fi
ss
e
c
t
i
o
ni
s1
f
t
e
x
e
r
c
i
s
ef
o
rt
h
e a
d
e
r
.Compar
dt
oeSOCf
o
rt
h
eu
n
c
o
n
s
t
r
a
i
n
e
dm
i
n
i
m
i
z
a
t
i
o
n
p
r
o
b
l
e
m
Equ
a
t
i
o
n
s(
4
.
5
7
)and(
4
.
5
8
) n
o
t syapply
e
s
p
e
c
i
a
l
l
yt
h
es
u
b
s
t
i
t
u
t
i
o
n
o
fE
qu
a
t
i
o
n(
4
.
5
8
)
.Fromap
r
a
c
t
i
c
a
lp
e
r
s
p
e
c
t
i
v
SOCi
sn
o
ti
m
p
o
s
e
df
o
r u
a
l
i
t
y
ti
sl
e
f
tt
ot
h
ed
c
s
i
g
n
c
rt
oe
n
s
u
r
ebyo
t
h
e
rmeans e
c
o
n
s
t
r
a
i
n
e
dp
r
o
b
l
e
m
s
.I
s
o
l
u
t
i
o
ni
saminimums
o
l
u
t
i
o
n
.

T
h
e
r
ea
r
ef
o
u
rv
a
r
i
a
b
l
e
s(
X
I
.
X
2
.Z
I
.Z
2
)andt
w
o
e
q
u
a
l
i
t
yc
O
l
l
s
t
r
a
i
n
t
s
.I
ti
sav
a
l
i
de
q
ua
1i
t
y
i
n
e
dp
r
o
b
l
e
m
. Th
eL
a
g
r
a
n
g
em
u
l
t
i
p
l
i
e
r mt
h
o
dc
a
nb
ea
p
p
l
i
e
dt
ot
h
i
s
c
o
n
s
mau--Tod n
g
u
i
s
hem
u
l
t
i
p
o
c
i
a
t
e
dw
i i
i
t
yc
o
n
s

es
ymbol~ i
su
s
e
d
.Th
i
si
ss
c
t
1
yf
o
rcIt
y
.

4
.
4
.
3 I
n
e
q
u
a
l
l
t
yConstrainedOptlmization

I
ft
h
eL
a
g
r
a
n
g
i
a
ni
sc
o
n
s
i
d
e
r
e
d a
nu
n
c
o
n
s
t
r
a
i
n
e
do
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
n
.t
h
eFOC
(
n
e
c
e
s
s
a
r
yc
o
n
d
i
t
i
o
n
s
)a
r
e

M
e
t
h
o
do
fLsg
r
s
n
g
e
:TheaugmentcdfunctionortheLagrangiani
s
:
M
i
n
i
m
i
z
e

F(x 1 ~ ZI Z2 I'2) =f(xl


.
X
)+s1
[
gI
(
X1

x2
)+z
i
l+s2
[
C
2
(
X
I
')+z~] (
4
.
61
)
2

Theproblem
dF

Minimize

f
(
X
I
Xz
}
:
X
IX
2

S
u
b
j
e
c
tt
o
: g
l
(
X
I
x
2
)
:20X
I+1
5X
2-30S0

df./l o
g
l ./l dg2

(
4
.
6
2
a
)

of./l o
g
l ./l ~g2

(
4
.
6
2
b
)

+s =0
-41
=axlaX12axl
a
x
l

(
4.
36
)

dF

= +I~Ul
+s
2
=0
dX dX2' .
d
x .r~x

(
4.
4
0
)

I
"
'

g2(xl xz):x114+~-1 SO

(
4.
41
)

OSXIS3;

(
4
.
3
7
)

dF

OSX2S3

Thenumbero
fv
a
r
i
a
b
l
e
s(
n=2
) dt
h
enumbero
fi
n
e
q
u
a
l
i
t
yc
o
n
s
t
r

i
n
t
s(m=2
)do
nf
a
c
t
.a
ni
n
e
q
u
a
l
i
t
yc
o
n
sa
i
n
e
dp
r
o
b
l
e
mi
nas
i
n
g
l
e
n
o
td
e
p
e
n
done
a
c
ho
t
h
e
r
.I
v
a
r
i
a
b
l
ec
a
nb
eu
s
e
f
u
l
l
yd
e
s
i
g
n
e
da
n
ds
o
l
v
e
d
.S
e
c
t
i
o
n4.
4
.
2s
o
l
v
e
dt
h
e u
a
l
i
t
y
s
f
o
r
m
e
dt
oa
ne
q
u
i
v
a
l
e
n
t
c
o
n
s
t
r
a
i
n
e
dp
r
o
b
l
e
m
.I
ft
h
eabovep
r
o
b
l
e
mc
a
nb
e n
n we h
a
v
ef
o
u
n
dt
h
es
o
l
u
t
i
o
n
.1
es
t
a
n
d
a
r
d
c
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
e
dp
r
o
b
l
e
m
. e
1
s
f
o
r
m
a
t
i
o
nr
e
qr
e
sas
l
a
c
kv
a
r
i
a
b
l
ez
jf
o
re
a
c
hi
n
e
q
u
a
l
i
t
yc
o
n
s
n
tg
U
n
l
i
k
e
L
P
j

p
r
o
b
l
e
m
s
t
h
es
l
a
c
kv
a
r
i
a
b
l
ef
o
rNLPi
sn
o
tr
e
s
t
r
i
c
t
e
di
ns
i
g
n
.

b
e
r
e
f
o
r
e

t
h
es
q
u
a
r
eo
f
t
h
enewv
a
r
i
a
b
l
ei
sa
d
d
e
dt
ot
h
el
e
f
t
h
a
n
ds
i
d
eo
ft
h
ec
o
r
r
p
o
n
d
i
n
gc
o
n
s
a
i
n
.
tT
h
ls
a
d
d
sap
o
s
i
t
i
v
ev
a
l
u
et
ot
h
el
e
f
t
h ds
i
d
et
ob
r
i
n
gt
h
ec
o
n
s
i
n
tupt
oz
e
r
o
.Ofc
o
u
r
s
e
az
e
r
ov
a
l
u
ew
i
l
lb
ea
d
d
e
di
ft
h
ec
o
n
s
t
r
l
l
i
n
ti
sa
l
r
e
l
l
d
yz
e
r
o
.

_n

(
4
.
6
2
c
)

. n

(
4
.
6
2
d
)

=2s
z
=0
1

1
;

aZ1

dF

?=
2s.z.=0

aZ
2

dF

i
JF

(
4
.
6
2
e
)

)
(
4
.
6
2t

=
g
.+zf=O

d
P
I

ul

=g.+z~=O

d
s
2

U "

188 NONL
lN
E
A
RPROGRAMMING

4
.
4

A
N
A
L
Y
T
'
C
A
LCONOtnONS 189

Eq
u
a
t
i
o
n
s(
4
.
6
2
e
)and(
4
.
6
2
t
)a
r
e u
a
l
i
t
yc
o
n
s
t
r
a
i
n
t
s
.Eq
u
a
t
i
o
ns
e
t(
4
.
6
2
)p
r
o
v
i
d
e
ss
i
x

Eq
u
a
t
i
o
n(
4
.
6
4
)
t
h
eLa
g
r
a
n
g
i
a
ni
sr
e
f
o
r
m
u
l
a
t
e
dw
i
t
h
o
u
tt
h
es
l
a
c
kv
a
r
i
a
b
l
e
s (
w
e
a
r
e

u
a
t
i
o
n
s
t
o
s
o
l
v
e
f
o
r
x
1

x
E

z
;
Z
5
.
9
1
.
.Bysimplerecombination
Eq
u
o
n
s
(
4
.
6
2
c
)

g
o
i
n
gt
op
r
e
t
e
n
dZn
e
v
e
re
x
i
s
t
e
d
)

w
h
i
l
et
h
es
l
a
c
kv
a
r
i
a
b
l
e
sZ
IandZ
2c
a
nb
e
t
o(
4
.
6
2
0c
a
nb
ec
o
l
l
a
p
s
e
dt
otwoe
q
u
a
t
i
o
n
s
e
l
i
m
i
n
a
t
e
d
f
r
o
mt
l
z
ep
r
o
b
l
e
m
.
F
i
r
s
tm
u
l
t
i
p
l
yE
q
u
a
t
i
o
n(
4
.
6
2
c
)byZ
IR
e
p
l
a
c
e by8
1f
r
o
mE
q
u
a
t
i
o
n(
4
.
6
2
e
)
.
Dropt
h
en
e
g
a
t
i
v
es
i
g
na
swellthec
o
e
f
f
i
c
i
e
n
t2t
oo
b
t
a
i
n

(
x
.
x
.
s.
1
3
)=.
f
{
x
.
x
)+s
.
[
g
.
(
x
.
.
x
2
)
]+s
[8T
.
X
)
] (4.65)
Minimize F
2
2
2
2
2

t
z

18.=0

(
4
.
6
3
a
)

1
32
8
2=0

(
4
.
6
3
b
)

w
h
i
c
hi
st
h
esamef
o
r
m
u
l
a
t
i
o
ni
nE
q
u
a
t
i
o
n(
4
.
5
3
)
.Thes
l
a
c
kv
a
r
i
a
b
l
ewasi
n
t
r
o
d
u
c
e
d
r
a
n
s
f
o
r
m
a
t
i
o
nt
oa
ne
q
ui
t
yc
o
n
s i
n.
tI
ti
sa
l
s
oe
v
i
d
e
n
t e
t
op
r
o
v
i
d
e t
c
o
n
s
c
t
i
o
no
fLa
g
r
a
n
g
i
a
nf
u
n
c
t
i
o
ni
si
n
s
e
n
s
i
t
i
v
et
o t
y
p
eo
fc
o
n
s
t
r
a
i
n
tS
i
n
c
e
em
u
l
t
i
p
l
i
e
r
st
i
e
dt
ot
h
ei
n
e
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
ta
r
er
e
q
u
i
r
e
dt
ob
cp
o
s
i
t
i
v
e
w
h
i
l
et
h
o
s
e

t
h
i
sbookw
i
l
lc
o
n
t
i
n
u
et
od
i
s
t
i
n
g
u
i
s
h
c
o
r
r
e
s
p
o
n
d
i
n
gt
ot
h
ee
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
t
sa
r
en
o
t
u
l
t
i
p
l
i
e
r
s
.Th
i
sw
i
l
ls
e
r
v
et
oe
n
f
o
r
c
cc
l
a
r
i
t
yo
fp
r
e
s
e
n
t
a
t
i
o
n
.TheFOC
b
e
t
w
e
e
n m
f
o
rt
h
ep
r
o
b
l
e
m
:

Equ
a
t
i
o
n(
4
.
6
3
b
)i
so
b
t
a
i
n
e
dbyc
a
r
r
y
i
n
go
u
ts
i
m
i
l
a
rm
a
n
i
p
u
l
a
t
i
o
n
sw
i
t
hE
q
u
a
t
i
o
n
s
(
4
.
6
2
d
)and(
4
.
6
2
0
.TheFOCc
a
nb
er
e
s
t
a
t
e
da
s

dF df. d
8
.. d
8
2

=;
J
+I~+ 1
3
2
=0
d
x
. d
x
. r1dx.'r~x.
n

(
4
.
6
2
a
)

e
)
F d
r ~d
8I ~ d
8

=L4144H=o
d
x d
x '
O
X
t.Z
d
x
.

(
4
.
6
2
a
)

dF df.n d
g
. .n d
g
2
+ I"\~I +1
3
2
=0

d
X
z=d
x
'
1
"
"
'
d
.
X
z
.
Z
d
.
X
z
2

(
4
.
6
2
b
)

'

I
"
"

I
"
"

dF d
r ~d
8I ~ d
8

=:
:
L+r
I~+ 1
3
2 =0
d
X
d
X
1
d
x
2
r~X2
2 2

(
4
.
6
2
b
)

1
3
1
8
.=0

(
4
.
6
3
a
)

Casea
: =0[
g
l<
0
]
;
2=0[
g2<0
]

1
3
2
8
2=0

(
4
.
6
3
b
)

Caseb
:1
3 =0[
8
.<0
]
;>0[
g
2=0
]

I
"
"

C
a
s
ec
:s
.>0[
g
.=0
]
;s
2=0[g2<0]

Thesef
o
u
re
q
u
a
t
i
o
n
sh
a
v
et
obes
o
l
v
e
df
o
rx
j
x
i

2
sNote tziZ
2a
r
eo
o
tb
e
i
n
g
yc
a
nb
ed
i
s
c
a
r
d
e
df
r
o
m p
r
o
b
l
c
ma
1t
o
g
c
e
r
.
determincd-whichs
u
g
g
e
s
t
se
I
twouldbeu
s
e
f
u
lt
op
r
e
t
c
n
dz
'
sncvcre
x
i
s
t
c
di
nt
h
ef
i
r
s
tp
l
a
c
e
.
E
q
u
a
t
i
o
n
s(
4
.
6
3
)l
a
yo
u
tad
c
f
i
n
i
t
ef
e
a
t
u
r
ef
o
ran
o
n
t
r
i
v
i
a
ls
o
l
u
t
i
o
n
:c
i
t
h
e
rs
ji
sz
e
r
o
(
a
n
d8
/ 0
)o
r8
/i
sz
e
r
o().S
i
n
c
es
i
m
u
l
t
a
n
e
o
u
sc
q
u
a
t
i
o
n
s b
e
i
n
gs
o
l
v
e
d
e
u
l
t
i
p
l
i
e
r
sa
n
dc
o
n
s
t
r
a
i
n
t
smustbes
a
t
i
s
f
i
e
ds
i
m
u
l
t
a
n
e
o
u
s
l
y
.F
o
r
c
o
n
d
i
t
i
o
n
sonem
Equ
a
t
i
o
n
s(
4
.
6
3
)i
s s
l
a

si
n
t
o ef
o
l
l
o
w
i
n
gf
o
u
rc
a
s
e
s
.Thci
n
f
o
r
m
a
t
i
o
non8i
n
b
r
a
c
k
e
t
si
st
oc
m
p
h
a
s
i
z
ea
naccompanyingc
o
n
s
e
q
u
e
n
c
e
.

8
.=0
]
;>0[
8
2=0
]
Cased
:s
.>0[

J
Casea
:s
.=0[
8
.<0
]
;1
3
2=0[
8
2<O
Caseb
:=0[8.<0];*0[
8
2=0
]
Casec:1
31*0[
8
.=
0
]
;1
3
2=0[82<0)
C d
:
I 0
[
8
.=
0
)
;*0[
8
2=0
]

(
4
.
6
4
)

I
n Eq
u
a
t
i
o
n(
4
.
6
4
)
i
fs
j*0 (
o
rc
o
r
r
e
s
p
o
n
d
i
n
g8
/=0
)t
h
e
nt
h
ec
o
r
r
e
s
p
o
n
d
i
n
g
aI
n
ti
sa
ne
q
u
a
l
i
t
y
.I
nt
h
ep
r
c
v
i
o
u
ss
c
c
t
i
o
nas
i
m
p
l
er
e
a
s
o
n
i
n
gwasusedshow
c
o
n
s
cs
i
g
no
fem
t
i
p
l
i
e
rmustbep
o
s
i
t
i
v
e(>0
)f
o
raw
e
l
l
f
o
r
m
u
l
a
t
e
dp
r
o
b
l
e
m
.
Whi
l
et
h
es
i
g
no
ft
h
em
u
l
t
i
p
l
i
e
rwasi
g
n
o
df
o
ree
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
e
dp
r
o
b
l
e
m

i
t
i
si
n
c
l
u
d
e
da
sp
a
r
to
ft
h
eFOCf
o
r i
n
e
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
e
dp
r
o
b
l
e
m
.B
e
f
o
r
er
e
s
t
a
t
i
n
g

(
4
.
6
4
)

E
q
u
a
t
i
o
ns
e
t(
4
.
6
2
)a
n
da
n
ys
i
n
g
l
ec
a
s
ei
nE
q
u
a
t
i
o
n(
4
.
6
4
)p
r
o
v
i
d
e
sf
o
u
re
q
u
a
t
i
o
n
st
o
s
o
l
v
f
o
rt
h
cf
o
u
runknownso
ft
h
ep
r
o
b
l
e
m
.A
l
lf
o
u
rs
c
t
sm
u
s
tb
es
o
l
v
e
df
o
rt
h
e
eb
e
s
td
e
s
i
g
ni
sd
e
c
i
d
c
dbys
c
a
n
n
i
n
gt
h
es
e
v
e
r
a
ls
o
l
u
t
i
o
n
s
.
s
o
l
u
t
i
o
n
.Th
o
tas
u
j
J
i
c
i
e
n
tc
o
n
d
i
t
i
o
nf
o
rt
h
e
Thcs
i
g
no
ft
h
em
u
l
t
i
p
l
i
e
ri
nt
h
es
o
l
u
t
i
o
ni
sn
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
e
dproblemTh
cv
a
l
u
ei
sun
p
o
r
t
a
n
tf
o
ro
p
t
i
m
i
z
a
t
i
o
nb
u
tmayb
e
m
u
l
t
i
p
l
i
e
ri
n
d
i
c
a
t
c
s
r
e
l
e
v
a
n
tf
o
rs
c
n
s
i
t
i
v
i
t
ya
n
a
l
y
s
i
s
.G
e
n
e
r
a
l
l
yap
o
s
i
t
i
v
ev
a
l
u
eo
f
t
h
es
o
l
u
t
i
o
ni
sn
o
tal
o
c
a
lmaximum.F
o
r
m
a
l
l
yv
e
r
i
f
y
i
n
gaminimums
o
l
u
t
i
o
n
sc
o
n
s
i
d
e
r
a
t
i
o
no
ft
h
es
c
c
o
n
dd
c
r
i
v
a
t
i
v
eo
ft
h
cLa
g
r
a
n
g
i
a
n
.I
np
r
a
c
t
i
c
a
l
r
e
q

t
h
cp
o
s
i
t
i
v
ev
a
l
u
eo
fcm
u
l
t
i
p
l
i
e
ru
s
u
a
l
l
y
s
i
t
u
a
t
i
o
n
s
.i
ft
h
cp
r
o
b
l
e
mi
sw
e
l
ld
e
f
i
n
e
d
s
u
g
g
c
s
t
saminimums
o
l
u
t
i
o
n
.Th
i
si
su
s
e
dc
x
t
e
n
s
i
v
c
l
yi
nt
h
cbookt
oi
d
e
n
t
i
f
yt
h
e
optimums
o
l
u
t
i
o
n
.

S
o
l
u
t
l
o
no
ft
h
eEx
ample:ThcLa
g
r
a
n
g
i
a
nf
o
rt
h
ep
r
o
b
l
e
mi
sd
e
f
i
n
c
da
s

F
(
x
.
x
.
s.
s
)
=
x
.
.
X
z+s(
2
C
k
.+1
5
X
z-30)+(0.i~+ .s -l)
2
2
TheFOCare

1
9
0 N
O
N
L
I
N
E
A
RP
R
O
G
R
A
M
M
I
N
G

4
.
4

dF

Z;=42+2ORI+05h=0

dF

55=41+15Pl+2h=0

a
s
eawasd
i
s
c
u
s
s
e
de
o
rC
a
r
l
i
e
randi
sc
Thes
o
l
u
t
i
o
nf
o
n
r
t
r
m
e
da
b
o
v
e
.
o
l
u
t
i
o
n
s
.Thef
n
ei
su
n
a
c
c
e
p
t
a
b
l
eb
e
c
a
u
s
ec
C
a
s
ebh
o
u
rs
i
r
s
to
n
sf
o
n
s
t
r
n
i
n
tg
sn
o
t
li
o
l
u
t
i
o
ni
st
h
ef
u
n
c
t
i
o
na
1c
e
c
o
n
ds
sf
e
a
s
i
b
l
en
a
ra
s
a
t
i
s
f
i
e
d
.Thes
sf
o
n
s
t
r
a
i
n
t
sg a
n
dg2
h
e
ydon
o
ts
a
t
i
s
f
yt
i
d
ec
o
n
s
t
r
a
i
n
t
s
.Th
et
u
tt
h
es
a
r
econcemedb
h
i
r
da
n
df
o
u
r
t
h
s
o
l
u
t
i
o
n
sa
n
a
c
c
e
p
t
a
b
l
ef
o
rt
h
esamer
e
a
s
o
n
.T
a
s
ebi
sn
r
eu
h
u
s
.C
o
tn
no
p
t
i
m
n
l
s
o
l
u
t
i
o
n
.
C
n
s
eci
n
a
c
c
e
p
t
a
b
l
eb
e
c
a
u
s
ec
o
n
s
t
r
a
i
n
tg2i
si
nv
i
o
l
a
t
i
o
n
.
su
C
n
s
edh
o
l
u
t
i
o
n
st
h
ef
i
r
s
to
h
i
c
hi
sn
o
ta
c
c
e
p
t
a
b
l
ef
n
stwos
fw
o
rs
e
v
e
r
a
lr
e
n
s
o
n
s
.The
r
s
e
c
o
n
ds
o
l
u
t
i
o
ns
l
lo
ft
h
er
e
m
e
n
t
s
n
t
i
s
f
i
e
sn
e
q

(
4
.
66
a
)

(
4
.
6
6
b
)

(
2
O
xI+15X
s1
z-30)=0

(
4
.
6
6
c
)

(
0

+4-1)=0
s2

(
4
.
6
6
d
)
I
ts
h
ec
o
n
s
t
r
a
i
n
t
s
.8
n
t
i
s
f
i
e
st
0
t
hc
o
n
s
t
m
i
n
t
s n
c
t
i
v
e
.
Them
o
s
i
t
i
v
e(maybeas
u
f
f
i
c
i
e
n
tc
o
n
d
i
t
i
o
n
)
.
u
l
t
i
p
l
i
e
r
s ep
h
es
i
d
ec
a
t
i
s
f
i
e
st
o
n
s
t
r
a
i
n
t
s
.
Its

Th
es
i
d
ec
o
n
s
t
r
a
i
n
t
s
OSX1S
3;

x
.
z

O S S3

(
4.
)
37

l
o
l
u
t
i
o
ni
s
es

Casea
:s
;s
:Thes
o
l
u
t
i
o
ni
l=0
si
2=0
v
i
a
la
n
dbyi
n
s
p
e
c
t
i
o
no
fE
q
u
a
t
i
o
n
s(
4
.
6
6
n
)
a
n
d(
4
.
6
6
b
)
.

x1=
0
.
8
1
5
1
;
.
7
4
4
3
;
1=0

.
9
1
3
2
; =
0
.
0
4
3
9
;
X
2=0
0
;
0
g
l
g
2

=0.0856;

si
sn
l
m
o
s
tc
o
n
f
i
r
m
e
dby n
u
m
e
r
i
c
n
ls
o
l
u
t
i
o
nw
h
i
c
hn
p
p
e
a
r
st
oh
n
v
eap
Thi
r
o
b
l
e
m
c
.I
l
m
o
s
ta
tt
h
es
o
l
u
t
i
o
n
.1
tmustb
e
r
et
h
eMATLAD
o
n
v
e
r
g
e
n
ti
w
i
sa
en
o
l
e
dh
h
a
tt
os
o
l
v
e p
r
o
b
l
e
mi
a
r
to
f
u
n
c
t
i
o
nu
s
e
dt
ef
u
n
c
t
i
o
n
si
sp
ft
h
es
t
a
n
d
a
r
dp
a
c
k
a
g
e
.Th
n
e
t
t
e
r
.N
e
v
e
r
t
h
e
l
e
s
s
.t
n
c
o
m
p
l
e
t
e
t
h
eO
p
t
i
m
i
z
a
t
i
o
nT
o
o
l
b
o
xw
i
l
lc
e
r
t
a
i
n
1
ydob
h
ei
s
o
l
u
t
i
o
ni
e
l
o
wb
u
tw
g
n
o
r
e
dd
oes
sr
e
p
o
r
t
e
db
i
l
lbei
u
et
y
m
b
o
l
i
cs
o
l
u
t
i
o
na
v
a
i
l
a
b
l
e
i
nC
a
s
ed
.

x =0.8187;

n
s
???E

X1=0; x
0
;
2=0; 1=0;
gJ=3
(
4
.
6
7
)
g
2=
Thei
n
e
q
u
a
1
i
t
yc
t
o
n
sa
r
es
a
t
i
s
f
i
e
d
.Thes
i
d
ec
a
o
n
i
n
t
sa
r
e i
s
f
i
e
d
.Thev
a
l
u
e
s
i
nE
q
u
a
t
i
o
n(
4
.
6
7
)r
s
e
p
e
n
tap
o
s
s
i
b
l
es
o
l
u
t
i
o
n
.
Thes
o
l
u
t
i
o
n
sf
o
rt
h
eo
t
h
e
rc
a
s
e
s'co
b
t
a
i
n
e
du
s
i
n
gMA
ec
11.ABl
o
d
ei
sa
v
a
i
l
a
b
l
e
i
nSec4_4_3
m.F
o
rC
a
s
e
sba
n
dct
h
ea

p
p
r
o
p
r
i
a
t
em
u
l
t
i
p
l
i
e
ri
ss
e
tt
oz
e
r
oa
n
dt
h
e
r
e
s
u
l
t
i
n
gt
h
r
E
eE
q
u
a
t
i
o
n
si
nt
h
r
e
eunbEownsa
r
es
o
l
v
e
d
.
F
o
rc
a
s
Edt
h
ec
o
m
p
l
e
t
es
e
t
o
ff
o
u
re
q
u
a
t
i
o
n
si
nf
o
u
runknownsi
ss
o
l
v
e
d
.
c
a
s
e
di
sa
l
s
oS
O
l
v
e
dn
u
m
e
r
i
c
a
l
l
y
.l
e
MA
T
I
.
ABc
o
d
ew
i
l
la
1
s
or
u
nc
o
d
eFlg4_8.mw
h
i
c
hc
o
n
t
a
i
n
st
c
h
ommandst
od
r
a
we
f
i
g
u
r
e
.Th
eo
u
t
p
u
tf
r
o
mt
h
eCommandwindowi
sp
a
t
c
h
e
db
e
1
0
w
.
The solution *** Case a ***(x1*
x2*f*9192)
o
0
0 30
1
The solution **Case (
** (x1*
b
)

b2* f*9192):
x2
1.4142
0.7071
1.0000 -1.0000
8.8909 0
-1.4142 -0.7071
1.0000 -1.0000 6.8909 0
-1.4142
0.7071 -1.0000
1.0000 -47.6777 0
1.4142 -0.7071 -1.0000
1.0000 -12.3223 0

Case (c)*** (x1*


The solution *
X2*b1* f*9192):
o.75001.oooo
o.0500-o.7500o
o.14o6
Thesolution***Case(d}***{xl*
x2*bl*b2*f*glg2):
1.8150 0.4200
0.0426 1.4007 0.7624 0 0
0.8151
0.9132
0.0439
0.0856 -0.7443 0 -0.0
Maximum nu
of function evaluations exceeded;
lncrease options.MaxFu
nEvals
Optimizer is stuck at a minimum that is not a
root
Try a9ain with a new startin9 guess
The numerical solution (

X2

b1*
x
1
b2*):
0.8187
0.9085
0.0435
0.0913

A
N
A
L
Y
T
I
C
A
L
C
O
N
D
IT
l
ONS 1
9
1

x
2=0.9085;

3
5
;
s =O.

=0.0913

o
l
u
t
i
o
n
:et
r
i
v
i
n
ls
o
l
u
t
i
o
ni
nC
a
s
eaa
n
dt
T
h
e
r
ea
r
enowt
w
oc
a
n
d
i
d
a
t
e
sf
o
res
h
e
s
o
l
u
t
i
o
ni
n
s
ed
.Thes
o
l
u
t
i
o
ni
nC
di
sf
a
v
o
r
e
dn
si
th
o
w
e
ro
b
j
e
c
t
i
v
ef
nC
n
s
n
sal
u
n
c
t
i
o
n
s
v
u
e
.A
x
p
l
o
r
e
dwi
a
1
p
e
c
tt
oF
i
g
u
r
e4
.
8
.I
ti
sa
n
l
lo
ft
h
ec
n
s
e
sa
b
o
v
ec
a
nb
ee
e
x
c
e
l
l
e
n
tf
a
c
i
l
i
t
a
t
o
rf
o
rc
o
m
p
r
e
h
e
n
d
i
n
gt
h
ec
a
s
e
sa
n
d s
o
l
u
t
i
o
n
.
4
.
4
.
4 GeneralOptimlzatlonProblem
e
s
c
r
i
b
e
di
nt
h
es
e
to
eg
e
n
e
r
r
o
b
l
e
mi
sd
fE
q
u
a
t
i
o
n
s(
4
.
1
)(

'
h
a
1o
4
.
8
)
.l
e
p
t
i
m
i
z
a
t
i
o
np

s
p
e
c
i
f
i
cp
.
3i
r
o
b
l
e
mi
i
n
e
di
nS
e
c
t
i
o
n4
s
h
a
p
t
e
rdf
nt
h
i
sc
M
i
n
i
m
i
z
e

X
l):'
j
(
X
2
X

x
S
u
b
j
e
c
tt
o
: h
1
(
X
l
V
:20x+15X2-30=0

x
g
l
(
X
V
:<4/4)+(~) IS0
OSXlS3;

OSX2S3

(
4
.
9
)
)
(
4.
10
(
4.
)
11
(
4
.
1
2
)

NONUNEARPROGRAMMING
4
.
4 ANAL
YTICALCONOITIONS

Z
A
i
i
l
;
;
3
5
5

M
u
/
i
od:Theproblcm

M
i
n
i
m
i
z
e

s
f
i

F
(
x1
X
n

A.
.
.
.

/
.s
I
'....s
r
n
)=

f
(
x
.
.
X
)+
h
+

+s.g
+ +sr
n
C
m
n

OSX2S3

(
4
.
7
0
)

Th
e
r
ea
r
en+1+mu
n
k
.
n
o
w
n
s
.Th
es
a
r
n
enumbero
fc
q
u
a
t
i
o
n
sa
r
e.
r
u
i
r
e
dt
os
o
l
v
e
ep
r
o
b
1
e
m
.
e
s
e "
ep
r
o
v
i
d
e
dbyt
h
eFOCo
reKuhnTuckerc
o
n
d
i
t
i
o
n
s
:
1
1c
q
u
a
t
i
o
n
sa
r
eo
b
t
a
i
n
e
da
s

aF
Z;=-X2+2
I+O.
5s =0

aF

~ =
-x.+x
2=o

aF

-=2 1+15~-30=0

(
4
.
6
8
a
)

a
t
i
o
n
sa
r
o
b
t
a
i
n
e
dd
i
r
c
c
t
l
yr
u
u
g
hecqu
i
t
yc
o
n
s
t
r
a
i
n
t
s
l u
(
4
.
6
8
c
)

=0;

g
l<0

>O;

(
4
.
6
9
a
)

Caseb
:

g
.=0

(
4
.
6
9
b
)

u
irthesoluuo

s
y
s
t
e
m

(
4
.
7
2
)

j<0
t
h
e
ng

npu

nu
U

--=

(
4
.
2
)

nu

"
k
(
X
I
X
2
.
.
..
.
x
n
)
=
O
.k=1
.
2
.
.
..
.
1

(
4
.
1
)

f
(
X
.
x
2

.
..x

;:

k=1
.
2
.
..
.
.
1

mc
q
u
a
t
i
o
n
sa
r
ea
p
p
i
1
edt
h
r
o
u
g
h 2m c
a
s
e
s
.T
h
i
si
m
p
l
i
e
s e
r
ea
r
e2mp
o
s
s
i
b
1
e
s
o
l
u
t
i
o
n
s
.Thescs
o
l
u
t
i
o
n
smusti
n
c
l
u
d
eE
q
u
a
t
i
o
n
s(
4
.
7
1
)a
n
d(
4
.
7
2
)
.Eachc
a
s
s
e
t
st
h
e
r c
o
r
r
e
s
p
o
n
d
i
n
gi
n
e
q
u
a
l
i
t
y
c
0
n
s
t

a
i

n
t
t

oz
e
I

f
t

em
l
t
m
u
l
t
i
p
l
i
e
ro
s
e
tt
oz
e
r
o
.e
nec
o
r
r
e
s
p
o
n
d
i
n
gc
o
n
s
i
n
tmustbef
e
a
s
i
b
1
ef
o
ra
na
c
c
c
p
t
a
b
l
e
o
n
s
t
r
a
i
n
ti
ss
e
tt
oz
e
r
o(
a
c
t
i
v
ec
o
n
s
t
r
a
i
n
t
)
.t
h
e
nt
h
ec
o
r
r
e
s
p
o
n
d
i
n
g
s
o
l
u
t
i
o
n
.I
fec
si
nmindcme
q
u
a
t
i
o
n
sc
a
nb

m
u
l
t
i
p
l
i
e
rmustb
ep
o
s
i
t
i
v
ef
o
raminimum.Withi
e
x
p
r
e
s
s
e
da
s
e
oJ

.
dd

i
2
3
2
2
2
!
1
2
2
3
;

hk (xI' X2 .. .X
0;
n)=

dF df.~ d
h
"
.
dhl.n dg
. .
n d
g
m
(
4
.
71
)
=+
A
.

:
;

+PI--++s
m
:
d
x
dx

"
'
d
x

'
d
x

-r
a
x

-rm
a
x
=0; i=1.2.....n

(
4
.
6
8
b
)

Casea
:

;;zzuyf

Su

l
t
o
:

(
4.
4
)

Minimize

TheFOC

Mi
n
i
m
i
z
e

=1.2.... n

TheL
a
g
r
a
n
g
i
a
n
:

S
u
b
j
e
c
t1
0
: "
2
)
:2
0
x
I+1
1
(
X

JX
5
x
2-30=0

3
2
;
:

~~X
I~.xl'

(
4
.
3
)

ym
i
n
i
m
i
z
i
n
gt
h
c

X.
F
(
X

J 2
A
.
.
.
s
U=-X.
X
2+1
(
2
0
x
.+1
5
x
2-3
0
)+s
.
(
0
.
2
5
x
?+xi-1)
OSXIS3;

;
(
XI
oX2

..
.

JSO
.j=1
.
2
.... m

1
9
3

>0

(
4
.
7
3
)

I
fcon onsi
nE
q
u
a
t
i
o
n
s(
4
.
7
3
)a
r
en
o
tm
e
t
.edigni
sn
o
ta
c
c
e
p
t
a
b
1
e
.I
n
i
m
p
l
e
m
e
n
t
i
n
gE
q
u
a
t
i
o
n(
4
.
7
3
)
.f
o
re
a
c
hc
a
s
eas
i
m
u
1
t
a
n
e
o
u
st
o o
fm v
a
1u
e
sa
n
d
c
q
u
a
l
i
t
i
e
smustbea
s
s
i
g
n
e
d
.Oncet
h
e
s
eFOCc
o
n
d
i
t
i
o
n
sd
e
t
e
r
m
i
n
eap
o
s
s
i
b
l
es
o
l
u
t
i
o
n
.
t
h
cs
i
d
ec
o
n
s
t
r
a
i
n
l
Sh
a
v
ct
ob
ec
h
e
c
k
e
d
.Asc
v
i
d
e
n
c
e
di
nce
x
a
m
p
1
e
se
a
r

ie
r
.t
h
i
si
s
ti
so
n
1
yc
o
n
f
t
r
m
e
da
f
t
e
rap
o
s
s
i
b
1
es
o
l
u
t
i
o
nh
a
sb
e
e
n
n
o
tb
u
i
l
ti
n
t
o FOC. l
i
d
e
n
t
i
f
i
e
d
.E
q
u
a
t
i
o
n
s(
4
.
7
1
)(
4
.
7
3
)a
r
er
e
f
e
r
r
e
dt
oa
st
h
eKuhn

T
u
c
k
e
r
c
o
n
d
i
t
i
o
n
s[
s
e
e
E
x
a
r
n
p
l
e4
.
3(
S
e
c
t
i
o
n4.
5.
2
)f
o
ra
d
d
i
t
i
o
n
a
ld
i
s
c
u
s
s
i
o
no
ft
h
eKuhnT
u
c
k
e
rc
o
n
d
i
t
i
o
n
s
]
.

i111114I11Jjq

192

4
.
5E
X
A
M
P
L
E
S 195

194 NONl
IN
E
A
RPROGRAMMING

Ex
p d
i
n
gt
h
ee
x
p
r
e
s
s
i
o
n
si
nt
h
eb
r
a
c
k
e
t
sa
n
dr
e
o
r
g
a
n
i
z
i
n
ga
sal
i
n
e
ma
i
x
e
q
u
o
n

4
.
5 EXAMPLES
ef
i
r
s
ti
sa
nu
n
c
o
n
s
t
r
a
i
n
e
dp
r
o
b
l
e
m
Twoe
x
a
r
n
p
l
e
sa
r
ep
r
e
s
e
n
t
e
di
nt
h
i
ss
e
c
t
i
o
n
.Th
h
a
ss
i
g
n
i
f
i
c
a
n
tu
s
i
nd
a
t
ar
e
d
u
c
l
i
o
n
.S
p
e
c
i
f
i
c
a
l
l
y
i
ti
l
l
u
s
t
r
a
t
e
st
h
ep
r
o
b
l
c
mo
fc
u
r
v
e
es
e
c
o
n
di
st
h
cb
e
a
r
nd
e
s
i
g
np
r
o
b
l
e
me
x
p
l
o
di
nC
h
a
p
t
e
r
s1
f
i
t
t
i
n
go
rr
e
g
r
e
s
s
i
o
n
.Th
a
n
d2
.

1
1

yf2Y;I~

Problem: A s
e
to
fy
Zd
a
t
ai
sg
i
v
e
n
.I
ti
sn
e
c
c
s
s
a
r
yt
of
i
n
dt
h
eb
e
s
ts
t
r
a
i
g
h
tl
i
n
e
r
o
u
g
ht
h
ed
a
t
a
.T
h
i
se
x
e
r
c
i
s
ei
st
e
r
m
e
dc
v
ef
i
t
t
i
n
go
rd
a
t
a d
u
c
t
i
o
no
rr
e
g
r
e
s
s
i
o
n
.
Pt
h
i
se
x
a
r
n
p
l
es
i
m
p
l
et
h
ed
a
t
ap
o
i
n
t
sa
r
eg
e
n
e
r
a
t
e
du
s
i
n
gt
h
ep
a
r
a
b
o
l
i
ce
q
u
a
t
i
o
n
Tok
Z=O
.
5
y
2
.Thedataa
r
eg
e
n
e
r
a
t
e
di
na
r
b
i
t
r
a
r
yo
r
d
e
ra
n
dnonun
ormi
n
t
e
r
v
a
l
s
.

LeastSquaredE
r
r
o
r
:Theo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
n d
r
i
v
e
sal
a
r
g
enumbero
fc
u
r
v
e
e
f
i
t
t
i
n
g
l
r
e
g
r
e
s
s
i
o
nm
e
t
h
o
d
si
st
h
em
i
n
i
m
i
z
a
t
i
o
no
ft
h
el
e
a
s
ts
q
u
a
r
e
de
r
r
o
r
. l
c
o
n
s
t
r
u
c
t
i
o
no
ft
h
eo
b
j
e
c
t
i
v
e i
r
e
s two e
n
t
i
t
i

t
h
ed
a
t
aa
n
d t
y
p
eo
f
m
a
t
h
e
m
a
l
i
c
a
lr
e
l
a
t
i
o
n
s
h
i
pb
e
t
w
nt
h
ed
a
t
a
.I
nt
h
i
sx
a
r
n
p
l
ei
ti
sas
t
r
a
i
g
h
tl
i
n
e
.
o
ra
n
yf
u
n
c
t
i
o
nr
o
ra
tm
a
n
n
e
r
.T
h
i
si
su
s
u
a
l
l
y
G
e
n
e
r
a
l
l
yi
tc
a
nb
eanyp
o
l
y
n
o
m
i
a
l
x
.p
e
r
i
m
c
n
t
a
lf
l
u
i
dd
y
n
a
m
i
c
sandh
e
a
t
t
h
ewayc
o
r
r
c
l
a
t
i
o
ne
q
u
a
t
i
o
n
sa
r
ed
e
t
c
r
m
i
n
e
di
ne
t
r
a
n
s
f
e
r
.F
o
rt
h
i
se
x
n
p
l
eed
a
t
aa
r
et
h
ec
o
l
l
e
c
t
i
o
no
fp
o
i
n
t
s(
Y
;
Z
/
)
i
=1

2

n
T
h
1
e
e

.
x
p
e
c

ds
t
r
a
i
g
h
tI
in
c
b
c

ha

ra

C~
.er

i
zedb

Z=X1Y+X
z
wherex
li
st
h
es
l
o
p
eandX2i
sei
n
t
e
r
c
e
pt
.U
singo
p
t
i
m
i
z
i
n
gt
e
r
m
i
n
o
l
o
g
y
t
h
e
r
ea
r
e
. F
o
rac
u
b
i
cp
o
l
y
n
o
m
i

I
at
h
e
r
ewi
l
bef
o
u
rd
e
s
i
g
nv
a
r
i
a
b
l
e
s
.
twod
e
s
i
g
nv
a
r
i
a
b
l
e
Z
p
ji
s
f
i
t
t
e
dv
a
l
u
ef
o
rei
n
d
e
p
e
n
d
c
n
tv
a
r
i
a
b
l
eY
;
eo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
n
whichi
s
e
o
ft
h
ee
r
r
o
ro
v
c
ra
I
lo
ft
h
ed
a
t
ap
o
i
n
t
s
.c
a
nb
cxp
s
s
e
d
a
s
t
h
eminimumo
ft
h
cs
q
u
n

1
1

;
=
1

1
=
>1

M
i
n
i
m
i
z
e f
(
x1
Z
/(
X +
X
2
)
]
2
)= [Z/-zpi= [

(
4
.
7
4
)

I
ti
sp
o
s
s
i
b
l
et
oc
o
m
p
l
e
t
e f
o
r
m
u
l
a
t
i
o
nwi
s
i
d
ec
o
n
s
n
t
sa
I
o
ugh i
sn
o
t
n
e
c
e
s
s
y
.l
eFOCa
r

=;2IZi-

2[Zj-(x

(
4
.
7
5
a
)

(
4
.
7
5
b
)

1
1

L
Y
; L11

/
=
1

4
.
5
.
1 Example4
.
2

"

I
=
J

IL
Z
;
Y
;

I
i
z
;
;
=
1

Not
t
h
a
tt
h
em
a
t
r
i
c
e
sc
a
nb
es
e
tupe
a
s
i
l
ya
n
ds
o
l
v
e
du
s
i
n
gMATLAB.TheH
e
s
s
i
a
n
em
a
t
r
i
xont
h
el
e
f
tTh
eSOCr
e
q
u
i
r
e
sa
ti
tb
ep
o
s
i
t
i
v
ed
e
f
i
n
i
t
e
.
m
a
t
r
i
xi
st
h
es
q
u

Code:Sec4_5_1.m: E
x
a
r
n
p
l
e4
.
2i
ss
o
l
v
e
du
s
i
n
gMATLAB dt
h
er
e
s
u
l
t
sf
r
o
me
ti
nC
h
a
p
t
e
r1
.The
Commandwindowa
r
eshownb
e
l
o
w
.Comparet
h
i
sc
o
d
ew
i
t
ha
r
e
s
u
l
t
sa
r
ed
i
s
p
l
a
y
e
donap
l
o
t(
n
o
ti
n
c
l
u
d
e
di
n b
o
o
k
)
.Newf
e
a
t
u
r
e
si
n c
o
d
ea
r
e
u
s
i
n
gt
h
esumcommand
andw
r
i
t
i
n
gt
ot
h
eCommand
t
h
eu
s
eo
frandomn
u
m
b
e
r
s
window.Fromt
h
ep
l
o
ta
n
d t
a
b
u
l
a
rv
a
Iu
e
si
ti
sc
l
e
a
rt
h
a
tal
i
n
e
a
rf
i
ti
sn
o
ta
c
c
e
p
t
a
b
l
e
f
o
ri
se
x
a
m
p
l
e
.Th
equa
Ii
t
yo
ft
h
ef
i
ti
sn
o
tu
n
d
c
rd
i
s
c
u
s
s
i
o
ni
ni
sx
a
r
n
p
l
ea
si
ti
s
o
n
l
i
n
e
a
rf
i
ts
h
o
u
l
dh
a
v
eb
c
e
nc
h
o
s
e
n
.T
h
isi
sl
e
f
la
sa
ne
x
e
r
c
i
s
c
.Thc
a
p
p
a
r
e
n
t an
o
d
ei
sr
u
n
.
f
o
l
l
o
w
i
n
gw
i
l
la
p
p
e
a
ri
neCommandwindowwhcnec
Results from Linear fit
objective function: 16.5827
design variables x1x2:
2.3458
-1.5305
diff
Zp
Zl
.
yi
1.6514
9.5951
4.7427 11.2465
7.2031 -0.2727
6.9304
3.7230
0.7714
8.6552
9.4266
4.3420
0.7179
0.0709 0.6470
0.3766
4.1089 -1.2193
2.8895
2.4040
0.3270
8.1002
8.4272
4.1054
0.3757
8.1646
8.5404
4.1329
0.7299
6.2973
5.5674
3.3369
4.9991 -1.1253
3.8739
2.7835
0.5039
8.3296
4.2032
8.8334
0.9267
5
.
7
7
2
3
3.1131
4.8456
-0.9241
5.7802
4.8561
3.1164
0.8199
6.0737
3.2415
5.2538
1.0911
0.0191
-1.0720
0.1954
0.0871
0.1781
0.2653
0.7284
1.1168
9.0450
4.5082 10.1619
-1.1247
2.9430
1.8183
1.9070
0.4874
-0.3638
0.1237
0.4974

196 NON
l
IN
EAAPAOGAAMMING
4
.
5E
X
A
M
P
L
E
S 197
1.9025

-1.1227
1.1356

I~I

429
EpdpfJSEJ

2.9325
-1.1203

X
r

a
M

4FEq4

1.8098
0.0153

s
e
u

a72
V04
n45
e
954

10
e2

0.1749

4
.
5
.
2 Example4
.
3
T
h
i
sexmnplci
st
h
esamea
sExample23.thef
l
a
g
p
o
l
epmMem.Iti
sq
u
i
t
ed
i
f
f
i
c
u
l
tt
o

3
3
1
5

i
t
z
i
F z
c
Minimize f
(
x
l)=6
.
0
5
5
9E05(x~ -x~)

(
4
.
7
6
)

S
u
b
j
e
c
tt
o
:
g
l
(
X
)
:
7
.
4
9
6
9E05
.
x
7+40000x1- 9.7418E06(x1-x~) s0
g
2
(
XI
')
:
(
5

+1
.4994E05xr
)
(
.
x
7+X) 17083E07(Xi-x~) S0(4.77b)

g
3
(
Xl
'x
2
)
:1
.909IE-03xl+6.1116E- -0.05(x1-xi
)S0
g4(XI'~) :X2 -x1+0
.
0
0
1S0

.0
;
0
.
0
2Sx1S1

i
7

(
4
.
7
7
a
)

(
4
.7
7
c
)
(
4
.
7
7
d
)

0
.
0
2Sx S1
.0

(
4
.
7
8
)

K
'
u

hn
-'
u
c
k
e

.
r'
o
n
d
i
i

on:
s
F

ou
r

La
g
r
a
n

g
cmL
t
p
l
i

r
s

a
r

i

I
r
r
u
t

y
c
o t

r
e
U
n
t

s dt

heLa
g
r

a
n

g
i

a
ni
s

F(X

s
)=6.0559E0
5
(
x
f-~) +

4
969E0
5
x
f+40

OXr-9.7418E06(x1-xi))+
s p.

2(
1
3
(
5
0 +1
.
4
994E05x
)
(
x
f+x+~)
l

1.7083E07(x1-4)) +

(
1
.9091E-03x.+6.1
l16E- 0.05(x1-4
)+

.
x

s
+0.001)
4(X
2-

(
4
.
7
9
)

whereX=[
X
oX
2]a
n
ds=[
1
3
1
1
3
2

s
4
]
T
.l
eKuhnT
u
c
k
e
rc
o
n
d
i
t
i
o
n
sr
e
q
u
i
r
etwo
u
a
t
i
o
n
su
s
i
n
gt
h
eg
r
a
d
i
e
n
t
so
fI
h
eL
a
g
r
a
n
g
i
a
nwi s
p
e
c
t
l
o ed
e
s
i
g
nv
a
r
i
a
b
l
e
s

a
F _a
F

=0:

a
X
a
X
2=0
r

(
4
.
8
0
)

Ie
x
p
r
e
s
s
i
o
nf
o
r eg
r
a
d
i
e
n
t
so
b
t
a
i
n
e
dfromE
q
u
a
t
i
o
n(
4
.
8
0
)i
sl
e
f
ta
s
Thea
c
a
c
x
e
r
c
i
s
ef
o
rt
h
es
t
u
d
e
n
t
.
l
eKuhnTuckerc
o
n
d
i
t
i
o
n
sa
r
ea
p
p
l
i
e
dbyi
d
e
n
t
i
f
y
i
n
gt
h
e
v
a
r
i
o
u
sc
a
s
e
sb
a
s
e
dont
h
ea
c
t
i
v
e
n
e
s
so
fv
a
r
i
o
u
ss
e
t
so
fc
o
n
s
t
r
a
i
n
t
s
.
4 6c
F
o
rt
h
i
sp
r
o
b
l
e
mn=2a
n
dm =4
.T
h
c
r
ea
r
e2
=2 = 1
a
s
e
st
h
a
tm
u
s
tb
e
i
n
v
e
s
t
i
g
a
t
c
d p
a
r
to
f Kuhn

Tuckerc
o
n
d
i
t
i
o
n
sg
i
v
e
nbyE
q
u
a
t
i
o
n(
4
.
7
3
)
.While
someo
ft
h
e
s
ec
a
s
e
sa
r
et
r
i
v
i
a
l
n
e
v
e
r
t
h
e
l
e
s
si
ti
saf
o
r
m
i
d
a
b
l
et
a
s
k
.T
h
i
sp
r
o
b
l
e
mwas
s
o
l
v
e
dg
r
a
p
h
i
c
a
l
l
yi
nC
h
a
p
t
e
r2andt
h
esameg
r
a
p
h
i
c
a
ls
o
l
u
t
i
o
nc
a
nb
ee
x
p
l
o
i
t
e
d1
0
i
d
e
n
t
i
f
yt
h
ep t
i
c
u
l
a
rc
a
s
e n
e
e
d
st
obes
o
l
v
e
df
o
rt
h
es
o
l
u
t
i
o
n
.V
i
s
i
t
i
n
gCh
a
p
t
e
r2
n
i
n
gF
i
g
u
r
e2
.
9
i
tc
b
ei
d
e
n
t
i
f
i
e
d a
tc
o
n
s
i
n
t
s8
1and8
3a
r
ea
c
t
i
v
e(
a
s
ands
c
c
o
n
f
i
r
m
e
dby ezoomeds
o
l
u
t
i
o
ni
nF
i
g
u
2
.
1
0
)
.Thes
o
l
u
t
i
o
ni
sx
T=0
.
6
8m a
n
d
x
i=0
.
6
5m.
l
f8
1a
n
d8
Ja
r
ea
c
t
i
v
ec
o
n
s
E
n

em
L
u

mer
e
a
s
o
n
i
n
gt
h
emu
l
t
i
p

l
i

e
r
sa
s
s
o
c
i
a
t
e
d
wi
i
n
a
c
t
iv
ec
o
n
s
t
r
a
i
n
t
s8
2 d
Byt
h
cs
a
E
4
't
i

S
1
3
2and1
3
4
mustbcs
e
tt
oz
e
r
o
.Th
i
si
n
f
o
r
m
a
t
i
o
nont
h
ea
c
t
i
v
ec
o
n
s
t
r
a
i
n
t
s
c
a
nbeu
s
e
dt
os
o
l
v
ef
o
rx
j
x
i i
si
sas
y
moftwo u
a
t
i
o
n
si
ntwou
n
k
n
o
w
n
s
.
T
h
isd

no

t however

c
o
m
p
l
e
t
e s
f
a
c
t
i
o
no
f eKuhn-T
k
e
rc
o
n
d
i
t
i
o
n
s
.1
3
1
e
.8
2a
n
d8
4mustb
ee
v
a
l
u
a
t
e
d
a
n
ds
Jmustbes
o
l
v
e
da
n
dv
e
r
i
f
i
e
dt
h
a
tt
h
e
ya
r
e v
ya
r
el
e
s
st
h
a
nz
e
r
o
.
a
n
dv
e
r
i
f
i
e
dt
h
a
te
Sec4_5_2.mi
st
h
ec
o
d
es
e
g
m
e
n
ta
ti
su
s
e
dt
os
o
l
v
e p
a
r
t
i
c
u
l
a
rc
a
s
ed
i
s
c
u
s
s
c
d
a
b
o
v
c
.Thcp
r
o
b
l
e
mi
ss
o
l
v
c
ds
y
m
b
o
l
i
c
a
l
l
y
.Twov
c
r
s
i
o
n
soft
h
ep
r
o
b
l
e
ma
r
es
o
l
v
c
d
:
t
h
eo
r
i
g
i
n
a
lp
r
o
b
l
e
mandas
c
a
l
e
dv
e
r
s
i
o
n
.1 o
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
n(
4
.
7
6
)i
slJlo
d
i
f
i
e
d
t
ohaveac
o
e
f
f
i
c
i
e
n
to
fu
n
i
t
y
.T
h
i
ss
h
o
u
l
dn
o
tc
h
g
et
h
ev
a
l
u
eo
ft
h
ed
e
s
i
g
nv
a
r
i
a
b
l
e
s
o
d
ea
n
dv
e
r
i
f
yi
f i
si
si
n
d
e
e
d e
.Notec
(
w
h
y
?
)
.Youc
a
ne
x
p
e
r
i
m
e
n
tw
i
t
h c
a
g
r
a
n
g
em
u
l
t
i
p
l
i
e
r
sd
u
r
i
n
gi
sc
x
e
r
c
i
s
ea
n
di
n
f
e
rt
h
er
e
l
a
t
i
o
nb
e
t
w
e
e
n
v
a
l
u
e
so
ft
h
eL
t
h
em t
i
p
l
i
e
r
sa
n
dt
h
es
c
a
l
i
n
go
f o
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
n
.Th
ef
o
l
l
o
w
i
n
gd
i
s
c
u
s
s
i
o
n
a
s
s
u
m
e
st
h
a
tt
h
ec
o
d
eh
a
sb
e
e
nr
u
na
n
dv
a
l
u
e
sa
v
a
i
l
a
b
l
e
.
R
e
f
e
r
r
i
n
gt
oes
o
l
u
t
i
o
no
ft
h
cc
a
s
ca
b
o
v
e
t
h
eo
p
t
i
m
a
lv
a
)
u
e
sf
o
r d
e
s
i
g
n
v
a
r
i
a
b
l
e
sa
r
ex
j:
:0
.
6
7
7
3m a
n
dx
i=0.6443m.Thisi
sv
e
r
yc
l
o
s
et
ot
h
eg
r
a
p
h
i
c
a
l
h
es
o
l
u
t
i
o
ni
sa
c
h
i
e
v
e
d

t
a
k
en
o
t
eo
ft
h
em
u
l
t
i
p
l
i
e
r
s
:
s
o
l
u
t
i
o
n
.B
e
f
o
r
ewec
o
n
c
l
u
d
e t
=5.2622e-010 d=0
.
12
7
0
.I
ft
h
em
u
l
t
i
p
l
i
e
r
sa
r
en
e
g
a
t
i
v
e
e
ni
ti
sn
o
ta
minimum.Thev
a
l
u
e
so
fec
o
n
s
t
r
a
i
n
t
sa
tt
h
eo
p
t
i
m
a
lv
a
l
u
e
so
ft
h
ed
e
s
i
g
na
r
e8
2=
-4.
5
5
5
3
e+ 007a
n
d8
4=
0.0320.Bothvaluessuggestthatthecons ntsare
i
n
a
c
t
i
v
e
.Atl
e
a
s
tes
o
l
u
t
i
o
ni
sf
e
a
s
i
b
l
e
a
l
lc
o
n
s
i
n
t
sa
r
es
a
t
i
s
f
i
e
d
.Thev
a
l
u
eo
f
s
ti
sz
e
r
oandi
ti
sa
1
s
oa
na
c
t
i
v
ec
o
n

198

199

4
.
5 EXAMPLES

NONL
lNEARPROGRAMMING

c
o
n
s
i
d
e
r
e
da
c
t
i
v
e
.1
es
o
l
u
t
i
o
n

however

v
i
o
l
a
t
t
h
es
i
d
ec
o
n
s
a
i
n
t
s
.
'
h
eg
r
a
p
h
i
c
a
l
s
o
l
u
t
i
o
ni
sr
a
t
h
e
re
x
p
l
i
c
i
ti
ni
d
e
n
t
i
f
y
i
n
gt
h
eminimuma
tepoin
e
s
t
a
b
l
i
s
h
e
di
nt
h
e
f
i
r
s
ts
o
l
u
t
i
o
n
.1
e v
i
a
lc
a
s
eo
ft
h
em
u
l
t
i
p
l
i
e
rb
e
i
n
gz
e
r
owhent
h
ec
o
n
s
t
r
a
i
n
ti
sa
c
t
i
v
e
i
so
b
s
e
r
v
c
dw
i
t
hr
e
s
p
e
c
tt
ot
h
ef
i
r
s
ttwoc
o
n
s
t
r
a
i
n
t
s e
x
p
r
e
s
ss
t
r
e
s
sr
e
l
a
t
i
o
n
s
.l
n
c
r
m
si
nt
h
e
s
ei
n
e
q
u
a
l
i
t
i
e
sh
a
v
el
a
r
g
eo
r
d
e
r
so
fm
a
g
n
i
t
u
d
ee
s
p
e
c
i
a
l
l
yi
f
g
e
n
e
r
a
lct
t
h
e
r
ea
r
eo
t
h
r
e
q
u
a
t
i
o
n
sa
te
x
p
r
e
s
sc
o
n
s
t
r
a
i
n
t
sond
i
s
p
l
a
c
e
m
e
n
t
sa
s
o
rt
h
el
i
k
e
.
L
a
r
g
ev
a
l
u
e
se
x
e
r
tal
a
r
g
ei
n
f
l
u
e
n
c
eont
h
ep
r
o
b
l
e
ms
omucha
st
oc
a
u
s
el
o
w
e
rv
a
l
u
e
c
o
n
s
t
r
a
i
n
t
st
ohavel
i
t
t
l
eo
rnos
i
g
n
i
f
i
n
c
cf
o
r ep
r
o
b
l
e
m
.Th
i
si
su
s
u
a
l
l
yas
e
v
e
r
e
eremedyf
o
ri
si
ss
c
a
l
i
n
go
ft
h
ef
u
n
c
t
i
o
n
s
.
p
r
o
b
l
e
mi
nn
u
m
e
r
i
c
a
li
n
v
e
s
t
i
g
a
t
i
o
n
s
.Th

S
c
a
l
l
n
g
:C
o
n
s
i
d
c
reo
r
d
e
ro
fm
a
g
n
i
t
u
d
ei
nE
q
u
a
t
i
o
n
s(
4
.
7
7
a
) and (
4
.
7
7
c
)
.
v
e
nbyl
a
r
g
e
rm
a
g
n
i
t
u
d
e
s
.l
n
e
q
u
a
l
i
t
y(
4
.
7
7
c
)w
i
l
lb
e
N
u
m
e
r
i
c
a
lc
a
l
c
u
l
a
t
i
o
n
sa
r
e i
i
g
n
o
r
e
di
nr
e
l
a
t
i
o
nt
ot
h
eo
t
h
c
rf
u
n
c
t
i
o
n
se
v
e
nt
h
o
u
g
ht
h
eg
r
a
p
h
i
c
a
ls
o
l
u
t
i
o
ni
n
d
i
c
a
t
e
s
i
sa
c
t
i
v
e
. si
saf
r
e
q
u
e
n
t c
u
r
r
e
n
i
na
l
lk
i
n
d
so
fn
u
m
e
r
i
c
a
lt
h
n
i
q
u

.Th
e
m
p
a
c
to
fl
a
r
g
ev
a
r
i
a
t
i
o
n
si
nm
a
g
n
i
t
u
d
e
samong
s
t
a
n
d
a
r
da
p
p
r
o
a
c
ht
om
i
n
i
m
i
z
e ei
d
i
f
f
e
1
te
q
u
a
t
i
o
n
si
st
on
o
r
m
a
l
i
z
eer
e
l
a
t
i
o
n
s
.I
np
r
a
c
t
i
c
et
h
i
si
sa
l
s
oe
x
t
e
n
d
e
dt
oe
si
sr
e
f
e
r
r
e
dt
oa
ss
c
a
l
i
n
gl
h
ev
a
r
i
a
b
l
e
sands
c
a
l
i
n
gt
h
ef
u
n
c
l
i
o
n
s
.Many
v
a
r
i
a
b
l
e
s
. i
c
u
r
r
e
n
ts
o
f
t
w
a
r
ewi
1
1s
c
a
l
et
h
ep
r
o
b
lmw
i
t
h
o
u
tu
s
e
ri
n
t
e
r
v
e
n
t
i
o
n
.
i
n
gV
a
r
i
a
b
l
e
s
:1
ep
r
c
s
e
n
c
co
fs
i
d
ec
o
n
s
t
r
a
i
n
t
si
np
r
o
b
l
e
mf
o
r
m
u
l
a
t
i
o
na
l
10ws
S l
an
a
t
u
r
a
1d
c
f
i
n
i
t
i
o
no
fs
c
a
l
e
dv
a
r
i
a
b
J
e
s
.Theu
s
e
r
d
e
f
i
n
e
du
p
p
e
ra
n
dl
o
w
e
rboundsa
r
e
.Th
e
r
e
f
o
r
e
u
s
e
dt
os
c
a
l
ee
a
c
hv
a
r
l
a
b
l
eb
e
t
w
e
e
n0 d1
-X~

=--'; x
:=scaledithvariable

chi
t
e
r
a
t
i
o
nu
s
u
a
l
l
yegr
i
e
n
to
ft
h
ef
u
n
c
t
i
o
n
sa
tt
h
cc
u
r
r
e
n
tv
a
l
u
eo
ft
h
ed
e
s
i
g
n
v
a
r
i
a
b
l
e
si
si
n
v
o
l
v
e
di
nt
h
ec
a
l
c
u
l
a
t
i
o
n
s
.T
h
isg
r
a
d
i
e
n
t
exp
s
s
e
d am
a
t
r
i
x
i
sc
a
1l
e
d
eJ
a
c
o
b
i
m
a
t
r
i
xo
rs
i
m
p
l
yt
h
eJ
a
c
o
b
i
a
l
l
.S
o
p
h
i
s
t
i
c
a
t
e
ds
c
a
l
i
n
gt
e
c
h
n
i
q
u
e
s[
7
8
]
employt
h
ed
i
a
g
o
n
a
len
e
so
ft
h
i
sm
a
t
r
i
xa
sm
e
t
r
i
c
st
os
c
a
l
et
h
er
e
s
p
e
c
t
i
v
ef
u
n
c
t
i
o
n
s
.
Th
eseen
aree
v
a
l
u
a
t
e
da
tt
h
es
t
a
r
t
i
n
gv
a
l
u
e
so
ft
h
ed
e
s
i
g
nv
a
r
i
a
b
l
e
s
.Th
ef
u
n
c
t
i
o
n
es
ca
1
edi
nesamemannera
sE
q
u
a
t
i
o
n
s(
4
.
8
1
)and(
4
.
8
2
)
.F
o
rt
h
ef
o
r
m
e
r
c
a
na
l
s
ob
a
ne
x
p
e
c
t
e
dl
o
w
e
randu
p
p
e
rv
a
l
u
e
so
ft
h
ec
o
n
s
t
r
a
i
n
t
sa
r
n
e
c
e
s
s
a
r
y
.I
nt
h
i
se
x
e
r
c
i
s
c
ec
o
n
s
t
r
a
i
n
t
sw
i
l
lbes
c
a
l
e
du
s
i
n
gr
e
l
a
t
i
o
n
ss
i
m
i
l
a
rt
ot
h
er
e
l
a
t
i
o
n
se
x
p
r
e
s
s
e
dby
u
a
l
i
t
i
e
s(
4
.
8
2
)
.Th
es
c
a
l
i
n
gf
a
c
t
o
rf
o
re
a
c
hc
o
n
s
n
tw
i
l
lb
ed
e
t
e
r
m
i
n
e
du
s
i
n
gE
s
t
a
r
t
i
n
gv
a
l
u
eo
rt
h
ei
n
i
t
i
a
lg
u
e
s
sf
o
rt
h
ev
a
r
i
a
b
l
e
s
.A s
t
a
r
t
i
n
gv
a
l
u
eo
f0
.
6f
o
rb
o
t
h
d
e
s
i
g
nv
a
r
i
a
b
l
e
si
ss
e
l
e
c
t
e
dt
ocomputeev
a
l
u
e
sn
c
e
s
s
a
r
yf
o
rs
c
a
l
i
n
gt
h
ef
u
n
c
t
i
o
n
s
.
Thes
c
a
l
i
n
gc
o
n
s
t
a
n
t
sf
o
rt
h
eeq
o
n
sa
r
ec
a
l
c
u
l
a
t
e
d

O=2
93888.
4
; g
2
0
=102561
.
12
; g
3
0
=1 70E-03; g
4
0
=1
1
ef
i
r
s
tt
h
r
e
ec
o
n
s
t
r
a
i
n
t
sa
r
ed
i
v
i
d
e
dt
h
r
o
u
g
hbyt
h
e
i
rs
c
a
l
i
n
gc
o
n
s
t
a
n
t
s
.Thel
a
s
t
u
a
t
i
o
ni
su
n
c
h
a
n
g
e
d
. 1
eo
b
j t
i
v
ef
u
n
c
t
i
o
nh
a
sac
o
e
f
f
i
c
i
e
n
to
fo
n
e
.Thes
c
a
l
e
d

p
r
o
b
l
e
mi
s
Minimize

1 4

(
4
.
8
3
)

8
.
:2.5509~+0.136lx.-33.148( -xj) O

(
4
.8
4
a
)

(
4
.
8
1
a
)

I
f
0
.
8
8+1
.
4
6
1
9
x

)
(
+x+) 166.
5641
(
x
i-x~) ~ 0 (
4
.8
4
b
)
2:(

(
48
I
b
)

:1
.086
8
x
.+0
.
3
4
8
2-28.
4
6
4
1(x1-x~) ~ 0

(
4
.
8
4
c
)

C4:X2-XI+
0.001~O

(
4
.8
4
d
)

Xi-Xi

=
x
;
(
.
x
:
)+
X
:

I
nt
h
eo
r
i
g
i
n
a
lp
r
o
b
l
e
mE
q
u
a
t
i
o
n(
4
.
8
1b
)i
su
s
e
dt
os
u
b
s
t
i
t
u
t
ef
o
rt
h
eo
r
i
g
i
n
a
lv
a
r
i
a
b
l
e
s
b
ee
x
p
r
e
s
s
e
di
nt
e
r
m
so
fs
c
a
l
e
dv
a
r
i
a
b
l
e
.Ana
l
t
e
m
a
t
e
a
f
t
e
rwhicht
h
ep
r
o
b
l
e
mc
1
yt
h
eu
p
p
e
rv
a
l
u
eo
ft
h
es
i
d
ec
o
n
s
t
r
a
i
n
tt
os
ca
1
et
h
cd
e
s
i
g
n
f
o
r
m
u
l
a
t
i
o
ni
st
ou
s
eon
v
a
r
i
a
b
l
e
:

s
i
=
x
i
/
4
e

(
4
.
8
2
a
)

x
j
=
x
j

(
4
.
8
2
b
)

sl
i
m
i
t
s h
i
g
h
e
rs
c
a
l
e
dva
1uet
o1

i
td
o
e
sn
o
ts
e
tel
o
w
e
rs
c
a
l
e
dv
a
l
u
et
o
While i
fi
ss
e
c
t
i
o
nt
h
e
r
ei
snon
e
c
e
s
s
i
t
yf
o
rs
c
a
l
i
n
gt
h
ed
e
s
i
g
n
z
e
r
o
.F
o
reexampleo
v
a
r
i
a
b
l
e
ss
i
n
c
et
h
e
i
ro
r
d
e
J
'o
fm
a
g
n
i
t
u
d
ci
sone
w
h
i
c
hi
se
x
a
c
t
l
ywhats
ca
1
i
n
ga
t
l
e
m
p
t
s
t
oa
c
h
i
e
v
e
.

S
c
a
l
i
n
gt
h
eC
o
n
s
t
r
a
i
n
t
s
:S
c
a
l
i
n
goft
h
ef
u
n
c
t
i
o
n
si
nep
r
o
b
l
e
mi
su
s
u
a
l
l
yc
r
i
t
i
c
a
l
f
o
ras
u
c
c
e
s
s
f
u
ls
o
l
u
t
i
o
n
.N
u
m
e
r
i
c
a
lt
e
c
h
n
i
q
u
c
su
s
e
di
no
p
t
i
m
i
z
a
t
i
o
na
r
ei
t
e
r
a
t
i
v
e
.I
n

0
.
0
2~ XI~ 1
.0
;

0
.
0
25
;x2~ 1
.0

(
4
.
7
9
)

Sec4_5_2_scaled.m: The c
o
d
e(
c
o
u
l
dh
a
v
eb
e
c
n p
a
r
to
fS
e
c
4
_
5
_
2
.
m
)
h
eo
r
i
g
i
n
a
lp
r
o
b
l
e
mwas
i
n
v
e
s
t
i
g
a
t
e
st
h
es
c
a
l
e
dp
r
o
b
l
e
mi
n sameway tt
i
n
v
e
s
t
i
g
a
.P
r
i
m
a
r
i
l
yi
se
x
p
l
o
i
t
s i
n
f
o
r
m
a
t
i
o
nfromt
h
eg
r
a
p
h
i
c
a
ls
o
I
o
n
.1
ti
s
e
x
p
e
c
t
e
dtt
h
es
o
l
u
t
i
o
nf
o
rt
h
ed
e
s
i
g
nv
a
r
i
a
b
1
e
sw
i
l
lbet
h
esamca
l
t
h
o
u
g
ht
h
e
1t
i
p
l
i
e
rv
a
1u
e
sa
r
ee
x
p
e
c
t
e
dt
obed
i
f
f
e
r
e
n
.
tFromt
h
ei
n
f
o
r
m
a
t
i
o
ni
neWorkspace
mu
windowt
h
eo
p
t
i
m
a
lv
a
l
u
e
sf
o
rt
h
ed
e
s
i
g
nv
a
r
i
a
b
l
e
sa
r
ex
i 0
.
6
7
7
4a
n
dx
i 0.6445.
lo
f c
o
n
s
t
r
a
i
n
t
sa
r
ef
e
a
s
i
b
l
e
.Theo
p
t
i
ma
1v
a
l
u
e
so
fem
u
1
t
i
p
l
i
e
r
sh
a
v
ec
h
a
n
g
e
d
.
AI
The
i
snol
o
o
g
e
raz
e
r
om
u
l
t
i
p
I
i
e
r
i
th
a
stumedp
o
s
i
t
i
v
e
.Th
e
r
ei
ss
t
i
l
lo
n
e
mu
1t
i
p
l
i
e
rw
i
t
han
e
g
a
t
i
v
es
i
g
ni
m
p
l
y
i
n
gt
h
ep
o
i
n
th
a
sn
o
tmet KuhnT
u
c
k
e
r

i
ts
h
o
u
l
dh
a
v
eb
e
e
na
n
t
i
c
i
p
a
t
e
ds
i
n
c
ec
h
a
n
g
i
n
go
f
l
h
es
i
g
l
lo
f
t
h
e
c
o
n
d
i
t
i
o
D
S
.A
c
t
u
a
l
l
y
m
u
l
t
i
p
l
i
e
ri
sn
o
tp
o
s
s
i
b
l
ew
i
t
hs
c
a
l
i
n
g
.

200

PROBLEMS

NONL
lNEARPROGRAMMING

Th
e
r
ei
ss
t
il
It
h
em
a
t
t
e
ro
fi
n
c
o
n
s
i
s
t
e
n
c
yb
e
t
w
nt
h
eg
r
a
p
h
i
c
a
ls
o
l
u
t
i
o
nandt
h
e
d
i
s
s
a
t
i
s
f
a
c
t
i
o
no
ft
h
eKuhnTuckerc
o
n
d
i
t
i
o
n
s
.E
x
p
l
o
r
i
n
gt
h
i
sp
a
r
t
i
c
u
l
a
rexampleh
a
s
b
e
e
nv
e
r
yt
o
r
t
u
r
o
u
sb
u
th
a
sp
r
o
v
i
d
e
deo
p
p
o
r
t
u
n
i
t
yt
oe
x
p
l
o
r
ei
m
p
o
r
t
a
n
tb
u
tr
e
l
a
t
e
d
ik
cs
c
a
l
i
n
g
.A
n
o
t
h
e
ri
m
p
o
r
t
a
n
tc
o
n
s
i
d
e
r
a
t
i
o
ni
s e s
c
o
p
co
ft
h
e
q
u
a
n
t
i
t
i
e
sI
Kuhn-Tuckerc
o
n
d
i
t
i
o
n
s
.Theseconditionsapp
onlyatreglllarpoints.
RegularP
o
l
n
t
s
:Regularp
o
i
n
t
s[
8
]a
r
i
s
ewhene
q
.
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
t
sh
(
X
)a
r
e
e
n
t e
p
r
o
b
1
e
m
.l
e
y
a
r
a
1
s
oe
x
t
e
n
d
e
dt
oa
c
t
i
v
ei
n
Iit
y
c
o
n
s
n
t
s
(
p
s
e
u
d
o
P s
n
t
s
)
.Kuhn-Tuckcrc
o
n
d
i
t
i
o
n
sa
r
cv
a
l
i
do
n
1
yf
o
rr
e
g
u
1
a
rp
o
i
n
t
s
. l
e
e
q
u
a
1
i
t
yc
o
n
s
1f
e
a
t
u
r
e
so
far
e
g
u
l
a
rp
o
i
n
tX.a
r
e
twoc
s
s
c
n
t
ia

i
sf
e
a
s
i
b
l
e(
s
a
t
i
s
f
i
e
sa
l
Ic
o
n
s
t
r
a
i
n
t
s
)
.
ThcpointX

2
0
1

4
.MATLAB
Demos S
y
m
b
o
l
i
cT
o
o
lB
o
x
I
n
t
r
o
d
u
c
t
i
o
n
.o
n
l
i
n
er
e
s
o
u
r
c
ei
n MATLAB.

M
a
t
h
W
o
r
k
sI
n
c
.
MA.
.L
.a
n
dF
a
i
r
e
s
J
.0
.
N
u
m
e
r
i
c
a
l '
s
i
s
.4
t
he
d
.
.PWS_KENTP
u
b
l
i
s
h
i
n
g
5
.B
u
r
d
e
nR
Company
B
o
s
t
o
n
.(
1
9
8
9
)
R
.L
.
.OptimiwtionMethodsforEl
I
g
i
n
e
e
r
i
n
gD
e
s
i
g
n
.A
d
d
i
s
o
n
W
e
s
l
e
y
R
e
a
d
i
n
g
6
.F
o
x
MA.(
1
9
7
1
)
7
.V d
e
r
p
l
a
a
t
sG
.N
.N/lm
e
r
i
c
a
lO
p
t
i
m
i
w
t
i
o
nT
e
c
h
l
l
i
q
u
e
sf
o
rEl
Ig
i
n
e
e
r
i
n
gD
e
s
i
g
l
l
.
McGmwH
i
1
1

NewY
o
r
k
.(
1
9
8
4
)
8
.A
r
or
a
J
.S
.
l
n
t
r
o
d
u
c
t
i
o
nt
oO
p
t
i
m
a
ID
e
s
i
g
n
.Mc
Gm
wH
il
NewY
o
r
k
.(
19
8
9
)
9
.Kuhn
H
.W.a
n
dT
u
c
k
c
r

A
.W.
N
o
n
l
u
l
e
a
rP
r
o
g
r
a
m
m
i
n
g
.P
r
o
c
e
e
d
i
n
g
s0
1t
h
eS
e
c
o
n
d
B
e
r
k
e
l
e
ySymposiumonM
a
t
h
e
m
a
t
i
c
a
lS
t
a
t
i
s
t
i
c
sa
l
l
dProbab
i
/
i
.
J
.Ncyman(
c
d
.
)
U
n
i
v
c
r
s
i
t
yo
fC
a
l
i
f
o
r
n
i
aP
r
c
s
s
1
9
5
1
.

. Th
cg
r
a
d
i
e
n
t
so
ft
h
e ua
1i
t
yn
s
t
r
i
n
t
s wc
l
Ia
st
h
ea
c
t
i
v
ei
n
l
i
t
y
i
n
d
e
p
e
n
d
e
n
ts
e
to
fv
e
c
t
o
r
s
.
c
o
n
s
t
r
a
i
n
t
sa
tX.mustf
o
r
mal
i
n
e

I
nt
h
es
c
a
1
e
dproblemo
fExamp1c4.
3t
h
cc
o
n
s
t
r
a
i
n
t
sgIand8
3a
r
ea
c
t
i
v
e
.The
j=0
.
6
7
7
4andx
i=0
.
6
4
4
5i
sf
e
a
s
i
b
l
e
. The g
r
a
d
i
e
n
to
fgI i
s Vg.=
s
o
l
u
t
i
o
nx
[-37.631435.
454T
.Theg
r
a
d
i
e
n
to
fg
2i
sVg1 [
3
4
.
3
1
1
930.
4
7
8
6
]
T
.l
ta
p
p
e
a
r
s
1c
lt
oe
a
c
ho
t
hr
.T
h
i
si
sa
l
s
oc
v
i
d
e
n
ti
nt
h
e
t
h
a
tt
h
etwog
r
a
d
i
c
n
t
sa
r
ca
1
m
o
s
tp
a
r
al
g
r
a
p
h
i
c
a
ls
o
l
u
t
i
o
n
. The g
r
a
d
i
e
n
t
st
h
e
r
e
f
o
r
ca
r
en
o
tI
in
e
a
r
l
y indepcndcntKuhn-Tuckcrc
o
n
d
i
t
i
o
n
sc
a
n
n
o
tb
ca
p
p
l
i
e
da
tt
h
i
sp
o
i
n.
tT
h
i
sd
o
e
sn
o
tmeant
h
e
r
i
sg
r
a
p
h
i
c
a
le
v
i
d
e
n
c
et
h
a
ti
ti
si
n
d
e
e
ds
o
.
p
o
i
n
ti
sn
o
tal
o
c
a
lminimum.Th

Tuckera
r
et
h
eon
1
ya
v
a
i
l
a
b
l
ef
o
r
ma
1c
o
n
d
i
t
i
o
n
sf
o
rr
e
c
o
g
n
i
t
i
o
no
ft
h
e
Kuhn
optimum v
a
1
u
e
s
.I
ti
su
n
i
v
c
r
sa
1l
ya
p
p
l
i
e
dw
i
t
h
o
u
tr
e
g
a
r
dt
or
e
g
u
l
a
r
i
.
Some
1c
o
n
s
i
d
c
r
a
t
i
o
n
ss
h
o
u
l
db
ck
e
p
ti
nmind[
8
]
.
a
d
d
i
t
i
o
na

l
1o
fei
n
e
q
ua
1i
t
yc
o
n
s
'
a
i
n
t
sa
r
e
I
fe
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
t
sa
r
ep
r
e
s
e
n
ta
n
da
nt
h
ep
o
i
n
t
ss
a
t
i
s
f
y
i
n
g cKuhnTucker c
o
n
d
i
t
i
o
n
s may be
i
n
a
c
t
i
v
ee
minimum
maximum
o
ras
a
d
d
1
ep
o
i
nt
.H
i
g
h
c
r
o
r
d
e
rc
o
n
d
i
t
i
o
n
sa
r
en
e
c
c
s
s
a
r
y
t
oi
d
c
n
t
i
f
yt
h
ct
y
p
eo
fs
o
l
u
t
i
o
n
.
I
ft
h
em
u
1
t
i
p
l
i
e
r
so
ft
h
ea
c
t
i
v
ei
n
e
q
u
a
1
i
t
yc
o
n
s
t
r
n
t
sa
r
ep
o
s
i
t
i
v
e
t
h
ep
o
i
n
t
c
a
n
n
o
tb
ea1
0
c
a
lmaximum.I
tmayn
o
tb
eal
o
c
a
lminimume
i
t
h
e
r
.Ap
o
i
n
tmay
u
l
t
i
p
l
i
c
ro
fa
na
c
t
i
v
ci
n
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
ti
sz
e
r
o
.
b
camaximumi
f m

REFERENCES
S
.K
.C
a
l
c
u
l
u
sandA
n
a
l
y
t
i
c
a
lG
e
o
m
e
t
r
y
.McGmw-H

I
i NewY
o
r
k
.(
1
9
8
7
)
1
.S
t
c
i
n
2
.H
o
s
t
e
t
l
e
r
G
.H
.S
a
n
t
i
n
a
M.S
.a
n
dM
o
n
t
n
l
v
oP
.0
.A
n
a
l
y
t
i
c
a
l
.N
w
n
e
r
i
c
a
l
.and
C
o
m
p
u
t
a
t
i
o
n
a
lMethodsf
o
rS
c
i
e
n
c
eandE
n
g
i
n
e
e
r
i
n
g
.P
r
e
n
t
i

H
a
1
I
.E
n
g
l
e
w
o
o
dCI
i

N
J
.(
19
9
1
)

C
.a
n
dC
o
s
t
a
.P
.J
.
S
y
m
b
o
l
i
cMathToolbox-forew
i
t
hMATUB-UsersG
u
i
d
e
3
.M
o
l
c
r
M
a
t
h
W
o
r
k
sI
n
c
.
.MA.

PROBLEMS
(
I
n many o
ft
h
ep
r
o
b
l
e
m
s be
1owyou a
r
er
e
q
u
i
r
e
dt
oo
b
t
a
i
nt
h
en
u
m
e
r
i
c
a
l
s
o
l
u
t
i
o
n
.
)
4
.
1 D
e
f
i
n
ctwon
o
n
l
i
n
e
a
rf
u
n
c
t
i
o
n
si
ntwov
a
r
i
a
b
l
e
s
.F
i
n
dt
h
c
i
rs
o
l
u
t
i
o
nt
h
r
o
u
g
h
c
o
n
t
o
u
rp
l
o
t
s
.
2 Fort
h
cf
u
n
c
t
i
o
n
si
nP
r
o
b
l
e
m4
.
1o
b
t
a
i
nt
h
eg
r
a
d
i
e
n
t
so
ft
h
ef
u
n
c
t
i
o
na
n
dU
I

4.
J
a
c
o
b
i
a
n
.Confirmthemu
s
i
n
gt
h
eS
y
m
b
o
l
i
cMathT
o
o
l
b
o
x
.

3 D
e
f
i
n
et
h
ed
e
s
i
g
ns
p
a
c
e(
c
h
o
o
s
e
ss
i
d
ec
o
n
s
t
r
a
i
n
t
s
)f
o
rat
w
o
v
a
r
i
a
b
l
ep
r
o
b
l
c
m
.
4.
1
in
e
a
rf
u
n
c
t
i
o
n
si
ntwov
a
r
i
a
b
l
e
st
h
a
tdon
o
th
a
v
eas
o
l
u
t
i
o
n
D
c
f
i
n
etwonon
w
i
t
h
i
nt
h
i
ss
p
a
c
e
.G
r
a
p
h
i
c
a
1l
yc
o
n
f
i
r
mt
h
r
c
s
u
l
.
t
4
.
4 D
e
f
i
n
e ad
e
s
i
g
ns
p
a
c
ef
o
rat
w
o
v
a
r
i
a
b
1
ep
r
o
b
l
e
m
.D
e
f
i
n
c two n
o
n
l
i
n
e
a
r
a
v
ea
tl
e
a
s
ttwos
o
l
u
t
i
o
n
sw
i
t
h
i
nes
p
a
c
c
.
f
u
n
c
t
i
o
n
si
ntwov
a
r
i
a
b
l
e
s h
4.
5 D
e
f
i
n
ean
o
n
l
i
n
e
a
rf
u
n
c
t
i
o
no
ftwov
a
r
i
a
b
l
e
s
.Chooseac
o
n
t
o
u
rv
a
l
u
ea
n
ddraw
t
h
cc
o
n
t
o
u
r
.I
d
c
n
t
i
f
yap
o
i
n
tont
h
cc
o
n
t
o
u
r
.C
a
l
c
u
l
a
t
et
h
ev
a
1
u
co
ft
h
eg
r
a
d
i
e
n
t
.Drawt
h
eg
r
d
i
e
n
ta
tt
h
ep
o
i
n
tu
s
i
n
gt
h
ecomputedv
a
l
u
e

C
a
l
c
u
l
a
t
c
a
tt
h
a
tp
o
i
nt
H
e
s
s
i
ca
b
o
v
ep
o
i
n
.
t
4
.
6 U
s
i
n
gt
h
er
e
l
n
t
i
o
ni
nEq
u
a
t
i
o
n(
4
.
2
3
)
e
s
t
a
b
l
i
s
h tt
h
eg
r
a
d
i
e
n
ti
sn
o
r
m
a
lt
ot
h
e
t
a
n
g
e
n
t[
ap
h
y
s
i
c
a
li
n
t
e
r
p
r
e
t
a
t
i
o
nf
o
rE
q
u
a
t
i
o
n(
4
.
2
3
)
]
.
4
.
7 D
e
f
i
n
ean
o
n
l
i
n
e
a
rf
u
n
c
t
i
o
no
f v
a
r
i
a
b
l
e
s
.Chooseap
o
i
n
ti
nt
h
ed
e
s
i
g
n
r
a
d
i
e
n
to
f f
u
n
c
t
i
o
na
tt
h
ep
o
i
n
.
tC
a
1
c
u
l
a
t
et
h
eH
e
s
s
i
a
n
s
p
a
c
e
.F
i
n
deg
m
a
t
r
i
xa
tt
h
esamcp
o
i
nt
.

4
.
8 E
x
p
r
c
s
st
h
eT
a
y
l
o
rs
e
r
i
e
se
x
p
a
n
s
i
o
n(
q
u
a
d
r
a
t
i
c
)o
ft
h
ef
u
n
c
t
i
o
n
f
t
x
) (
2-3x+
x
2
)s
i
nxa
b
o
u
t p
o
i
n
tx=
0
.
7
0
7
.Confirmy
o
u
rr
e
s
u
l
t
st
h
r
o
u
g
ht
h
eS
y
m
b
o
l
i
c
MathT
o
o
l
b
o
x
.P
l
o
tt
h
eo
r
i
g
i
n
a
lf
u
n
c
t
i
o
nandea
p
p
r
o
x
i
m
a
t
i
o
n
.

9 Expandt
h
ef
u
n
c
t
i
o
n
f
t
x
.
y
)=1
0(
1-x
2
)
2+(
y-2
)
2q
u
a
d
r
a
t
i
c
a
1l
ya
b
o
u
tt
h
ep~int
4.
(
1
1
)
.Howwi
l
1youdisplayt
h
ei
n
f
o
r
m
a
t
i
o
n
?Drawt
h
ec
o
n
t
o
u
r
so
ft
h
eo
r
i
g
i
na
1
f
u
n
c
t
i
o
na
n
dt
h
ea
p
p
r
o
x
i
m
a
t
i
o
n
.

202 NONL
lN
E
A
RPROGRAMMING

q
u
a
t
i
o
n(
4
.
1
0
)i
sa
d
d
e
dt
o
4
.
1
0O
b
t
a
i
nt
h
es
o
l
u
t
i
o
ng
r
a
p
h
i
c
a
l
l
yf
o
r
t
h
e
c
a
s
ewhenE
p
r
o
b
l
e
md
e
f
i
n
e
db
yE
q
u
a
t
i
o
n
s(
4
.
3
6
)
(
4
.
3
9
)
and(
4
.
3
7
)
.O
b
t
a
i
nes
o
l
u
t
i
o
n
y
m
b
o
l
i
cMathT
o
o
l
b
o
x
.
u
s
i
n
geS

4
.
1
1O
b
t
a
i
ncs
o
l
u
t
i
o
nt
ot
h
co
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
e
mi
nS
c
c
t
i
o
n4.
4
.
4u
s
i
n
gMATLAB.
4
.
1
21
nS
e
c
t
i
o
n4
.
5
.
1o
b
t
a
i
nt
h
ec
o
e
f
f
i
c
i
e
n
t
sf
o
raqua
Cf
i
tt
oep
r
o
b
l
e
m
.
4
.
1
3 Usea
nexamplet
oshow optimumv
a
l
u
e
sf
o
rt
h
ed
e
s
i
g
nv
a
r
i
a
b
l
e
sdon
o
t
c
b
g
e t
h
eo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
ni
sm
u
l
t
i
p
l
i
e
dbyac
o
n
s
t.P
r
ov
et
h
esamei
fa
u
n
c
t
i
o
n
.
c
o
n
s
t
a
n
ti
sa
d
d
e
dt
oef
4
.
1
4 Howd
o
e
ss
c
a
Ii
n
go
ft
h
eo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
na
f
f
c
c
tt
h
ev
a
l
u
co
fem
u
l
t
i
p
l
i
e
r
s
?
u
lt
.Howd
o
e
ss
c
a
l
i
n
gt
h
ec
o
n
s
t
r
a
i
n
ta
f
f
e
c
tt
h
e
Usca
ncxamplct
oi
n
f
e
rt
h
e s
m
u
l
t
i
p
l
i
s?
Checkw
i
t
ha
ne
x
a
m
p
l
e
.

NUMERICALTECHNIQUE8THEONE-DIMENSIONAL
PROBLEM

'
heo
n
e
d
i
m
e
n
s
i
o
n
a
lu
n
c
o
n
s
t
r
a
i
n
c
do
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
e
mi
so
f
t
e
ni
n
t
r
o
d
u
c
e
d d
s
o
l
v
e
dn
u
m
e
r
i
c
a
l
l
yi
nmostc
o
u
r
s
e
sa
td
e
a
Iw
i
t
hn
u
m
e
r
i
c
a
la
n
a
l
y
s
is
J
t
e
c
h
n
i
q
u
e
s
.I
t
n
o
t
a
b
l
ya
sf
i
n
d
i
n
gt
h
er
o
o
to
fan
o
n
l
i
n
e
a
re
q
u
a
t
i
o
n
whichi
se
a
p
p
e
a
r
si
n
d
i
r
e
c
t
l
y
samea
ss
a
t
i
s
i
n
geFOCf
o
ra
nu
n
c
o
n
s
t
r
a
i
n
e
do
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
e
m
.Twoo
ft
h
e
a
rt
e
c
h
n
i
q
u
e
sa
r
eeNewton-Ra
p
h
s
o
na
n
deb
i
s
c
c
t
i
o
nt
e
c
h
n
i
q
u
e
.
'
h
e
y E
pop
i
n
t
r
o
d
u
c
e
di
nS
e
c
t
i
o
n5
.
2t
op
r
o
v
i
d
e ac
o
m
p
a
r
i
s
o
nw
i
t
ht
h
em
e
t
h
o
d
su
s
e
di
n
o
p
t
i
m
i
z
a
t
i
o
n
.I
ne
n
g
i
n
e
e
r
i
n
gd
c
s
i
g
n es
i
n
g
l
e
v
a
r
i
a
b
l
eo
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
e
mi
s
p
r
o
b

l
ya v
i
a
l
i
t
yb
u
toned
im
e
n
s
i
o
n
a
lo
p
t
i
m
i
z
a
t
i
o
ni
sac
r
i
t
i
c
a
Icomponento
f
mt
i
v
a
r
i
a
b
l
ed
e
s
i
a
n
di
sd
i
s
c
u
s
s
e
di
nSon5.
3
.M
eanwhilet
h
ec
I
a
s
s
i
c
a
lr
o
o
t
f
i
n
d
i
n
gm
e
t
h
o
d
si
d
e
n
t
i
f
i
e
da
b
o
v
edon
o
tp
l
a
ya
n
ys
i
g
n
i
f
i
c
a
n
tp
a
r
ti
nn
u
m
e
r
i
c
a
l
o
p
t
i
m
i
z
a
t
i
o
nt
e
c
h
n
i
q
u
e
sa
st
h
e
ya
r
ec
o
n
s
i
d
e
r
e
dc
o
m
p
u
t
a
t
i
o
n
a
l
l
ye
x
p
e
n
s
i
v
ea
n
dn
o
t
r
o
b
u
s
tenougho
v
e
ral
a
r
g
ec
I
a
s
so
fp
r
o
b
l
e
m
s
.Ani
m
p
o
r
t
a
n
tc
o
n
s
i
d
c
r
a
t
i
o
ni
nno
n
I
i
n
e
a
r
d
e
s
i
g
no
p
t
i
m
i
z
a
t
i
o
ni
st
h
a
td
u
r
i
n
gt
h
ei
n
i
t
i
a
li
t
e
r
a
t
i
o
n
s

whenyoua
r
et
y
p
i
c
a
l
l
yf
a
raway
omes
o
l
u
t
i
o
na
c
c
u
r
a
c
yc
b
cd
e
df
o
rs
p
e
e
d
.I
nt
h
e
s
ei
t
e
r
a
t
i
o
n
si
ti
sa
l
s
o
c
x
li
t
c
r
a
t
i
o
n
i
n
s
t
c
a
do
ff
a
l
t
e
r
i
n
ga
tec
u
r
r
e
n
to
n
e
.
i
m
p
o
r
t
a
n
tt
omovet
h
r
o
u
g
ht
ocn
u
i dt
h
a
tt
h
co
n
e
.
d
i
m
e
n
s
i
o
n
a
lt
e
c
h
n
i
q
u
e
sb
es
i
m
p
l
ei
nc
o
n
c
e
p
t
Th
e
s
e
.
c
r
i
t
e
r
i
ah
a
v
e q
a
sw
e
l
la
se
a
s
i
l
yi
m
p
l
e
m
c
n
t
a
b
l
e
. Two d
i
f
f
e
r
e
n
tt
e
c
h
n
i
q
u
e
st
h
ep
o
l
y
n
o
m
i
a
l
l
a
t
i
o
nandeg
o
l
d
e
ns
e
c
t
i
o
n
p
r
o
v
i
d
ep
o
p
u
l
a
rs
u
p
p
o
r
tf
o
rm
o
s
to
p
t
i
m
i
z
a
t
i
o
n
a
p
p
r
o
x
s
o
f
t
w
e
.Veryo
f
t
e
nt
h
etwot
e
c
h
n
i
q
u
e
sa
r
ei
m
p
l
e
m
e
n
t
e
di
nc
o
m
b
i
n
a
t
i
o
n
.Th
e
ya
r
e
d
i
s
c
u
s
s
e
di
nS
e
c
t
i
o
n5
.
2
.
203

5
.
2 SOLUnONTOT
H
EP
R
O
B
L
E
M 205

204 NUME
I
C
A
LT
E
C
H
N
I
O
U
E5-T
H
EO
N
E
.
O
I
M
E
N
S
I
O
N
A
LP
R
O
B
L
E
M
5
.
1 PROBLEMDEF
lNITION
I
no
r
d
e
rt
oc
o
n
n
e
c
tw
i
t
ht
h
ed
i
s
c
u
s
s
i
o
n
sl
a
t
e
r
t
h
eo
n
e
d
i
m
e
n
s
i
o
n
a
lv
a
r
i
a
b
l
ew
i
l
lb
e
rX
II
nt
h
ef
o
l
1owingc
h
a
p
t
e
r
si
si
sa
1
s
or
e
f
e
r
r
e
dt
on
s
i
d
e
n
t
i
f
i
e
dn
s

i
n
s
t
e
a
do
fX o
t
h
eo
n
c
d
i
m
e
n
s
i
o
n
a
ls
t
e
p
s
i
z
ec
o
m
p
u
t
a
t
i
o
n
.

Th
es
e
c
o
n
ds
o
l
u
t
i
o
ni
su
s
e
f
u
lo
n
l
yi
fe
x
p
r
e
s
s
i
o
n(
5
.
3
)i
sa
na
c
t
i
v
ec
o
n
s
t
r
a
i
n
tO
t
h
e
r
w
i
s
e
t
h
es
o
l
u
t
i
o
nt
oExample5
.
1wi
l
1p
r
e
v
a
i
1
.
l
ef
i
n
a
ls
o
l
u
t
i
o
nc
h
o
s
e
ne
n
s
u
s
e
x
p
r
e
s
s
i
o
n
s
(
5.
1
)
(
5.
3
)a
r
es
a
t
i
s
f
i
e
d
.Ag
r
a
p
h
i
c
a
li
n
v
e
s
t
i
g
a
t
i
o
ni
nen
e
x
ts
e
c
t
i
o
ns
h
o
u
l
di
l
l
u
s
t
r
a
t
e
t
h
i
sf
e
a
t
u
r
ec
1
e
a
r
l
y
.

lONTOTHEPROBLEM
5
.
2 SOLUT

Ex
ample5
.
1 Thi
sexamplei
ss
i
m
i
l
a
rt
oonei
nC
h
a
p
t
e
r4
:
Minimize

f
(
)=(_1
)
2
(

-2)(-3
)

(
5
.
1
)

S
u
b
j
e
c
t
t
o
:

0SS4

(
5
.
2
)

Th
es
o
l
u
t
i
o
nt
oExamples5
.
1 d5
.
1
a
i
sexaminedgr
p
h
i
c
a
I
I
y
.N
e
x
tt
h
ec
1n
s
s
i
c
a
lr
o
o
t
f
i
n
d
i
n
gt
e
c
i
q
u
c
sofNewton-Raphsona
n
deb
i
s
e
c
t
i
o
nmethoda
r
ep
r
e
s
e
n
t
e
d
.The
s
e
M f
o
l
l
o
w
e
dbyt
h
ep
o
l
y
n
o
m
i
a
la
p
p
r
o
x
i
m
a
t
i
o
na
n
deg
o
l
d
e
ns
e
c
t
i
o
nm
c
t
h
o
d
.

5
.
2
.
1 GraphlcalSolutlon

twasconstructedtohavemultiple
Theproblemdoesnotrepresentanyrealdesign.I
l
o
c
a
l minimums w
i
t
h
i
nt
h
ea
r
e
ao
fi
n
t
e
r
e
st
.E
x
p
r
e
s
s
i
o
n(
5.
1
)r
e
p
r
c
s
e
n
t
san
u
n
c
o
n
s
t
r
n
i
n
e
d
p
r
o
b
l
e
m
.
T
h
e
s
i
d
e
c
o
n
s
t
r
a
i
n
t
s
a
c
c
o
m
p
a
n
y
al
1problcmsinthisbookto
conv
e
yt
h
e
i
d
e
at
h
a
tt
h
e
r
e
a
r
e
n
o
t
r
uI
yun
c
o
n
s
t
r
a
i
n
e
dprobl
e
m
s
.S
i
d
e
c
o
n
s
t
ra
i
nt
s
a
l
s
o
d
e
f
i
n
eana
c
c
e
p
t
a
b
l
ed
e
s
i
g
nr
e
g
i
o
nf
o
ra
l
lp
r
o
b
l
e
m
s
.S
e
c
t
i
o
n5
.
2w
i
l
le
x
p
l
o
r
et
h
i
s
probl
e
m
t
h
r
o
u
g
h
t
h
c
v
a
ri
o
usmethodsment
i
o
n
e
d
c
a
ri
I
e
r.

Th
eexampleh
a
sa
l
r
e
a
d
ymadei
t
sa
p
p
e
a
r
a
n
c
ei
nC
h
a
p
t
e
r4
.F
i
g
u
5
.
1i
sap
l
o
to
ft
h
e
o
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
ni
nEx
ample5
.1
.I
ta
l
s
oi
n
c
l
u
d
e
scs
o
l
u
t
i
o
nExample5
.
1
a
.I
f
e
x
p
r
e
s
s
i
o
n(
5
.
3
)r
e
p
r
e
s
e
n
t
sa
ni
n
c
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
tt
h
ef
e
a
s
i
b
l
er
e
g
i
o
nf
o
ro
n
e
v
a
r
i
a
b
l
ef
u
n
c
t
i
o
n
si
sq
u
i
t
ed
i
f
f
e
r
e
n
tf
r
o
mt
h
et
w
o
v
r
u
i
a
b
l
ep
r
o
b
l
e
m
se
n
c
o
u
n
t
e
r
e
di
n

-DlmenslonalProblem
5
.
1
.
1 ConstralnedOne

Theo
n
l
ycons~raint t
h
a
tExample5
1
. as
i
n
g
l
e
v
a
r
i
a
b
l
ep
r
o
b
l
e
m
c
a
na

ommodatei
s
i
n
e
q
u
a
l
i

c
o
n
s
t
r
a
i
n
.
tFromt
h
cp
r
e
v
i
o
u
sc
h
a
p
t
c
rt
h
e
r
ec
a
nb
emo
o
ncs
u
c
h
c
o
n
s
i
n.
tAs
i
n
g
l
e
v
a
r
i
a
b
l
ep
r
o
b
l
e
mc
n
n
n
o
th
a
v
e u
a
l
i
t
yc
o
n
s
t
r
a
i
n
t(
C
h
a
p
t
e
r4
)
.
dm
u
l
t
i
v
a
r
i
a
b
l
eo
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
e
mi
tw
i
l
lo
f
t
e
nben
e
c
e
s
s
a
r
yl
o
.
f
i
n
d
I
nac
o
n
s
t
r
n
i
n
es
t
e
p
s
i
z
e

s
u
c
hecons
i
n
tj
u
s
tbecomesa
c
t
i
v
e
.I
ti
sp
o
s
s
i
b
l
et
odefme
p
r
o
b
l
e
m
.

S
u
b
j
e
c
tt
o
: g(
)
:0
.
7
5

2
-1
.5
IS0
OSS4

(
5
.
1
)
(
5.
3
)

(
5
.
2
)

I
na
c
t
u
a
la
p
p
l
i
c
a
t
i
o
n
si
ti
smorei
I
k
e
l
ya
tc
x
p
r
e
s
s
i
o
n
s(
5
.
1
)
(
5.
3
)e
s
t
a
b
l
i
s
htwo
p
r
o
b
l
e
m
st
h
a
tp
r
o
v
i
d
ed
i
f
f
e
r
e
n
ts
o
l
u
t
i
o
n
st
ocs
t
e
p
s
i
z
c

.Thef
i
r
s
ti
ses
o
l
u
t
i
o
nt
o
)a
n
d(
5
.
2
)whichi
st
h
esamca
sExnmplc5
.1
.Thesecondi
st
h
e
e
x
p
r
e
s
s
i
o
n
s(
5
.1
s
o
l
u
t
i
o
nt
o
g
(

.
)
:
0
.
7
5

2
-1
.5
1=0

(
5.
3
)

OSS4

(
5
.
2
)

g
(

'h

Minimize f
(

.
)
=
(
1
)
2
(

-2)(-3
)

{}()-

Example5
.
1
a

" '
J

c
o
n
s
t
r
a
l
n
e
dminimum

1 .

;
.
.

feiI~ .r~"

H J " ' l g l o b a l~Inlmum


u
n
c
o
n
s
t
r
a
i
n
e
dminimum
1
.6

0
.
5
g
u
r
e5
.
1

G
r
a
p
h
i
c
a
ls
o
l
u
l
i
o
n
:E
x
a
m
p
l
e5
.
1
.

3
.
5

2
0
6 N
U
M
E
A
I
C
A
LT
E
C
H
N
I
Q
U
E
S

T
H
EO
N
E
.
O
I
M
E
N
S
I
O
N
A
LP
A
O
B
L
E
M

5
.
2S
O
L
U
T
1
0
NTOT
H
EP
A
O
B
L
E
M 2
0
7

:
:
t
z
;
;
:;3
2
1
i
2
2
2
2
1
1

1FZZ;itmt

E
x
a
m
p
l
e
5
.
1

=2.6484;

f*=
0
.
6
1
9
5

(
5.
4a
)

Exa
m
p
l
e5
.
1
a

*=2
.
5
2
7
5
;

f
.=-0.5815

(
5.
4b
)

5
.
2
.
2 Newton-RaphsonTechnique
TheN
c
w
t
o
n
R
a
p
h
s
o
nt
e
c
h
n
i
q
u
e

a
l
s
or
c
f
e
r
r
d
t
oa
sN
c
w
t
o
n
'
st
e
c
h
n
i
q
u
e
i
sag
r
a
d
i
e
n
t
-

b
a
s
c
dnt
of
i
n
d
i
n
gcr
o
o
to
fas
i
n
g
l
cn
o
n
l
i
n
e
a
re
q
o
n
.R
o
o
tf
i
n
d
i
n
g

l
l
itU122ZZ;::rh

j
l

i
sf
a
i
r
l
ys
t
a
n
d
a
r
di
ni
t
e
r
a
t
i
v
et
e
c
h
n
i
q
u
e
sf
o
rmanyc
l
a
s
s
e
so
f
T
h
ea
b
o
v
es
e
q
u
e
n
p
r
o
b
l
e
m
s
.I
ta
l
s
oc
a
p
l
u
r
e
st
h
ee
s
s
e
n
c
ef
o
rm
o
s
to
ft
h
et
e
c
h
n
i
q
u
e
st
of
o
l
l
o
wi
nt
h
i
sb
o
o
k
.
f

rco
fi
t
e
r
a
t
i
v
eme
o
ds-thes
t
a
r
t
i
n
gs
o
l
u
t
i
o
n
.Thi
s
S
l
e
p1i
n
d
i
c
a
t
e
st
h
ep
r
i
m
a
r
yi
i
sa
ni
n
i
t
i
a
lg
u
e
s
sp
r
o
v
i
d
e
db
yt
h
eu
s
e
r1
0s
l
a
r
tt
h
ei
t
e
r
a
t
i
v
ep
r
o
c
e
s
s
.Thei
t
e
r
a
t
i
v
e
p
r
o
c
e
s
si
sc
o
n
t
i
n
u
e
d1
0t
h
es
o
l
u
t
i
o
nb
yc
a
l
c
u
l
a
t
i
n
gc
h
a
n
g
e
si
nt
h
ev
a
r
i
a
b
l
e(
.
1

.
)
.The
v
a
l
u
eo
ft
h
ev
a
r
i
a
b
l
ei
su
p
d
a
t
e
d
.C
o
n
v
e
r
g
e
n
c
ei
sc
h
e
c
k
e
d
.I
fEq
u
a
t
i
o
n(
5
.
5
)i
ss
a
t
i
s
f
i
e
d
t
h
e
nt
h
es
o
l
u
t
i
o
nh
a
sb
e
e
nf
o
u
n
d
.I
fc
o
n
v
e
r
g
e
n
c
ei
sn
o
ta
c
h
i
e
v
e
d t
h
eu
p
d
a
t
e
d
v
a
l
u
ep
r
o
v
i
d
e
st
h
ev
a
l
u
et
op
r
o
c
e
e
dw
i
t
h n
e
x
ti
t
e
r
a
t
i
o
n
.Ofc
o
u
r
s
e
.
1
canb
e
c
a
l
c
u
l
a
t
e
di
nmanyw
a
y
sa
n
dN
e
w
t
o
n
R
a
p
h
s
o
ni
sj
u
s
to
n
eo
ft
h
ew
a
y
s
e
s
t
a
b
l
i
s
ht
h
e
c
h
a
n
g
ei
nt
h
ev
a
r
i
a
b
l
e
.

C
a
l
c
u
l
a
t
l
o
n01A
(Newton-Raphson):L
e
t be t
h
ec
u
r
r
e
n
tv
a
l
u
eo
ft
h
e
v
a
r
i
a
b
l
e
.I
ti
sa
s
s
u
m
e
dt
h
a
tE
q
u
a
t
i
o
n(
5
.
5
)i
sn
o
ts
a
t
i
s
f
i
e
da
t i
sv
a
l
u
eo
ft
h
ev
a
r
i
a
b
l
e
(
e a
s
o
nf
o
ru
st
oi
t
e
r
a
t
e
)
.Let
abeaneighboringvalue.Ideallywewouldlike10
(
e
v
e
n wea
r
ea
w
a
r
ei
tr
n
i
g
h
tt
a
k
eu
ss
e
v
e
r
a
l
a
c
h
i
e
v
ec
o
n
v
e
r
g
e
n
c
ea
tt
h
i
sv
a
l
u
i
t
c
r
a
t
i
o
n
s1
0a
c
h
i
e
v
cc
o
n
v
e
r
g
n
c
e
)
.U
s
i
n
gT
a
y
l
o
r
'
st
h
e
o
r
e
me
x
p
a
n
d
e
dt
ot
h
el
i
n
e
a
r
t
e
r
r
no
n
l
y

;
:
;
L

ee
q
u
a
t
i
a

()=0

d
(

)=c
t
(

+)=()+25A=0

(
5
.
5
)

A=1=-' c
b
(
o
.
)

I
d

w
h
e
r

)i
san
o
n
l
i
n
e
a
re
q
u
a
t
i
o
ni
nt
h
ev
a
r
i
a
b
l
e
.
l
et
e
c
h
n
i
q
u
eh
a
sa
d
d
i
t
i
o
n
a
lf
e
a
t
u
r
c
s
:
I
thasag
m
e
t
r
i
cu
n
d
e
i
n
n
i
n
g
.
I
tustheT
a
y
l
o
rs
e
r
i
e
se
x
p
a
n
d
e
di
I
n
e
a
r
Iy
.
I
ti
si
t
e
r
a
t
i
v
e
.

I
th
a
st
h
ep
r
o
p
e
r
t
yo
fq
u
a
d
r
a
t
i
cc
o
n
v
e
r
g
e
n
c
e
.

Anyi
t
e
r
a
t
i
v
ct
c
c
h
n
i
q
u
ea
d
d
r
c
s
s
i
n
gE
q
u
a
t
i
o
n(
5
.
5
)c
a
nb
es
u
m
m
a
r
i
z
e
da
s

G
e
n
e
r
i
cA
l
g
o
r
i
t
l
r
m(
A
5
.
1
)
S
t
e
p1
.Assum

S
t
e
p2
.C
a
l
c
u
l
a
t
e.
1

S
t
e
p3
.U
p
d
a
t
e =+.
((

)=
:
0
)e
ne
x
i
t
I
fc
o
n
v
e
r
g
e
d
I
fn
o
tc
o
n
v
e
r
g
e
d(

(
:
;
f
:0
)e
n
a
got
oS
t
e
p2

(
5
.
6
)

(
5
.
7
)

F
o
rEq
u
a
t
i
o
n(
5
.
7
)t
ob
ce
f
f
c
c
t
i
v
e
eg
r
a
d
i
e
n
to
ft
h
ef
u
n
c
t
i
o
ns
h
o
u
l
dn
o
tb
ez
c
r
o
.l
t
i
sa
1
s
oi
n
e
v
i
t
a
b
l
ea
tc
h
a
n
g
e
si
n

w
i
l
lb
el
a
r
g
ew
h
e
r
ec
ti
sf
l
a
t

a
n
ds
m
a
l
1w
h
e
r
et
h
e
s
l
o
p
ei
sl
a
r
g
e
.F
o
rt
h
eN
e
w
t
o
n
R
a
p
h
s
o
nm
e
t
h
o
dt
ob
ee
f
f
c
c
t
i
v
et
h
ei
t
e
r
a
t
i
o
n
ss
h
o
u
l
d
a
v
o
i
d g
i
o
n
sw
h
e
r
et
h
es
l
o
p
ei
ss
m
a
l
.
lT
h
i
si
sas
c
r
i
o
u
sd
i
s
a
d
v
a
n
t
a
g
eo
fi
sm
c
t
h
o
d
.
F
i
g
u
r
e5
.
2i
l
Iu
s
t
r
a
t
e
st
h
eg
e
o
m
e
t
r
i
c
a
lc
o
n
s
t
r
u
c
t
i
o
no
ft
b
et
e
c
h
n
i
q
u
ca
tt
h
ep
o
i
n
t =3
.

Example5
.
1 TheN
e
w
t
o
n
R
a
p
h
s
o
nt
c
c
h
n
i
q
u
ei
sa
p
p
l
i
e
dt
ot
h
ef
i
r
s
t
o
r
d
e
rc
o
n
d
i
t
i
o
n
s
(FOC)o
fExample5
.1
.
M
i
n
i
m
i
z
c

f(

)
=
(
1)2
(
2)(-3
)

(
5
.
1
)

p
I
1
Ii
0b
ed
o
w
n
l
o
a
d
e
df
w
mt
h
cw
e
bs
i
l
e i
n
d
i
c
a
l
e
db
yb
o
l
d
f
n
c
cs
n
n
ss
e
r
i
fI
y
p
e
.

S
u
b
j
e
c
tt
o
: 0SS4

(
5
.
2
)

TheFOCis

==

-2

3
)
)
+
(

)
1
3)
+
(

)
l
r

2
(
5
.
8
)

208 N
U
M
E
R
I
C
A
LT
E
C
H
N
I
Q
U
ES-T
H
EO
N
E
.
O
I
M
E
N
S
I
O
N
A
LPROBLEM

5
.
2S
O
L
UTlO
NTOTHEP
R
O
B
L
E
M 209

1
0

'
h
ec
o
d
eSe
c5
_2_2.mr
u
i
r
e
sf
i
v
ei
t
e
r
a
t
i
o
n
st
oc
o
n
v
e
r
g
et
h
ev
a
1
u
eo
ft
It
o
1
.
0E
8(
c
o
n
s
i
d
e
r
e
dz
e
r
oh
e
r
e
)
.Tsva
Iueof1
.0E

08i
st
e
r
m
e
dt
h
ec
o
n
v
e
r
g
e
n
c
e(
o
r

-2
0

0
.
5

1
.
5

2
.
5

3
.
5

F
l
g
u
r
e5
.
2D
e
v
e
l
o
p
m
e
n
t0
1N
e
w
t
o
n
.
R
a
p
h
s
o
nt
e
c
h
n
i
q
u
e
.
JeT
echnlque:Twoi
t
e
r
a
t
i
o
n
sa
r
eshownh
e
r
e
.Thec
o
m
p
l
e
t
ea
p
p
l
i
c
a
t
i
o
ni
sdone
t
h
r
o
u
g
hMA11ABcodesecs-2-2.m.Thec
o
d
eu
s
e
ss
y
m
b
o
l
i
cc
a
l
c
u
l
a
t
i
o
n
.
I
nrumino
c
o
d
et
h
ef
i
g
u
r
ewindowd
i
s
p
l
a
y
s r
n
a
t
i
o
no
ft
h
ei
t
e
r
a
t
i
r
a
v
e
l
60;
A
l
s
oshowni
st
h
ec
h
a
n
g
ei
n6withe
a
c
hi
t
e
r
a
t
i
o
n
. Th
ei
t
e
r
a
t
i
v
er
e
s
u
l
t
sa
r
e
n
d
.Thec
a1
cu
l
a
t
i
o
n
sr
e
q
u
i
r
ct
h
eg
r
a
d
i
e
n
to
f(

)
:
s
u
m
m
a
r
i
z
c
di
nat
a
b
l
e8
tec

s
t
o
p
p
i
n
g
)c
r
i
t
e
r
i
o
n
.Thca
c
c
c
l
e
r
a
t
i
o
no
fc
o
n
v
e
r
g
e
n
c
eo
rr
a
t
eo
fc
o
n
v
e
r
g
e
n
c
et
ot
h
c
s
o
l
u
t
i
o
ni
sf
o
r
m
a
l
l
yd
e
f
i
n
c
di
nmanyd
i
f
f
e
r
e
n
tw
a
y
s
.Oneo
ft
h
cwaysi
st
om
o
n
i
t
o
rt
h
e
r
a
t
i
oo
f
(+6)
1
(

)
.Ano
t
h
c
ristrack6.l
na
p
p
l
y
i
n
ga
n
du
n
d
e
r
s
t
a
n
d
i
n
g
It
e
c
h
n
i
q
u
e
si
ti
si
m
p
o
r
t
a
n
tt
od
e
v
e
l
o
pa
ni
n
s
t
i
n
c
ta
n
df
e
e
lf
o
rt
h
enumbers
n
u
m
e
r
i
ca
c
o
r
r
e
s
p
o
n
d
i
n
gt
ov
a
r
i
o
u
sf
u
n
c
t
i
o
n
s ei
t
e
r
a
t
i
o
n
sp
r
o
g
r
e
s
s
.Thes
p
e
e
do
re
q
u
a
l
i
t
y
o
fc
o
n
v
e
r
g
e
n
c
ec
a
nbeg
a
u
g
e
dbyes
h
i
f
ti
nt
h
edccim
a
Ip
o
i
n
ti
nt
h
eq
u
a
n
t
i
t
yo
f
I
ueo
ft
Ithrought
h
ef
i
v
ei
t
e
r
a
t
i
o
n
s
i
n
t
e
r
e
s
tt
h
r
o
u
g
ht
h
ei
t
e
r
a
t
i
o
n
s
.C
o
n
s
i
d
e
rt
h
eva
[
4
.
0
0
0
0
0
.
8
7
5
0
0.
1040
0
.
0
0
2
3
0
.
0
0
0
0
]
.D
e
f
i
n
i
t
e
l
yec
h
a
n
g
c
si
na
r
en
o
n
l
i
n
e
a
r
.
Th
eNewton-Raphsont
e
c
h
n
i
q
u
ec
a
nb
ed
i
r
e
c
U
ya
p
p
l
i
e
dt
oE
qu
a
t
i
o
n(
5
.
3
)
a
f
t
e
r
a
I
l
t
h
emeUlodi
sd
e
s
i
g
n
e
dt
oo
p
e
r
a
t
eont
h
e
s
ek
i
n
d
so
fp
r
o
b
l
e
m
s
.I
ti
sl
e
f
ta
sa
ne
x
e
r
c
i
s
e
f
o
r r
d
e
r
.l
e
r
ei
sav
e
r
yi
m
p
o
r
t
a
n
tf
e
a
t
u
r
eo
fmosti
t
c
r
a
t
i
v
et
e
c
h
n
i
q
u
e
st
h
a
t
e
x
p
!
o
r
e m t
i
p
l
es
o
l
u
t
i
o
n
s
a
si
s ec
a
s
eo
f Example 5
.
1
. Rerun t
h
ec
o
d
e
Sec5_Z2.mw
i
t
has
t
a
r
t
i
n
gva
Iueof0
.
5f
o
ra
.
.Afters
i
xi
t
e
r
a
t
i
o
n
s

convergest
o
ev
a
Iueof1
.
0
whichi
sal
o
ca
Iminimum.Whe
r
ewouldi
t
e
r
a
t
i
o
n
sl
e
a
di
ft
h
es
t
a
r
t
v
a
Iuei
s1
.
5?Thesea
n
s
w
e
r
sa
r
e s
yt
os
g
r
a
p
h
i
c
a
Il
yi
nao
n
e
-o
rmaybea
t
w
o
v
a
r
i
a
b
l
ep
r
o
b
l
e
m
.I
nat
y
p
i
c
a
Ie
n
g
i
n
e
e
r
i
n
gd
e
s
i
g
np
r
o
b
l
e
m

s
i
g
n
i
f
i
c
a
n
te
x
p
e
r
i
e
n
c
e
a
n
di
n
t
u
i
t
i
o
na
r
er
e
q
u
i
r
e
dt
or
e
c
o
g
n
i
z
eU
l
e
s
cs
i
t
u
a
t
i
o
n
swhenc
yo
c
c
u
r
.A h
e
a
l
t
h
y
S
p
i
c
i
o
no
ft
h
es
o
l
u
t
i
o
nm
u
s
ta
Iwaysbei
nt
h
eb
a
c
ko
fo
n
e
'
smindwhene
x
p
l
o
r
i
n
g
m
u
l
t
i
v
a
r
i
a
b
l
ep
r
o
b
l
e
m
s
.I
n summarym
o
s
ti
t
e
r
a
t
i
v
en
u
m
e
r
i
c
a
lt
e
c
h
n
i
q
u
e
sa
r
e

1
St
h
a
ta
r
ec/o
s
e10wheret
h
e
ys
t
a
r
t
f
r
o
m
.
d
e
s
i
g
n
e
dt
oc
o
n
v
e
r
g
et
os
o
l
u
t
i
o
5
.
2
.
3 BisectlonTechnique
Th
i
si
sa
n
o
t
h
e
rp
o
p
u
l
a
rt
c
c
h
n
i
q
u
et
of
i
n
d t
o
faf
u
n
c
t
i
o
n
.I
ti
sa
l
s
oc
a
Il
e
dab
i
n
a
r
y
s
e
a
r
c
ho
ri
n
t
e
r
va
Iha
Iv
i
n
gp
r
o
c
e
d
u
r
e
.U
n
l
i
k
et
h
eNewton-Raphson method t
h
i
s
p
r
o
c
e
d
u
r
ed
o
e
sn
o
t q
u
i
r
et
h
ee
v
a
Iu
a
t
i
o
no
feg
r
a
d
i
e
n
to
fef
u
n
c
t
i
o
nwhoser
o
o
t
i
sb
e
i
n
gs
o
u
g
ht
.Oneo
ft
h
er
e
a
s
o
n
st
h
emethodi
sb
e
i
n
gi
n
c
l
u
d
e
di
st
ocomparei
tw
i
t
h
t
h
eg
o
l
d
e
ns
c
c
t
i
o
nmethodd
i
s
c
u
s
s
e
dl
a
t
c
r
.Then
u
m
e
r
i
c
a
l
t
e
c
h
n
i
q
u
ei
ni
smethodi
s
basedonthe o
n
i
n
g t
h
e
r
st
r
a
p
p
e
do
rboundb
e
t
w
e
e
n
t(orzero)ofafunctioni
ap
o
s
i
t
i
v
ea
n
dan
e
g
a
t
i
v
eva
Iueoft
h
ef
u
n
c
t
i
o
n
.Th
ea
c
t
u
a
Is
o
l
u
t
i
o
na
c
h
i
e
v
e
dbyE
methodi
sa
ni
n
t
c
r
v
a
li
nwhichez
e
r
oo
rt
h
es
o
l
u
t
i
o
nc
a
nb
el
o
c
a
t
e
d
.Thcf
i
n
a
Ii
n
t
e
r
v
a
l
i
st
y
p
ilIyd
c
t
c
r
m
i
n
e
dt
ob
eav
c
r
ysma
l
Iv
a
l
u
e

c
a
l
l
c
dt
h
ct
o
l
e
r
a
n
c
e
o
ft
h
eo
r
d
c
ro
f
l
O
E
0
8
.S
c
ei
si
sar
o
o
tf
i
n
d
i
n
gp
r
o
c
e
d
u
r
e
whena
p
p
l
i
e
dt
oo
p
t
i
m
i
z
a
t
i
o
ni
ti
s
a
p
p
l
i
c
dt
o g
r
a
d
i
e
n
to
f o
t

t
i
v
ewhichwei
d
e
n
t
i
f
i
c
da
s
(

.
)Eq
u
o
n(
5.
5
)
.
Est
a
b
l
i
s
h
i
n
gt
h
eminimumo
ft
h
eo
b
j
e
c
t
i
v
ei
st
h
e
r
e
b
yt
r
a
n
s
l
a
t
c
dt
ol
o
c
a
t
i
n
ger
o
o
t
o
f
eg
r
a
d
i
e
n
tf
u
n
c
t
i
o
n
.I
ti
sp
o
s
s
i
b
l
ct
od
e
v
c
l
o
pab
i
s
e
c
t
i
o
nt
c
c
h
n
i
q
u
ef
o
rd
i
r
e
c
t
l
y
h
a
n
d
l
i
n
gt
h
em
i
n
i
m
i
z
a
t
i
o
np
r
o
b
l
e
m
.
Themethodr
u
i
r
e
stwop
o
i
n
t
st
os
t
a
r
tsaynandbThev
a
Iueo
fI
ta
tE
h
e
s
cc
a
n be t
h
es
i
d
cc
o
n
s
t
r
a
i
n
t
so
ft
h
e
p
o
i
n
t
s must b
eo
p
p
o
s
i
t
ei
ns
i
g
n
.
'
o
n
e
d
i
m
e
n
s
i
o
na
Io
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
e
m
.l
ti
sassumed a
tI
c
a
s
tones
o
l
u
t
i
o
ne
x
i
s
t
s
si
n
i
t
i
i
n
t
e
r
v
a
l
.D
u
r
i
n
ge
a
c
hi
t
e
r
a
t
i
o
ni
si
n
t
e
r
va
1i
sha
Ivedwithemidpoint
i
n i
o
ft
h
ei
n
t
e
r
v
a
lr
e
p
l
a
c
i
n
gc
i
t
h
e
r orb
w
h
i
l
ek
c
c
p
i
n
gt
h
cr
o
o
ts
t
i
l
lt
r
a
p
p
c
db
c
t
w
e
e
n
l

'()=2(-2)(-3
)+
4
*
(
1
)
*
(
3
)+4
*
(
1)*(-2
)+
2
*
(
1
)
^
2

I
t
e
r
a
t
i
o
n1
:
=3;

f(
)=0.0;

)=4

t
I
'()=16

6=-0.
2
5

t
I(+L
1

.
)=0.875-notc
o
n
v
e
r
g
e
d
a
tl
e
a
s
ta
n
o
t
h
e
ri
t
e
r
a
t
i
o
nn
e
c
c
s
s
a
r
y

I
t
e
r
a
t
i
o
n2
:
2
.
7
5
;
=

f
(
1

.
)
=0
.
5
7
4
2
;

.
)
=
0
.
0
8
7
5
;

.
)
=
9
.
2
5
;

= O
.

4
6

(+6)=0.1040
n
o
tc
o
n
v
e
r
g
e
d
a
tl
e
a
s
ta
n
o
t
h
c
ri
t
e
r
a
t
i
o
nn
c
c
e
s
s
a
r
y

210 N
U
M
E
R
I
C
A
LT
E
C
H
N
I
Q
U
ES-T
H
EO
N
E
D
I
M
E
N
S
I
O
N
A
LP
R
O
B
L
E
M

5
.
2S
O
L
U
T
I
O
NTOT
H
EP
R
O
B
L
E
M 2
1
1

a
m
c
t
i
o
nv
a
l
u
e
so
f
o
p
p
o
s
i
t
es
i
g
n
.
T
h
eitemtiveMmiquei
se
x
p
r
e
s
s
e
da
st
h
ef
o
l
l
o
w
i
n
g

a
Igorithm(
as
e
to
fp
r
o
c
e
d
u
r
e
s er
e
p
e
a
t
e
d
)
.

odi
ss
i
m
p
l
ci
nc
o
n
c
e
p
t
.I
n
s
t
e
a
do
fm
i
n
i
m
i
z
i
n
gad
i
f
f
i
c
u
l
tf
u
n
c
t
i
o
no
fo
n
e
Theme
v
a
r
i
a
b
l
em
i
n
i
m
i
z
eap
o
l
y
n
o
m
i
a
la
ta
p
p
r
o
x
i
m
a
t
e
st
h
ef
u
n
c
t
i
o
n
.Thco
p
t
i
m
a
lva
Iu
eo
f
ev
a
r
i
a
b
l
m
i
n
i
m
i
z
e
st
h
ep
o
l
y
n
o
m
i
a
li
st
h
c
nc
o
n
s
i
d
c
r
e
dt
oa
p
p
r
o
x
i
m
a
t
ct
h
e
a
Ju
co
fev
a
r
i
a
b
l
ef
o
rt
1
J
eo
r
i
g
i
na
If
u
n
c
t
i
o
n
.I
ti
sr
a
r
ef
o
rt
1
J
ed
c
g
r
e
eo
ft
1
J
e
o
p
t
i
m
a
lv
a
p
p
r
o
x
i
m
a
t
i
n
gp
o
l
y
n
o
m
ia
Jt
oe
x
c
e
c
dt
h
r
e
c
.A q
u
a
d
r
a
t
i
ca
p
p
r
o
x
i
m
a
t
i
o
ni
ss
t
a
n
d
a
r
d
l
e
s
sei
r
dd
e
g
r
e
ei
sw r
a
n
t
e
d
.I
ti
sc
1e
a
rt
h
a
ts
c
r
i
o
u
se
r
r
o
r
si
na
p
p
r
o
x
i
m
a
t
i
o
na
r
e
e
x
p
e
c
t
e
di
fep
o
l
y
n
o
m
i
a
li
st
os
i
m
u
l
a
t
et
h
eb
e
h
a
v
i
o
ro
feo
r
i
g
i
n
a
lf
u
n
c
t
i
o
no
v
c
ra
.
g
er
a
n
g
eo
fva
Iu
e
so
f v
a
r
i
a
b
l
e
.M
a
t
h
e
m
a
t
i
c
a
lt
1
Je
oremse
x
i
s
t j
u
s
t
i
f
ya
l
1
Jef
u
n
c
t
i
o
n
w
i
t
hap
r
e
s
c
r
i
b
e
dd
e
g
r
e
eo
fe
r
r
o
r
wit
I
Ji
na
q
u
a
d
r
a
t
i
cr
e
p
r
e
s
e
n
t
a
t
i
o
no
ft
maJln
e
i
g
h
b
o
r
h
o
o
do
ft
1
Jem
inimum.What i
se
n
s
u
r
e
si
s a
tt
1
J
ep
o
l
y
n
o
m
i
a
l
s
a
p
p
r
o
x
i
m
a
t
i
o
ng
e
t
sb
e
t
t
e
ra
s eminimum i
sb
e
i
n
ga
p
p
r
o
a
c
h
. E
xample5
.
1
.
r
e
p
f
u
c
e
df
o
rc
o
n
v
e
n
i
e
p
c
e

i
su
s
e
df
o
ri
I
1u
s
t
r
a
t
i
o
no
ft
h
ep
r
o
c
e
d
u
r
e
.

A
l
g
o
r
i
t
I
J
m
!
o
rB
i
s
e
c
t
i
o
nM
e
t
l
l
o
d(
A
S
.
2
)
S
t
e
p1
: Choose a
n
d t
os
t
a
r
.
tL
et
S
t
e
p2
:S
e
t= + ( )
1
2

:I
f
(

.
)=0
.
0-ConvergedS
o
l
u
t
i
o
n-e
x
i
t
S
t
e
p3
s
e I
f(b-n)<l
OE04-t
o
l
e
r
a
n
c
emet-e
x
i
t
E
l
s
e I
f
(

.
)
*()
>
0
.
0
; en
Elseb
got
oS
t
e
p2

Wec
o
n
s
i
d
e
rExample5
.
1whichi
sr
e
p
r
o
d
u
c
e
dh
e
r
e
Minimize

f
{
)=(_1
)
2
(

-2)(-3
)

S
u
b
j
e
c
tt
o
: 0S 4

2.
4 PolynomlalApproxlmatlon
5.

Example5.1
(
5.
1
)

M
i
n
i
m
i
z
e

f
(
.)=(_1
)
2
(

-2
)
(
3
)

(
5
.
2
)

ThcFOCis

S
u
b
j
c
c
tt
o
: 0SS4

(
5
.
1
)
(
5
.
2
)

Aq
u
a
d
r
a
t
i
cp
o
l
y
n
o
m
ia
JP
(
)i
su
s
e
df
o
rt
h
ea
p
p
r
o
x
i
m
a
t
i
o
n
.Th
i
sp
o
l
y
n
o
m
i
a
li
s
e
x
p
r
e
s
s
e
da
s

(
5
.
8
)
S 5_.
3
.
m
: Thi
s ABc
o
d
er
u
n
st
h
eb
i
s
e
c
t
i
o
nmethodon p
l
e5
.1
.I
t
i
I
lu
s
t
r
a
t
e
st
h
et
r
a
p
p
i
n
go
femullrnumbyt
h
etwov
a
l
u
e
sone
i
t
h
e
rs
i
d
eOft
h
em
o
t
.
l
e
1
7i
t
e
r
a
t
i
o
n
sf
o
rat
o
l
e
r
a
n
c
eo
f1
.0E04whicht
e
r
m
i
n
a
t
e
dt
h
cp
r
o
g
r
a
m
.
Compa
h
i
sw
i
t
h5i
t
c
r
a
t
i
sf
o
rt
h
eNewton-Raphson1
h
o
d
.f
o
rab
e
t
t
c
rs
o
l
u
t
i
o
n
.
U l
o
w
e
rnumbero
fi
r
a
t
Jo
n
sc
b
e t
o qu
a
Ji
t
yo
fi
r
'
n
n
a
t
i
o
n edt
o
ed
v
a
r
i
a
b
l
echangcncg
r
a
d
i
e
n
t
b
a
s
c
di
n
f
o
r
m
a
t
i
o
nu
s
c
di
nt
h
c
upd
N
e
w
t
o
n
R
a
P
I
E
B
o
ntEChmquel
smores
o
p
h
i
s
t
i
c
a
t
e
d
.
t
h
o
u
g
hcg
r
a
d
i
e
n
tc
o
m
p
u
t
a
t
i
o
ni
s
a
na
d
d
i
t
i
o
na
Jw
orkl
o
a
d
.
T Kp
r
o
b
l
e
md
e
s
c
r
i
b
e
dbyE
q
u
a
t
i
o
n(53)cmbeh
a
n
d
l
e
dd
i
r
e
c
t
l
ybyt
h
E
b
i
s
e
c
t
i
o
n
o
d(

.
)i
s same(

)
.I
naw
a
Y
.
t
h

e
o
p

p
o u
t

Yt

o
r
e

v
1
1

tp
r
i
o
rk
n
o
w
l
c
d
g
e
.p
a
r
t
i
a
r
l
yw
i
t
hr
e
s
p
d
e
f
i
n
i
n
ga
n
d
s
l
a
t
i
n
ga
na
l
g
o
r
i i
n
t
oI
U
D
ngcode.Therei
s
u
s
u
as
t

r
o

ng
co
r

r
e
s
p

o
n

d
l
b
e
t
w
e
c
nes
t
c
p
b
y
s
t
c
pdc
r
i
p
t
i
o
no
fen
u
m
e
r
i
ca
Jt
e
c
h
n
i
q
u
e(
t

hea
l
g
o
r
i
t

h
r

e
c
o
d
e a

tw
i
l
li
m
p
l
e
m
e
n
t e p
s e
f
f
e
c
tas
o
l
u
t
i
o
n(
f
o
re
x
a
m
p
l
e
.

:
2
2
t
i
t
B
;
E
P
;
5
2
1
1
1
1
:
Z
Z
2
2
0
:
2
2
L
t
J
b
r
e

f
o
l
l
o
w
l
n
gtwomethodsa
r
eu
s
e
de
x
t
e
n
s
i
v
e
l
yf
o
r
o
n
e
d
i
m
e
n
s
i
o
n
a
lmim
i
z
a
t
i
o
n
.They
h
a
n
d
l
ct
h
em
i
n
i
m
i
z
a
t
i
o
np
r
o
b
l
e
md
i
r
e
c
t
l
y
.

2
P
(
)
=bo+b+b
2

(
5
.
9
)

Twoe
l
c
m
e
n
t
sn dt
o u
n
d
e
r
s
t
o
o
dp
r
i
o
rt
ot
1
Jef
ol
1owingd
i
s
c
u
s
s
i
o
n
.l
ef
i
r
s
t
v
a
l
u
a
t
i
o
no
ft
1
Jep
o
l
y
n
o
m
i.andt
1
Jes
e
c
o
n
dc
o
n
c
e
r
n
sei
n
c
l
u
s
i
o
no
f
conccmsee
e
x
p
s
s
i
o
n(
5
.
2
)i
ned
i
s
c
u
s
s
i
o
n
.Thcp
o
l
y
n
o
m
ia
Ji
sc
o
m
p
l
e
t
e
l
yd
e
f
i
n
e
di
ft
h
e
.andb
r
cknown.Tod
e
t
e
r
m
i
n
e m.t
1
Jr
e
ed
a
t
ap
o
i
n
t
s[
(

1
'
/
.
)
c
o
e
f
f
i
c
i
e
n
t
sbo b
2a
2
'!
z
)
.(
.
3
1
3
)
]a
r
cg
e
n
e
r
a
t
e
df
r
o
mE
q
u
a
t
i
o
n(
5
.
1
)
.T
h
i
ss
e
t
supai
In
c
a
rs
y
s
t
e
mo
f
t
h
r
e
e
(
.
e
q
u
a
t
i
o
n
si
neunknownsby q
u
i
r
i
n
gt1Jeva
J
u
c
soft
1
Jef
u
n
c
t
i
o
na
n
dt
h
e
Jm
u
s
tb
et
h
esamea
tt
1
Jet
1
J
r
c
ep
o
i
n
t
s
.Th
es
o
l
u
t
i
o
nt
ot
1
Ji
ss
y
s
t
e
mo
f
p
o
l
y
n
o
m
ia
a
l
u
e
so
ft
1
Jcc
o
e
f
f
i
c
i
e
n
t
s
. 1
c
o
n
s
i
d
e
r
a
t
i
o
no
fe
x
p
r
e
s
s
i
o
n(
5
.
2
)
e
q
u
a
t
i
o
n
si
s v
1
J
et
y
p
eo
fo
n
e
d
i
m
e
n
s
i
o
n
a
Jp
r
o
b
l
c
mb
e
i
n
gs
o
l
v
e
d
.I
ft
h
eo
n
e
d
i
m
c
n
s
i
o
na
I
d
e
p
e
n
d
sont
p
r
o
b
l
e
mi
sag
e
n
u
i
n
es
i
n
g
l
e
v
a
r
i
a
b
l
ed
e
s
i
g
np
r
o
b
l
e
me
ne
x
p
r
e
s
s
i
o
n(
5
.
2
)n
e
e
d
st
o
t
. I
f e o
n
e
d
i
m
e
n
s
i
o
na
I problem i
sas
u
b
p
r
o
b
l
e
m m e
b
ep
r
e
s
e
n
m
u
J
t
i
d
i
m
e
n
s
i
o
n
a
lo
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
e
m

t
1
Je
ne
x
p
r
e
s
s
i
o
n(
5
.
2
)i
sn
o
ta
v
a
i
l
a
b
l
e
.I
nt
h
a
t
ceas
c
a
n
n
i
n
gp
r
o
c
c
d
u
r
ei
su
s
e
dt
od
e
f
i
n

a
n
d3

ScannlngProcedure: l
ep
r
i
ss
t
a
r
t
e
df
r
o
m l
o
w
e
ri
Im
i
tf
o
r

.Ava
Ju
e
o
f
z
e
r
of
o
ri
sva
Ju
e b
ej
u
s
t
i
f
i
e
ds
i
n
c
ei
tr
e
f
e
r
st
ova
Ju
e
s c
u
r
r
e
n
ti
t
e
r
a
t
i
o
n
.
1
s
c
a
l
e
dp
r
o
b
l
e
mt
1
J
i
sva
I
u
e
Ac
o
n
s
t
a
n
ti
n
t
c
r
v
a
lf
o
r
6
.
isa
l
s
oi
d
e
n
t
i
f
i
e
d
.F
o
rawel
l
1y1
.S
t
a
r
t
i
n
ga
tt
1
J
el
o
w
e
ri
Imit

i
n
t
c
r
va
Ii
sd
o
u
b
l
e
du
n
t
i
lt
l
J
r
e
ep
o
i
n
t
sa
r
e
i
su
s
u
a

.
5
.
2S
O
L
U
T

O
NTOT
H
EP
R
O
B
L
E
M 213

212 N
U
M
E
R

C
A
LT
E
C
H
N

Q
U
E
S
T
H
EO
N
E
.
O

M
E
N
S

O
N
A
LP
R
O
B
L
E
M
d
e
t
e
n
n
i
n
e
ds
u
c
ha
tt
h
eminimumi
sb
r
a
c
k
e
t
e
db
e
t
w
e
e
nt
h
e
m
.Wi r
e
s
p
e
c
tt
o
Example5

1
.t
h
es
c
a
n
n
i
n
gproceduregenerat
t
h
ef
o
l
l
o
w
i
n
gv
a
l
u
e
s
:

1
0

=0
;j
{O
)=6
;
=0; /
=6
=1
;j
{
l
)=0
; =1
; =0

=2
;j
(2
)=0
; t
h
i
sc
a
n
n
o
tbe theminimumi
sn
0
1y
e
t pped

/
2=18

=4
;j
(
4
)=1
8
; =
4
;

di
se
s
s
e
n
t
i
a
ls
Si
ti
sb
o
t
hi
n
d
i
f
f
e
r
e
n
tt
ot
h
ep
r
o
b
l
e
m
Ap
r
o
c
e
s
ss
u
c
ha
st
h
a
ti
l
l
u
s
t
r
a
t
b
e
i
n
gs
o
l
v
e
dandc
a
nbeprogrammed s
i
l
y
.Thei
m
p
o
r
t
a
n
tr
e
q
u
i
r
e
m
e
n
to
fanys
u
c
h
si
st
oe
n
s
u a
tt
h
eminimuml
i
e
sb
e
t
w
e
e
neI
i
m
i
t
se
s
t
a
b
l
i
s
h
e
dbyt
h
e
p
r
o
c
p
r
o
c
e
d
u
r
e
.l
i
sp
r
o
c
e
d
u
r
ei
sd
e
v
e
l
o
p
e
da
sMA
T
I
.
A
Bmf
i
)eU
pperBound_1Var.m.
Th
es
e
to
fi
I
n
e
a
re
q
u
a
t
i
o
n
st
oe
s
t
a
b
I
is
ht
h
ec
o
e
f
f
i
c
i
e
n
t
so
ft
h
ep
o
l
y
n
o
m
i
a
li
n
E
quation(
5
.
9
)i
s

6=
bo
+b(O)+b2(0)

(
5
.
1
0
8
)

+b(
0
:
:
:bo
l
)+b2(
l
)

(
5
.l
O
b
)

1
8=bo
+b

(
4
)+b
(
1
6
)
2

(
5
.l
O
c
)

0
.
5

1
.
5

3
.
5

o
l
y
n
o
m
i
a
la
p
p
r
o
ma
1
i
o
n
.
R
g
u
r
e
5
.
3p

UsingMA
TI.ABe
soh
o
nt
oee
q
u
a
t
i
o
n
si
s

h
o
=
6
.
0
;b
=-9; b2=3

(
5.
11
)

a
n
dt
h
ep
o
l
y
n
o
m
i
a
li
s

dedevelopmenti
st
e
r
m
e
dab
l
o
c
ks
t
r
u
c
t
u
r
e
da
p
p
r
o
a
c
h
I
nt
h
i
sw
a
y
l
a
r
g
c
o
d
e
d
e
v
e
l
o
p
m
e
n
ti
sp
o
s
s
i
b
l
eu
s
m
g
s
m
a
l
l
e
r
p
i
e
c
e
so
f
e
x
i
s
t
i
n
gc
o
d
e
.

P
(
.}=6-9+32
Theminimumf
o
rt
h
ep
o
l
y
n
o
m
i
a
li
st
h
es
o
l
u
t
i
o
nt
o
dP

=-9+6=0; :=
=1
.5
0 r
I
Ca
p
p
r
o
x
i
m
a
t
es
o
l
u
t
i
o
nt
oE
q
u
a
t
i
o
n(
5
.
1
)i
s1
.
5
.F
i
g
u 5.
3dcribest
h
ep
o
l
y
n
o
m
i
a
l
a
p
p
r
o
x
i
m
a
t
i
o
np
r
o
c
e
d
u
r
e
.l
ti
sc
l
e
a
ra
tt
h
ea
p
p
r
o
x
i
m
a
t
i
o
nl
e
a
v
e
smucht
obed
e
s
i
r
e
d
.

i
ft
h
ed
a
t
nf
o
rt
h
ep
o
l
y
n
o
m
i
a
lc
o
e
f
f
i
c
i
e
n
t
sw
e
r
ea
r
o
u
n
das
m
a
l
l
e
r
Ont
h
eo
t
h
e
rhand
r
e
g
i
o
nnear=2
.
6
t
h
e
n r
e
s
u
l
t
swouldbemorei
m
p
r
e
s
s
i
v
e
.N
o
t
ea
tr
a
m
u
s
cah
i
g
h
e
r
o
r
d
e
rp
o
l
y
n
o
m
i
a
l
ab
e
t
t
e
rs
e
tofd
a
t
ai
sp.
Thes
c
a
n
n
i
n
gp
r
o
c
e
d
u
r
eandep
o
l
y
n
o
m
i
a
la
p
p
r
o
x
i
m
a
t
i
o
nt
e
c
h
n
i
q
u
e
sw
i
l
lf
i
n
d
s
i
g
n
i
f
i
c
a
n
tu
s
ee
x
p
l
o
r
i
n
gm
u
l
t
i
d
i
m
e
n
s
i
o
n
a
lo
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
e
m
s
.Th
es
t
e
g
yf
o
r
o
s
s
i
b
i
l
i
t
yo
fc
o
d
er
e
u
s
e
.l
e
d
e
v
c
l
o
p
i
n
gMATLABc
o
d
cf
r
o
mnowonmusti
n
c
l
u
d
eep
r
e
n
tn
u
m
e
r
i
c
a
lt
e
c
h
n
i
q
u
e
s
.Suchg
e
n
e
r
i
c
samec
o
d
es
e
g
m
e
n
t
sc
a
nbeu
s
e
di
nmanyd
ie

e
t
c
n
n
i
n
a
t
i
o
no
fe
upperBound1Var.m:Thiscodcsegmenti
m
p
l
e
m
e
n
t
sed
upperboundoilhE
o
no
fonev
a
r
i
a
b
l
e
.Thei
n
p
u
tt
ot
h
ef
u
n
c
t
i
o
ni
st
h
e ncofthc

:
Z3;42;;;22dEmusthaM

cotZ

Bm-f
i
)
e

i
nesamed
i
r
e
c
t
o
r
y

t
Z
T
f
t
U
3
2
3
1
:
;
:
;
Z
?
5
5
1
2

tw
h
i
c
ht
h
cf
u
n
c
t
i
o
n

p
0
1
10)
U
s
a
g
e
: upperBound_1Var('EXam le5_1'

(
5.
12
)

cod

s
e
g
m
e
n
t i
m
p
l
e
m
e
n
t
s e p
o
l
y
n
o
m
i
a
l
i
s
PolyApprox_1Var.m: l
f
u
n
c
t
i
o
no
f one v
a
r
i
a
b
l
T
h
i
sf
u
n
c
t
i
o
nu
s
c
s
a
p
p
r
o
x
i
m
a
t
i
o
n method f
o
ra

I
l
J
'

5
.
2S
O
L
U
T
I
O
NT
OT
H
EP
R
O
B
L
E
M 2
1
5

2
1
4 N
U
M
E
A
I
C
A
LT
E
C
H
N
I
Q
U
E5-T
H
EO
N
E
.
O
I
M
E
N
S
I
O
N
A
LP
A
O
B
L
E
M
a
r
i
a
b
l
e
.Th
ei
n
p
u
tt
oe
UpperBound_1Var.mt
od
e
t
e
n
n
i
n
et
h
er
a
n
g
eo
fev
ft
h
ef
u
n
c
t
i
o
n
;t
h
eo
r
d
e
r(
2o
r3
)o
ft
h
ea
p
p
r
o
x
i
m
a
t
i
o
n
;
f
u
n
c
t
i
o
ni
st
h
enameo
lowbound-the s
t
a
r
tv
a
l
u
co
ft
h
es
c
a
np
a
s
s
e
dt
o UpperBound_1Var.m;
i
n
t
v
l
s
t
e
p s
c
a
n
n
i
n
gi
n
t
e
r
v
a
lp
a
s
s
e
dt
oUpperBound_1V
a
r
.
m
;i
n
i
a
l
s

n
u
m
b
e
ro
fs
c
a
n
n
i
n
gs
t
e
p
sp
a
s
s
e
dt
oUpperBound_1Var.m.Theo
u
t
p
u
to
fe
p
r
o
g
r
a
mi
sav
e
c
t
o
ro
ft
w
ov
a
l
u
e
s
.Thef
i
r
s
te
l
e
m
e
n
to
ft
h
ev
e
c
t
o
ri
st
h
el
o
c
a
t
i
o
no
ft
h
e
minimumo
ft
h
ea
p
p
r
o
x
i
m
a
t
i
n
gp
o
l
y
n
o
m
i
a
l
a
n
dt
h
es
e
c
o
n
di
st
h
ef
u
n
c
t
i
o
nv
a
l
u
ea
t
t
h
i
sl
o
c
a
t
i
o
n
.
Th
ef
u
n
c
t
i
o
n f
e
r
e
n
c
e
d
b
yec
o
d
em
u
s
tbeaMA
TI.A
Bmf
i
l
e

i
nt
h
es
a
m
ed
i
r
e
c
t
o
r
y
(Examp
5
_
1
.
m
)
.1
ei
n
p
u
tf
o
rE
x
a
m
p
l
e
5
_
1i
st
h
ev
a
l
u
ea
tw
h
i
c
ht
h
ef
u
n
c
t
i
o
n
n
e
e
d
st
ob
ec
o
m
p
u
t
e
d
.a
n
di
t
so
u
t
p
u
ti
st
h
ev
a
l
u
eo
ft
h
ef
u
n
c
t
i
o
n
.
U
s
a
g
c
:Value

:
2
;

1;12A
up/

1=(
1-'
t
)
a
!
o
w+
;
/
1=
/
(

.
)
;
ri+1
GoToS
t
e
p3

i
s
i
Z
!
o
n
l
yc
a
l
lateevalues.Thealgo mi
si
m
p
l
e
m
e
n
t
e
di
nad
i
f
f
e
r
e
n
tf
i
l
e
.

2
0
1
1
0
)
(
5.
13
)
PolyApprox_1Var('Example5_1'

ExampleS.l

Mi
n
i
m
i
z
e /()=(_1
)
2
(

2
)
(

3
)

Thet
w
op
r
o
g
r
a
m
su
s
es
w
i
t
c
h
/
c
a
s
ea
n
dt
h
e/
e
v
a
ls
t
a
t
e
m
e
n
t
sf
r
o
mM
A
T
L
A
B
.T
h
e
l
Iu
s
a
t
e
sc
a
l
l
i
n
ga
n
dr
e
t
u
m
i
n
gf
r
o
mo
t
h
e
rn
c
t
i
o
n
s
.
c
o
d
ea
l
s
oi

(
5
.
1
)

G
o
l
d
e
nS
e
c
t
i
o
nMethod
P=
1
.(1.10" o
.
o
;
l
o
w 6
;
u
4;!"'=1
8
S
t
L

<
X
j
!
n
o
l=0
.
0
0
1
;E=(
0
.
0
0
1
1
4
)
;N=1
7

5
.
2
.
5 GoldenS
e
c
t
i
o
nMethod
T
h
m
e
t
h
o
do
fg
o
l
d
e
ns
e
c
t
i
o
ni
st
h
cc
r
c
a
mo
ft
h
ef
a
m
i
l
yo
fi
n
t
e
r
v
a
l d
u
c
i
n
gm
e
t
h
o
d
s
i
s
e
c
t
i
o
nm
e
t
h
o
d
)
.I
tr
e
d
u
c
e
sei
n
t
e
r
v
a
lb
yt
h
cs
a
m
ef
r
a
c
t
i
o
nw
i
t
h
(
f
o
re
x
a
m
p
l
e
.eb
e
a
c
hi
t
e
r
a
t
i
o
n
.Thei
n
t
e
r
v
a
l
s ed
e
r
i
v
e
df
r
o
mt
h
eg
o
l
d
e
ns
e
c
l
i
o
nr
a
t
i
o
.1
.6
1
8
0
3
.T
h
i
s
s
t
h
e
t
i
c
sa
sw
e
l
la
sm
a
t
h
e
m
a
t
i
c
s[
3
4
]
.T
hem
e
t
h
o
di
ss
i
m
p
l
e
r
a
t
i
oh
a
ss
i
g
n
i
f
i
c
a
n
c
e a
.
tI
ti
si
n
d
i
f
f
e
r
e
n
tt
oes
h
a
p
ea
n
dc
o
n
t
i
n
u
i
t
yp
r
o
p
e
r
t
i
e
so
fef
u
n
c
t
i
o
n
t
oi
m
p
l
e
m
e
n
b
e
i
n
gm
i
n
i
m
i
z
e
d
.Mosti
m
p
o
r
t
a
n
t L
h
en
u
m
b
e
ro
fi
t
e
r
a
t
i
o
n
st
oa
c
h
i
e
v
eap
r
e
s

i
b
e
d
hei
t
e
r
a
t
i
o
n
ss
t
a
r
t
.
t
o
l
e
r
a
n
c
ec
a
nb
ce
s
t
a
b
l
i
s
h
e
db
e
f
o
r
eL

.5
2
7
9
;
/
.=0
.
1
9
3
7
; =2.
4
7
2
1
;
1
2
=0
.
5
4
0
1
S
t
e
p2..=1

F
1
g
u
r
e
54i
1
lu
s
t
r
a
t
e
st
h
l
a
y
o
u
to
fi
n
f
o
r
m
a
t
i
o
na
tt
h
i
ss
t
a
g
e
.
I
tc
a
nbCVErinda
t
i
c
a
l
l
yw
i
t
h
m
s
p
e
c
tt
oo
u
t
e
rb
o
u
n
d
sd
O

"
p
.

.
.

2a
r
el
o
c
a
t
e
dsymm

S
t
e
p3
.i
=1
/
1
>
1
2
(
1
.
l
o
lV=1
.5
2
7
9
;
l
o
w=0.
19
3
7
; =2.
4
7
2
1
;
/
.=-0.
5
4
0
1

A
l
g
o
r
i
t
h
m
f
o
rG
o
l
d
e
nS
e
c
t
i
o
l
lMethod(
A
S
.
3
)
S
t
e
p1
.C
h
o
o
s
e
l
o
w

.
0
5
5
7
;
1
2=0
.
2
4
8
6

2=3

=0
.
3
8
1
9
7(
f
r
o
mG
o
l
d
e
nR
a
t
i
o
)
e=t
o
l
e
r
a
n
c
e=(
L

)
O
n
a
l
(

u
p
_

l
o
w
)
/
N=n
u
m
b
e
ro
fi
t
e
r
a
t
i
o
n
s=2
.
0
7
8l
ne

i=2
S
t
e
p3
.c
o
n
t
i
n
u
e
d
.
.
.

1
2
>
1
3

up=3
a
.
.
0
5
5
7
; =
0
.
2
4
8
6
;

2=2.
4
7
2
1
;=-0
.
5
4
0
1

I
f
h
>
/
.

dijjY35233:

)
l
o
w
+
;
f
l=
/
(

1
)
Sp2.1=(1OW
P
f
2=
=
r
i +(l
)
U ;
)
ncp
o
i
n
t
sa
r
ee
q
d
i
s
t
a
n
tf
r
o
mb
o
u
n
d
s
S
t
e
p3
.I
f(i<N)
I
f
/
/
2
I>
10"
.
;

1
2
;
f
i1
2
~=Iow +(
1
)
=/()
ii
+1
GoToS
t
e
p3

.
1
1
1
5
;
1
2=
0
.
1
2
2
4

1=2
i=3
GoToS
t
e
p3e
t
c
I
ne
a
c
hi
t
c
r
a
t
i
o
nL
h
e
r
ei
so
n
l
yo
n
e o
f
c
a
l
l
a
t
i
o
nt
oe
v
a
a
evariableand
v
a
l
u
eo
f
t
h
ef
u
n
c
t
i
o
n
.
T
h
er
e
s
to
f
t
h
E
C
o
m
p
u
t
a
u
o
n
t
sm
a
s
S
I
g
n
s
n
ge
x
t
s
t
i
n
gi
n
f
o
r
m
a
t
I
o
n
.
Th
ed
i
f
f
e
r
c
n
o
u
b
omdsi
s
a
l
s
o
c
h
a
n
g
i
n
g
w
i
t
h
e
a
c
hkmuon-hExmpleal

S
t
a
r
t
:

=4.0

=4
.
0-1
.
5
2
7
9=2.
4
7
2
1
I
t
e
r
a
t
i
o
n1
: L

2
1
6 N
U
M
E
A
I
C
A
LT
E
C
H
N
I
Q
U
ES-T
H
EO
N
E
.
D
I
M
E
N
S
I
O
N
A
LP
A
O
B
L
E
M

5
.
3I
M
P
O
A
T
A
N
C
EO
FT
H
EO
N
E
.
D
I
M
E
N
S
I
O
N
A
LP
A
O
B
L
E
M 2
1
7

2
0

v
e
c
t
o
ro
ff
o
u
rp
a
i
r
so
fv
a
r
i
a
b
l
ea
n
df
u
n
c
t
i
o
nv
a
l
u
e
sa
f
t
e
rt
h
ef
i
n
a
li
t
e
r
a
t
i
o
n
.W
i
t
ha
s
m
a
l
lt
o
l
e
r
a
n
c
e yo
n
eo
fl
h
e
s
ep
a
i
r
sp
r
o
v
i
d
e
s minimuml
o
c
a
t
i
o
no
ft
h
ef
u
n
c
t
i
o
n
.
tt
h
es
e
c
o
n
dp
a
i
rb
ea
c
c
e
s
s
e
df
o
rt
h
i
sp
u
o
s
e
.1
ef
u
n
c
t
i
o
n
I
ti
sr
e
c
o
m
m
e
n
d
e
da
i
nt
h
es
a
m
ed
i c
t
o
r
y
r
e
f
e
r
e
n
c
e
db
yt
h
ec
o
d
em
u
s
tb
e a MATLAB m
f
i
l
e
(
E
x
a
m
p
l
e
5
_
1
.
m
)
.Thei
n
p
u
lf
o
rExample5_1.mi
st
h
ev
a
l
u
ea
lw
h
i
c
ht
h
ef
u
n
c
t
i
o
n
n
e
e
d
st
ob
ec
o
m
p
u
t
e
d
a
n
di
t
so
u
t
p
u
ti
st
h
ev
a
l
u
eo
ft
h
ef
u
n
c
t
i
o
n
.

1
8
1
8

1
4

'
t

1
2
{6}h

1
0

U .
g
e
:

Value

'

T
h
e
r
ea
r
en
onewc
o
m
m
a
n
d
s
.Thco
r
g
a
n
i
z
i
n
ga
n
dd
i
s
p
l
a
y
i
n
go
fi
n
f
o
r
m
a
t
i
o
ni
nt
h
e
ft
h
i
sc
o
d
ei
sp
a
r
to
fal
a
r
g
e
rp
r
o
g
r
a
m t
h
e
Commandwindowi
sw
o
r
t
hn
o
t
i
n
g
.I
p
r
i
n
t
i
n
go
fi
n
f
o
r
m
a
t
i
o
ni
nt
h
ecommandwindown
e
e
d
st
ob
es
w
i
t
c
h
e
do
f
f
.I
tc
b
e
m
edo
nf
o
rd
e
b
u
g
g
i
n
g
.

'J

up
2

F
l
g
u
r
5
.
4G
o
l
d
e
n5
e
c
t
l
on-s
t
a
r
t
i
n
gv
a
l
u
a
s
.
I
t
e
r
a
t
i
o
n2
: d=3
.5
2
7
8
.
0
5
5
7-15279=1

Thef
i
n
a
li
t
e
r
a
t
i
o
ni
sr
e
p
r
o
d
u
c
e
db
yr
u
n
n
i
n
gt
h
ef
o
l
l
o
w
i
n
gc
o
d
e
:
i
t
e
r
a
t
i
o
n1
7
2
.
6
3
9
5
2
.
6
3
9
9
0
.
6
1
9
7 -0.
6
1
9
7

2
.
6
4
0
2
-0.
6
1
9
7

GoldSection_1Var(
'ExampleS_1'
0.001
0
1
1
0
)(
5
.
1
4
)

2
.
6
4
0
6
0
.
6
1
9
7

l
el
o
l
e
r
a
n
c
e

2
Ii
nt
h
ef
i
n
a
li
t
e
r
a
t
i
o
ni
sa
b
o
u
t0
.
0
0
1a
se
x
p
e
c
l
e
d
.

Gold
Se
c
t
i
o
n
_
1
V
a
r
.
m
: Thec
o
d
cl
r
s
l
a
t
e
sea
l
g
o
r
i
t
h
mf
o
rt
h
eg
o
l
d
e
ns
t
i
o
n
m od(A5
.
3
)i
n
TLABcode.Theinput1
0t
h
ef
u
n
c
t
i
o
ni
senameo
ft
h
cf
u
n
c
t
i
o
n
tnCt
n
a
m
e
)w
h
o
s
eminimumi
sb
e
i
n
gs
o
u
g
h
t
;et
o
l
c
r
a
n
c
e(
t
o
l
)o
ft
h
ea
p
p
r
o
x
i
m
a
t
i
o
n
;
value(

l
o

w
bo
und)o
fes
c
p
a

e
dt

oU
p
p
e
r
B
ound_1V
a
r
.
m

1
;
t
h
e
s
c
a
n
mn
g
i

n
t

e
r
v
(

u
t
1
)pa
5
5
e

dtoU

p
perBo

u
md_1V
a
r

'
.
m
;
t

he

u
m

E
p
s

1
1
p
t
oUpperBound_1V .
m
.Theo
u
t
p
u
to
f p
r
o
g
r
a
mi
s;

Compar/sonw
l
t
hPolynom/alApproxlmatlon:Thet
w
op
o
p
u
l
a
rm
e
t
h
o
d
sf
o
r
m
i
n
i
o
u
z
i
n
gaf
u
n
c
t
i
o
no
fo
n
ev
a
r
i
a
b
l
ea
r
ep
o
l
y
n
o
m
i
a
la
p
p
r
o
x
i
m
a
t
i
o
no
rg
o
l
d
e
n
o
l
d
e
ns
e
c
t
i
o
nh
a
v
e
s
e
c
t
i
o
n
.Thea
l
g
o
r
i
l
h
m
sf
o
rp
o
l
y
n
o
o
u
a
la
p
p
r
o
x
i
m
a
t
i
o
na
n
d g
o
r
m
e
ri
sao
n
e
s
h
o
ta
p
p
r
o
a
c
ha
n
di
sa
c
c
o
m
p
a
n
i
e
db
y
s
i
g
n
i
f
i
c
a
n
td
i
f
f
e
r
e
n
c
e
s
.Thef
w
h
i
c
hw
i
l
li
m
p
r
o
v
ea
st
h
ed
a
t
af
o
r
s
i
g
n
i
f
i
c
a
n
te
r
r
o
ri
nt
h
ee
s
t
i
m
a
t
i
o
no
ft
h
eminimum
a
p
p
r
o
x
i
m
a
t
i
o
ng
e
tb
e
t
t
e
r
.Imp
l
i
e
di
nt
h
ei
m
p
l
e
m
e
n
t
a
t
i
o
no
ft
h
ep
o
l
y
n
o
m
i
a
la
p
p
r
o
a
c
h
i
st
h
ec
o
n
t
i
n
u
i
t
yo
fef
u
n
c
t
i
o
n
.Theg
o
l
d
e
ns
e
c
t
i
o
nm
e
t
h
o
d

o
nt
h
eo
t
h
e
rh
a
n
d

i
s
i
t
e
r
a
t
i
v
et
e
c
h
n
i
q
u
e
.Then
u
m
b
e
ro
fi
t
e
r
a
t
i
o
n
sd
e
p
e
n
d
so
ne
t
o
l
e
r
a
n
c
e
x
p
e
c
t
e
di
ne
u
l
ta
n
di
sknownp
r
i
o
rt
o s
t
a
r
to
ft
h
ei
t
e
r
a
t
i
o
n
s
as
i
g
n
i
f
i
n
ti
m
p
r
o
v
e
m
e
n
t
f
i
n
a
l s
u
m
b
e
ro
fi
t
e
r
a
t
i
o
n
sc
a
n
n
o
tb
e
r
e
l
a
t
i
v
et
ot
h
eN
e
w
l
o
n
R
a
p
h
s
o
nm
e
t
h
o
dw
h
e
r
e n
p
r
e
d
i
c
t
e
dap
r
i
o
r
i
.Thei
m
p
l
e
m
e
n
t
a
t
i
o
ni
ss
i
m
p
l
ea
n
dt
h
er
e
s
u
l
t
si
m
p
r
e
s
s
i
v
ea
si
ti
sa
b
l
e
1
u
m
.I
ti
si
n
d
i
f
f
e
r
e
n
tt
ot
h
en
a
t
u
r
eo
ft
h
ef
u
n
c
t
i
o
n
.
tohomeinon oun
T
h
e
r
ei
sn
or
e
a
s
o
nwhyt
h
et
w
oc
a
n
n
o
tb
ec
o
m
b
i
n
e
d
.

b
eg
o
l
d
e
ns
e
c
t
i
o
nc
a
nb
e
in
s
t
e
a
do
fal
o
wv
a
l
u
e
)
u
s
e
dt
oe
s
t
a
b
l
i
s
hf
o
u
rd
a
t
ap
o
i
n
t
sw
i
t
har
e
a
s
o
n
a
b
l
et
o
l
e
r
a
n
c
e(
a
n
dac
u
b
i
cp
o
l
y
n
o
o
u
a
lc
a
nb
ef
i
t1
0i
d
e
n
t
i
f
yt
h
ea
p
p
r
o
x
i
m
a
t
el
o
c
a
t
i
o
no
fm
i
n
i
m
u
m
.

5
.
3 IMPORTANCEOFTHEONE-DIMENSIONALPROBLEM
1

o
n
e
d
i
m
e
n
s
i
o
n
a
ls
u
b
p
r
o
b
l
e
mi
n am
u
l
t
i
v
a
r
i
a
b
l
eo
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
e
mi
s
e
t
e
r
m
i
n
a
t
i
o
no
ft
h
es
t
e
p
s
i
z
ea
f
t
e
rt
h
cs
e
a
r
c
hd
i
r
e
c
t
i
o
nh
a
sb
e
e
n
e
m
p
l
o
y
e
df
o
red
m
i
d
e
n
t
i
f
i
e
d
.I
ti
se
a
s
i
e
r1
0u
n
d
c
r
s
t
a
n
dt
h
e
s
el
e
r
m
sb
yr
e
c
o
g
n
i
z
i
n
gt
h
eg
e
n
e
r
i
ca
l
g
o
f
o
ru
n
c
o
n
s
t
r
a
i
n
o
p
t
i
o
u
z
a
t
i
o
n
.I
no
r
d
e
rt
of
o
c
u
st
h
ed
i
s
c
u
s
s
i
o
nr
a
t
h
e
rt
h
a
ne
g
e
n
e
r
a
lo
b
j
c
c
t
i
v
ef
u
n
c
t
i
o
nr
e
p
r
e
s
e
n
t
a
t
i
o
n
as
p
e
c
i
f
i
co
n
e
E
x
a
m
p
l
e5
.
2
i
si
n
t
r
o
d
u
c
e
d
.

Example5.2

M
i
n
i
m
i
z
e

f
(
x
l
X2'X
)=(
x
l)
2+2{~ -x3)2+3(x
)
2
3- 1
3

(
5
.
1
5
)

218

NUMERICALTECHNIQUE5THEONE-DIMENSIONALPROBLEM

219

5
.
4 AODITIONALEXAMPLES

Ac
o
r
yg
l
a
n
c
ea
tExample5
.
2i
n
d
i
c
a
t
e
sa
t
t
h
e
s
o
l
u
t
i
o
n
i
s
a
t
x
l=l;x2=I;X3=1
;
andt
h
eminimumv
a
l
u
eo
f
f
i
sO
.Thr
e
ev
a
r
i
a
b
l
e
sa
r
ec
h
o
s
e
nb
e
c
a
u
s
ewea
r
n
o
t
g
o
i
n
g
iti
se
s
s
e
n
t
i
a
lt
of
o
l
l
o
wt
h
ev
e
c
t
o
ra
l
g
e
b
r
a
)
.
t
ou
s
ea
n
yg
r
a
p
h
i
c
sf
o
ri
l
l
u
s
t
r
a
t
i
o
n(

GenerlcAlgorlthm(
A
5
.
4
)
:Theg
e
n
e
r
i
ca
l
g
o
r
i
t
h
mi
sa
ni
t
e
r
a
t
i
v
eoneandi
sa
l
s
o
r
e
f
e
r
d
toas
e
a
r
c
ha
l
g
o
r
i
t
h
m
.a
st
h
ei
t
e
r
a
t
i
o
n
st
a
k
p
l
a
c
eb
ymovinga
l
o
n
gas
e
a
r
c
h
d
i
r
e
c
t
i
o
n
.Theses
e
a
r
c
hd
i
r
e
c
t
i
o
n
sc
a
nb
ed
e
t
e
r
m
i
n
e
di
ns
e
v
e
r
a
lways(
C
h
a
p
t
e
r6
)
.The
a
l
g
o
r
i
t
h
mw
i
t
h
o
u
tanyc
o
n
v
e
r
g
e
n
c
e
l
s
t
o
p
p
i
n
gc
r
i
t
e
r
i
ac
a
nb
ee
x
p
r
e
s
s
e
da
s
S
t
e
p1
.ChooseXo
S
t
e
p2
.F
o
re
a
c
hi
t
e
r
a
t
i
o
ni
D
e
t
e
r
m
i
n
es
e
a
r
c
hd
i
r
e
c
t
i
o
nS
Jv
e
c
t
o
r
S
t
e
p3
.C
a
l
c
u
l
a
t
es
X;=jSj

X;i
snowaf
u
n
c
t
i
o
no
ft
h
es
c
a
l
a
rc
l
ja
sS
ji
sknownfromS
t
e
p2
N
o
t
e
:s
c
l
ji
sc
a
l
l
e
dt
h
es
t
e
ps
i
z
ea
si
te
s
t
a
b
l
i
s
h
e
st
h
el
e
n
g
t
ho
fsX
j

j1=X/+sX
X
i
sd
e
t
e
r
m
i
n
e
dbyM
i
n
i
m
i
z
i
n
gF
(
X
i
+
l
)
i

Xl=lL

(
5
.
1
6
)

a
n
ds
oo
n
.

5
.
4 ADDITIONALEXAMPLES
h i
ss
e
c
t
i
o
na
d
d
i
t
i
o
n
a
le
x
a
m
p
l
e
so
fs
i
n
g
l
e
v
a
r
i
a
b
l
eo
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
e
m
sa
r
c
e
x
p
l
o
r
e
d
.Thea
v
a
i
l
a
b
l
ec
o
d
ef
o
rt
h
ep
o
l
y
n
o
m
i
a
la
p
p
r
o
x
i
m
a
t
i
o
na
n
dg
o
l
d
e
ns
e
c
t
i
o
n
s
h
o
u
l
dh
e
l
ps
o
l
v
ea
n
ya
n
da
l
ls
i
n
g
l
e
v
a
r
i
a
b
l
eo
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
e
m
s
.Thi
si
sl
a
r
g
e
l
y
.
tI
nt
h
i
ss
e
c
t
i
o
no
n
l
ye
x
t
e
n
s
i
o
n
s
m
o
d
i
f
i
c
a
t
i
o
n
s
o
rc
r
e
a
t
i
v
e
l
e
f
tt
ot
h
er
e
a
d
e
rt
oe
x
p
l
o
i
a
p
p
I
i
c
a
t
i
o
n
so
ft
h
eo
n
e
d
i
m
e
n
s
i
o
n
a
lo
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
e
ma
r
c
o
n
s
i
d
e
r
e
d
.F
i
r
s
t

Example5
.
2i
sr
e
v
i
s
i
t
e
dt
om
o
d
i
f
yt
h
eg
o
l
d
e
ns
e
c
t
i
o
nmethods
oa
ti
tc
a
nbeu
s
e
dt
o
l
c
u
l
a
t
e S p
s
i
z
ei
nm
u
l
t
i
d
i
m
e
n
s
i
o
n
a
lp
r
o
b
l
e
m
s
.Example5.
3i
sas
o
l
u
t
i
o
nt
ot
h
e
i
si
sa
ne
x
a
m
p
l
eo
ft
h
ee
x
a
c
ts
o
l
u
t
i
o
nt
oeN
a
v
i
e
r
S
t
o
k
e
s
B
l
a
s
s
i
u
sp
r
o
b
l
e
m
. l
u
a
t
i
o
nf
o
rf
l
o
wo
v
e
raf
l
a
tp
l
a
t
c
.Example5.
4i
sa
ne
x
a
m
i
n
a
t
i
o
no
ft
h
ei
n
e
q
u
a
l
i
t
y
c
o
n
s
t
r
a
i
n
ti
nExamplc5
.
l
aby t
h
eg
o
l
d
e
ns
e
c
t
i
o
n methods
ot
h
a
te
q
u
a
l
i
t
ya
n
d
i
n
e
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
t
sc
a
nb
eh
a
n
d
l
e
d
.

Iisi
sr
e
f
e
r
r
e
dt
oa
so
n
e
d
i
m
n
s
i
o
n
a
ls
t
e
ps
i
z
ec
o
m
p
u
t
a
t
i
o
n

ii+1
;00ToS
t
e
p2

ample5.2I
l
l
u
s
t
r
a
t
l
o
nofGeneralGoldenSectlonMethod
5
.
4
.
1 Ex

'
h
ee
x
a
m
p
l
ef
r
o
mep
r
e
v
i
o
u
ss
e
c
t
i
o
ni
sv
i
s
i
t
e
da
g
a
i
nt
om
o
d
i
f
yeg
o
l
d
e
ns
e
c
t
i
o
n

A
p
p
l
i
c
a
l
i
o
n0
/GenericA hm(A5.4)EXIlITlp
l
e5
.
2

~[g}
..

n
u

a
u

rJ

et
o

t
n9uw

eb

1Ill1
t
10
10

EBEBEE--EEEJ

drllML

.006.

grill-L
n .

E -'EEEEEL

LzrE

+
J

MMM

U11111111

='J

F1111111L=

J
dJ

guAU

-a
-n a
p

n
er
enr

f(X1
(

1
)=(
0-0
)
2+2
(
0-6

1
)
2+3
(
6

1 1
)
2=72l+3(6J-1
)
2
)=f

methodtod
t
c
r
m
i
n
ees
t
e
p
s
i
z
ef
o
rm
u
l
t
i
d
i
m
e
n
s
i
o
n
a
lp
r
o
b
l
e
m
. i
sbooki
sa
l
s
o
a
b
o
u
tg
e
t
t
i
n
gMATLABt
oworkf
o
ry
o
u
.I
nas
c
n
s
e ss
u
b
s
e
c
t
i
o
ni
sl
a
r
g
e
l
ya
ne
x
e
r
c
i
s
e
i
nd
e
v
e
l
o
p
i
n
gt
h
ec
o
d
c w
i
l
ls
e
ec
o
n
s
i
d
e
r
a
b
l
eu
s
ei
nt
h
en
e
x
tc
h
a
p
t
e
r
.An u
i
v
a
l
e
n
t
o
l
y
n
o
m
i
a
la
p
p
r
o
x
i
m
a
t
i
o
nmethodi
sl
e
f
ta
sa
n
d
e
v
e
l
o
p
m
e
n
tw
i
t
hr
e
s
p
e
c
tt
oep
.
t
e
x
e
r
c
i
s
ef
o
rt
h
es
t
u
d
e
n

'
h
eg
o
l
d
c
ns
e
c
t
i
o
n method r
e
q
u
i
r
e
st
h
ee
s
t
a
b
l
i
s
h
m
e
n
to
ft
h
eu
p
p
e
rb
o
u
n
d
.

'
h
r
e
f
o
r
etwof
u
n
c
t
i
o
n
swi
l
1n
e
e
dt
ob
ec
h
a
n
g
e
dt
oh
a
n
d
l
ed
c
s
i
g
nv
e
c
t
o
r
sa
n
dt
h
e
s
e
a
r
c
hd
i
r
e
c
t
i
o
nv
e
c
t
o
r
s
.Thec
o
d
ei
sa
v
a
i
l
a
b
l
e UpperBound_nVar.m d
GoldSectlon_nVar.m.Them
o
d
i
f
i
c
a
t
i
o
n
sa
r
en
O
lv
e
r
yc
h
a
l
l
e
n
g
i
n
ga
n
dc
a
nb
e
i
n
f
e
r
r
e
dbyc
o
m
p
a
r
i
n
gt
h
eu
s
a
g
eo
fetwof
u
n
c
t
i
o
n
s
.
Fort
h
es
i
n
g
l
ev
a
r
i
a
b
l
e
:
0
1
10) (
5.
12
)
U
s
a
g
e
:UpperBound_1Var(
'Example5_1'
F
o
rmanyv
a
r
i
a
b
l
e
s
:

M
i
n
i
m
i
z
i
n
gj
{
1
)cv
a
l
u
eo
f
1=0
.1
.T
h
i
si
sao
n
e
d
i
m
e
n
s
i
o
n
a
l
o
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
e
m
.
Whilet
h
es
o
l
u
t
i
o
na
b
o
v
ewaso
b
t
a
i
n
e
da
n
a
l
y
t
i
c
a
l
l
yb
e
c
a
u
s
ei
twasas
i
m
p
l
el
i
n
e
a
r
e
x
p s
s
i
o
nep
o
l
y
n
o
m
i
a
la
p
p
r
o
x
i
m
a
t
i
o
no
rt
h
eg
o
l
d
e
ns
e
c
t
i
o
nmethodc
o
u
l
dh
a
v
e
a
d
.
b
e
e
nu
s
e
di
n
s
t

x
s
O
l
0
) (
5.
17
)
U ge:UpperBound_nVar('Example5_2'
x: current position vector or design vector
s
: prescribed search direction

220 N
U
M
E
R
I
C
A
LT
E
C
H
N
I
O
U
E5--T
H
EO
N
E
D
I
M
E
N
S
I
O
N
A
LP
R
O
B
L
E
M
Ofc
o
u
r
s
eExample
5
_2.mr
e
t
u
m
st
h
ev
a
l
u
o
fef
u
n
c
t
i
o
nc
o
r
r
e
s
p
o
n
d
i
n
gt
oa
v
e
c
t
o
ro
fd
e
s
i
g
nv
a
r
i
a
b
l
e
s
.l
i
si
m
p
l
i
e
s t
h
ef
u
n
c
t
i
o
nc
a
l
lt
oExample5_2.mc
a
n
o
n
l
yt
a
k
ep
l
a
c
a
f
t
e
rt
h
ed
e
s
i
g
nv
a
r
i
a
b
l
e
sa
r
ee
v
a
l
u
a
t
e
dfromev
a
l
u
eo
fs
t
e
p
s
i
z
e
a
p
p
l
i
e
da
l
o
n
gep
r
e
s
c
r
i
b
e
ds
e
a
r
c
hd
i c
t
i
o
nf
r
o
mt
h
ec
u
r
r
e
n
tl
o
c
a
t
i
o
no
fed
e
s
i
g
n
.
17
)ande
Theu
s
a
g
eo
fG
o
l
d
S
e
c
t
i
o
n
_
n
V
a
r
.
mmustaccommodatef
u
n
c
t
i
o
n(
5.
n
e
e
dt
od
e
a
lw
i
t
had
e
s
i
g
nv
e
c
t
o
r
.Comparingw
i
t
ht
h
es
i
n
g
l
e
v
a
r
i
a
b
l
ec
a
s
e

U g
e
:
Value

GoldSection_1Var('Example5_1'
0.001
0
1
10)
(
5
.
1
4
)

U g
e
:

0.001
x
s
0
1
10)
GoldSection_nVar('Example5_2'
(
5
.
1
8
)
V
i
s
i
t c
o
d
et
os
t
hed
1
5
.
Value

Compar/sonw
/
t
hS
e
c
t
l
o
n5
.
3
:Th
ec
o
d
eGoldSectlon_nVar.mwasr
u
nf
r
o
m
eCommandwindowu
s
i
n
gt
h
ef
ol
1owingl
i
s
t
i
n
g
:

>>x
[
000]
;
>>S
[0 0 6];
>> GoldSection_nVar('Example5_2'
0.001
x
s
0
1
10)
in
ei
nt
h
eCommandwindowwas
A
f
t
e
r1
1i
t
e
r
a
t
i
o
n
st
h
ef
i
n
a
lJ

ans CI
0.1000

5
.
4 A
D
D
I
T
I
O
N
A
LE
X
A
M
P
L
E
S 2
2
1

MathematlcalFormulatlon

f
f
"+2f"=
0

(
5
.
2
0
)

x=O; f
(
O
)
= 0
; f(O)=O

(
5
.
2
1
a
)

x=

f( }=1

(
5
.
2
1
b
)

Th
es
o
l
u
t
i
o
n
n
o
n
d
i
m
e
n
s
i
o
n
a
lv
e
l
o
c
i
t
yi
nt
h
eb
o
u
n
d
a
r
yl
a
y
e
r

i
so
b
t
a
i
n
e
da
s

=f(x)

(
5
.
2
2
)

Th
es
o
l
u
t
i
o
nt
oE
q
u
a
t
i
o
n
s(
5
.
2
0
)t
o(
5
.
2
1
)i
sl
a
r
g
e
l
yt
h
r
o
u
g
hs
p
e
c
i
a
li
t
e
r
a
t
i
v
e
t
e
c
h
n
i
q
u
c
s
. These t
e
c
h
n
i
q
u
e
su
s
e an
u
m
e
r
i
c
a
li
n
t
e
g
r
a
t
i
o
np
r
o
c
e
d
u
r
e li~e
R
u
n
g
e
K
u
t
t
amethodt
oi
n
t
e
g
r
a
t
et
h
es
y
s
t
e
mbyg
u
e
s
s
i
n
ganda
d
j
u
s
t
i
n
gt
h
em
i
s
s
i
n
g
b
o
u
n
d
a
r
yc
o
n
d
i
t
i
o
n
sa
tt
h
ei
n
i
t
ia
1p
o
i
n
ts
u
c
h t
h
ef
i
n
a
lp
o
i
n
tb
o
u
n
d
a
r
yc
o
n
d
i
t
i
o
n
s
e
a
1
i
z
e
d
.Thi
si
sduet
oef
a
c
tt
h
a
tt
h
ei
n
t
e
g
r
a
t
i
o
np
r
o
c
e
d
u su
s
u
a
l
l
ys
o
l
v
ei
n
i
t
i
a
l
a
r
er
n
l
y
.I
nt
h
eB
l
a
s
s
i
u
sp
r
o
b
l
e
m i
swouldi
m
p
l
yt
h
em
i
s
s
i
n
gi
n
i
t
i
a
l
v
a
l
u
eproblemso
e
g
a
r
d
e
d ad
e
s
i
g
nv
a
r
i
a
b
l
e
.
v
a
l
u
ec
a
nber
('(0)=

(
5
.
2
3
)

Theo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
nt
h
e
r
e
f
o
r
ew
i
l
lbe
M
i
n
i
m
i
z
e
: () 1
)
2

(
5
.
2
4
)

Im
p
l
i
e
di
nt
h
ef
o
r
m
u
l
a
t
i
o
ni
s l
f
(
)i
so
b
t
a
i
n
e
dbyi
n
t
e
g
r
a
t
i
n
gt
h
ed
i r
e
n
t
i
a
l u
a
t
i
o
n
s
1.2000

o0.5973

Th
ea
b
o
v
ei
si
n
t
e
r
p
r
e
t
e
da
s
:

1=0
.
1
0
0
0
; f(
1
)=1
.2

5
.
1
9
)
x1=0; A
2=0.0; x3=0.5973 (

whichm
a
t
c
h
e
sS
e
c
t
i
o
n5.
3a
n
dEq
u
a
t
i
o
n(
5
.
1
6
)
.

ec
a
nb
ec
o
n
s
i
d
e
r
e
da
sd
i
f
f
e
r
e
n
t
i
a
lc
o
n
s'
a
i
n
).l
e
r
ei
so
n
eo
t
h
c
r
(
5
.
2
0
)(
c
o
n
s
i
d
e
r
a
t
i
o
nt
ot
a
k
ei
n
t
oa
c
c
o
u
nt
.
l
ei
n
t
e
g
r
a
t
i
o
nm
e
t
h
o
d
sr
e
q
u
i
r
et
h
eublemt
ob
c
e
x
p
r
e
s
s
e
di
ns
t
a
t
es
p
a
f
o
r
m
whichr
n
-o
r
d
d
i
f
f
e
r
e
n
t
ia1u
a
t
i
o
n1
0be
e
x
p
r
e
e
da
sas
y
s
t
e
mo
fn l
o
r
d
c
r u
a
t
i
o
n
s
. l
ec
o
n
v
e
r
s
i
o
ni
sf
a
i
r
l
ys
t
a
n
d
a
r
da
n
di
s
o
r
m
a
t
i
o
na
n
di
n
u
u
c
i
n
ga
d
d
i
t
i
o
n
a
lv
l
e
s
.F
o
rt
h
i
se
x
a
m
p
l
e
:
d
o
n
et
h
r
o
u
g
hulSf

y =f

5
.
4
.
2 Ex
ample5
.3Two-PolntBoundaryValueProblem
Ar
e
a
1examplef
o
ro
n
e
v
a
r
i
a
b
l
eo
p
t
i
m
i
z
a
t
i
o
nc
a
nb
ef
o
u
n
di
nen
u
m
e
r
i
c
a
ls
o
l
u
t
i
o
n
o
ft
h
el
a
m
i
n
a
rf
l
o
wo
v
e
raf
l
a
tp
l
a
t
e[
5
]
. l
ep
r
o
b
l
e
m
u
s
u
a
l
l
ya
t
t
r
i
b
u
t
dt
oB
l
a
s
s
i
u
s
r
e
p
r
e
s
e
n
t
sa
nexampleo
ft
h
ee
x
a
c
ts
o
l
u
t
i
o
no
ft
h
ef
o
r
m
i
d
a
b
l
eN
a
v
i
e
r
S
t
o
k
e
sc
q
u
a
t
i
o
n
.
l
em
a
t
h
e
m
a
t
i
c
a
lfom
lUl
a
t
i
o
n
a
l
l
o
w
i
n
gf
o
rs
i
m
i
l
a
r
i
t
ys
o
l
u
t
i
o
n
s
i
se
x
p
s
s
e
dbya
t
h
i
r
d
o
r
d
e
rn
o
n
l
i
n
e
a
ro
r
d
i
n
a
r
yd
i
f
f
e
r
e
n
t
ia
1e
q
u
a
t
i
o
nw
i
t
hb
o
u
n
d
a
r
yc
o
n
d
i
t
i
o
n
s
s
p
i
f
i
e
da
ttwop
o
i
-at
w
o
p
o
i
n
tb
o
u
n
d
a
r
yv
a
l
u
eproblemPBVP). Wh
a
t
f
o
l
I
owsi
st
h
ee
s
s
e
n
t
i
a
1m
a
t
h
e
m
a
t
i
c
a
ld
e
s
c
r
i
p
t
i
o
n
.l
ei
n
t
e
t
e
dr
e
a
d
e
rc
a
nf
o
l
l
o
wt
h
e

i
n
c
1
u
d
i
n
g s
u
g
g
e
s
t
e
dr
e
f
e
n
c
e
.
d
e
t
a
i
l
si
nmostbooksonf
l
u
i
dm
e
c
h
a
n
i
c
s

y~ =f=Y2

Y;=f
'=Y3
Y;=f"='/2
Th
eo
p
t
i
m
i
z
a
t
i
o
nc
a
nb
ef
o
r
m
u
l
a
t
e
d:F
i
n
d
M
i
n
i
m
i
z
e
:

[
Y
2
(

)
lf

u
s
i
n
gt
h
eR
u
n
g
e
K
u
t
t
amethod(
o
r yo
t
h
e
rm
e
t
h
o
d
)t
oi
n
t
e
g
r
a
l
e

(
5
.
2
5
)

222

EEEEEEaEEEEEJ

rEE E -E E -EL

o
o

1BEEt-EEEBEEEEJ

)
)
)
n
vn
U4
vi
(
(
(
-

'

fwlw
l
v

v
d3

-EEEEEEEEEEE

AU

. 3
EJ
UJUJgJI

JI11111J

r
i
I
r

N
U
M
E
R
I
C
A
LT
E
C
H
N
I
Q
U
E8-T
H
EO
N
E
.
D
I
M
E
N
S
I
O
N
A
LP
R
O
B
L
E
M

5
.
4A
D
D
IT
lO
N
A
LE
X
A
M
P
L
E
S

Ther
e
s
u
l
te
x
p
e
c
t
e
dis=0.
33
2
0
6
.Howevert
h
et
o
l
e
r
a
n
c
es
p
e
c
i
f
i
e
dwaso
n
l
y
0
.
0
01
. The i
n
i
t
i
a
l bounding i
n
t
e
r
v
a
l was o
n
l
y 0.
4
. Th
eg
o
l
d
e
ns
e
c
t
i
o
nh
a
s
l
s
on
o
t
et
h
ec
o
n
v
e
n
i
e
n
c
eo
f
p
r
o
d
u
c
e
da
ni
m
p
r
e
s
s
i
v
es
o
l
u
t
i
o
nt
oeTPBVP.A
r
u
m
m
i
n
g
. No c
h
a
n
g
e
s we
m
adet
o UpperBound_1Var.m o
r
modular pl'Og
GoldSectlon_1Var.mt
or
u
nt
h
i
se
x
a
r
n
p
l
e
.

(
5
.
2
6
)

=0.33206.Insleadofenalpointat i
t
Fromp
u
b
l
i
s
h
e
dwork[
5
Jt
h
ev
a
l
u
eo
f

a
v
i
n
gi
sv
a
l
u
ea
sav
a
r
i
a
b
l
ec
a
nd
e
f
i
n
eac
o
r
r
e
s
p
o
n
d
i
n
g
i
sc
o
n
s
i
d
e
r
e
d1
0b
e5
.
0
.Le
m
p
l
e
m
e
n
t
a
t
i
o
nt
of
o
l
l
o
wepower d
t
w
o
v
a
r
i
a
b
l
eo
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
e
m
.I
nei
1LA
Ba
r
r
e
a
d
i
l
y
a
p
p
a
r
e
n
t
c
o
n
v
e
n
i
e
n
c
eo
fMA

.
4RootFindlngwithGoldenSectlon
5
.
4
.
3 Example5
Th
ea
p
p
l
i
c
a
t
i
o
no
ft
h
eg
o
l
d
e
ns
e
c
t
i
o
nmethodt
ot
h
er
o
o
tf
i
n
d
i
n
gp
r
o
b
l
e
mi
si
m
p
o
r
t
a
n
t
t
oaccommodatec
q
u
a
l
i
t
ya
n
di
n
e
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
t
s
.C
o
n
s
i
d
e
rExample5
.
4
:

Is
e
dt
oi
m
p
l
e
m
e
n
te
x
p
r
e
s
s
i
o
n(
5
.
2
5
)
.I
tc
a
l
l
st
h
e
Example5_3.m: T
h
i
sm
f
i
l
ei
sl

K
u
t
t
a4
/
5o
r
d
e
rm
e
t
h
o
d
.The
MATLABb
u
i
l
t
i
nf
u
n
c
t
i
o
node451
0u
s
et
h
eRunge
ta
1s
o
f
u
n
c
t
i
o
node45r
e
q
u
i
r
e
st
h
es
t
a
t
ee
q
u
a
t
i
o
n
st
ob
ea
v
a
i
l
a
b
l
ei
na
n
o
t
h
e
rm
f
i
l
e
.I
u
i
r
e
sed
e
f
i
n
i
t
i
o
no
ft
h
ei
n
t
e
r
v
a
lo
fi
n
t
e
g
r
a
t
i
o
na
sw
e
l
la
st
h
ei
n
i
t
i
a
lc
o
n
d
i
t
i
o
n
.Th
e
E
d
c
s
c
r
i
b
e
st
h
eu
s
a
g
e(
ec
o
d
ei
swel
Icommentedt
o
o
)
f
o
l
l
o
w
i
n
gl
i
n
efromt
h
ecod

[t y ]

g
(
)
:
0
.
7
5

2_1
.5
1=0

0$$4

Ex
5_3_state.m: Thes
t
a
t
ee
q
u
a
t
i
o
n
si
n(
5
.
2
6
)a
r
ea
v
a
i
l
a
b
l
ei
ni
sf
i
l
e
.
y(3)

(
5
.
2
)

tI
ni
sformt
h
e

n
e
c
e
s
s
a
r
yt
omakei
ti
n
t
oa
na
c
t
i
v
ec
o
n
s
t
r
a
i
n
l
e
q
u
a
l
i
t
yc
o
n
s
t
m
i
n.
c
a
nb
ee
s
t
a
b
l
i
s
h
e
d
.Th
es
i
m
p
l
e
s
twayt
oa
c
c
o
m
p
l
i
s
ht
h
i
si
st
oc
o
n
v
e
r
tt
h
ep
r
o
b
l
e
mt
o
am
i
n
i
m
i
z
a
t
i
o
np
r
o
b
l
e
mf
o
rwhi
en
u
m
e
r
i
c
a
lt
e
c
h
n
i
q
u
e
sh
a
v
ea
l
r
e
a
d
yb
c
e
n
u
a
t
i
o
n(
5
.
2
7
)e
n
s
u
r
c
st
h
a
tt
h
eminimumv
a
l
u
ewouldb
ez
e
r
o
c
s
t
a
b
l
i
s
h
e
d
.S
q
u
a
r
i
n
gEq
(
i
fi
te
x
i
s
t
s
)
.T
h
e
r
e
f
o
r
et
h
es
o
l
u
t
i
o
n1
0E
q
u
a
t
i
o
n(
5
.
2
7
)i
st
h
esamea
s

ode45('Ex5 3 state'
tintval
bcinit);

[ y(
2
)

(
5
.
2
7
)

i
si
sei
n
e
q
u
a
l
i
t
yc
o
n
s
i
n
to
fE
x
a
r
n
p
l
e5
.
l
ae
x
c
e
p
t ti
ti
st
r
a
n
s
f
o
r
m
e
dt
oa
n

ti
st
h
ei
n
d
e
p
e
n
d
e
n
lv
a
r
i
a
b
l
ea
n
dyt
h
ed
e
p
e
n
d
e
n
tv
e
c
t
o
r
.

stst

223

2- 1
Mm
i
z
e
:
)
:[
0
.
7
5
a
.
5-1]2

0.5*y(1)*y(3)]';

(
5
.
2
8
)

a
n
d
l
e
dbyGoldSection_1Var.m.Wh
i
l
ei
swasq
u
i
c
kandp
a
i
n
l
e
s
s
whichc
a
nbeh
t
h
e
r
e
a
r
s
omea
t
t
e
n
d
a
n
tp
r
o
b
l
e
m
su
s
i
n
gt
h
i
sa
p
p
r
o
a
c
h
. t
h
en
o
n
l
i
n
r
i
t
yh
a
s
i
n
c
r
e
a
s
e
d
.Thi
si
su
s
u
a
l
l
yn
o
te
n
c
o
u
r
a
g
e
di
nn
u
m
e
r
i
c
a
li
n
v
c
s
t
i
g
a
t
i
o
n
s
.S
e
c
o
n
dt
h
e
numbero
fs
o
l
u
t
i
o
n
sh
a
ss
a
t
i
s
f
y
i
n
gFOCi
n
c
r
e
a
s
e
d
.F
i
g
u
r
e5
.
5r
e
p
s
e
n
t
st
h
ef
u
n
c
t
i
o
n
s
i
nE
q
u
a
t
i
o
n
s(
5
.
2
7
)a
n
d(
5
.
2
8
)
.

R
e
f
e
rt
ol
h
ec
o
d
ef
o
rd
e
t
a
i
l
s
.GoldSectlon_1Var.mi
su
s
e
df
r
o
mt
h
ecommandl
i
n
e
f
o
ri
sp
r
o
b
l
e
ma
s

U g
e
: GoldSection_1Var('Example5_3'
0.001
0
0.1
10)
S
t
a
r
ti
t
e
r
a
t
i
o
n
:
start
alphal(low)

1.0000

alpha()
0.1528
0.1904

alpha(2) alpha(up)
0.2472
0.4000
0.0378
0.0161

alpha(l)
0.3364
0.0000

alpha(2) alpha(up)
0.3367
0.3371
0.0000
0.0000

'

Thef
i
n
a
li
t
e
r
a
t
i
o
n
:
iteration 12
alphal(low)
0.3359
0.0000

"
4

S
o
l
u
t
l
o
nUslngGoldSection_1Var.m:Thef
o
l
l
o
w
i
n
gd
i
s
c
u
s
s
i
o
ni
sb
a
s
e
do
n
t
h
ee
x
p
e
r
i
e
n
c
eo
fa
t
t
e
m
p
t
i
n
gt
or
u
nt
h
cg
o
l
d
e
ns
e
c
t
i
o
nmethodw
i
t
ht
h
el
o
w
e
rbound
a
tt
h
ev
a
l
u
eo
f0ec
a
s
ew
i
t
heo
t
h
c
re
x
a
m
p
l
e
s
.The
i
sap
r
o
b
l
e
ms
i
n
c
et
h
e
s
o
l
u
t
i
o
nc
a
n
n
o
tb
ef
o
u
n
d
.Th
ep
r
o
b
l
e
ma
p
p
r
st
obei
nt
h
ec
a
1
c
u
l
a
t
i
o
no
ft
h
eu
p
p
e
r
b
o
u
n
d
.l
nmanyo
ft
h
en
u
m
e
r
i
c
a
lt
e
c
h
n
i
q
u
e
st
h
e
r
ei
sa
na
s
s
u
m
p
t
i
o
na
tt
h
eo
b
j
e
c
t
i
v
e
f
u
n
c
t
i
o
nb
e
i
n
gd
e
a
1tw
i
t
hi
smodala
ti
s
h
a
sas
i
n
g
l
eminimumi
nt
h
er
e
g
i
o
no
f
i
a
t
e
r
e
s
t
.F
ort
h
i
st
oh
a
p
p
e
n
.i
ti
sassumeda
tef
u
n
c
t
i
o
nb
e
i
n
gm
i
n
i
m
i
z
e
dwi
1
l
a
l
w
a
y
sd
e
c
r
e
a
s
ef
r
o
mt
h
el
o
w
e
rb
o
u
n
d
.Takeal
o
o
ka
tef
u
n
c
t
i
o
n
f
i
nF
i
g
u
r
e5
.
5
.I
t
n
c
r
e
a
s
i
n
ga
tt
h
cva
1ueofz
e
r
o
.UpperBound_1Var.ma
s
s
u
m
e
st
h
a
tei
n
i
t
ia
1
i
si
i
n
t
e
r
va
1i
st
o
ol
a
r
g
eandc
o
n
t
i
n
u
e
st
ob
i
s
e
c
ti
tt
of
i
n
dan
c
t
i
o
nv
a
1u
e i
sl
e
s
sa
n
r
i
v
e
st
h
ei
n
t
e
r
v
a
It
oz
e
r
oandt
h
ef
u
n
c
t
i
o
ni
se
x
i
t
e
d
t
h
eonea
tt
h
el
o
w
c
rb
o
u
n
d
.Thisd
w
i
t
ht
h
el
o
w
e
rbounda
n
dw
i
t
hnos
o
l
u
t
i
o
n
.
Remedy1
: Thee
a
s
i
e
s
twayo
u
to
ft
h
ed
i
f
f
i
c
u
l
t
yi
st
oc
h
a
n
g
ct
h
ev
a
l
u
eo
ft
h
el
o
w
e
r
bounda
n
dt
r
ya
g
a
i
n
.I
ti
su
s
e
f
u
lt
ou
n
d
c
r
s
t
a
n
dt
h
a
t p
r
o
b
l
e
mi
sw
i
t
ht
h
ec
a
l
c
u
l
a
t
i
o
n

P
R
O
B
L
E
M
S 225

224 N
U
M
E
R
I
C
A
LT
E
C
H
N
I
Q
U
E
5T
H
EO
N
E
D
I
M
E
N
S
I
O
N
A
LP
R
O
B
L
E
M

1
6
1
4

start
alphal(low)

1.0000
iteration 16
alphal(low)
2.5266
0.0000

a
v

J-Mhe-e
s'p-

1
2

U g
e
:GoldSection_1Var('Example5_4'0.0010.
120)

E
e
a

'
a
a

a
'
'
o
-

1
8

1
0

t
'--e

(HM)

3
.V

2
.
5

3
.
5

alpha(up)
2.6000
0.0289

alpha(l)
2.5270
0.0000

alpha(2)
2.5273
0.0000

alpha(up)
2.5278
0.0000

1
.B
u
r
d
e
n
R
.L
.
a
n
dF
a
i
r
e
s
J
.D
.
N
u
m
e
r
i
c
a
lA
n
a
l
y
s
i
s
.4
t
he
d
.PWS_KENTPu
b
l
i
s
h
i
n
g
o
s
t
o
n1
9
8
9
.
CompanyB
2
.H
o
s
l
e
l
l
e
rGH..samina-M.s..andM
o
n
t
a
I
O
U
p
.
D
.A
n
a
l
y
t
i
c
a
l
N
w
n
e
r
i
c
d
d

p
r
m
t
i
c
c
H
a
l
l
E
n
s
l
e
w
o
o
dC
l
i
f
f
s
.
c
o
m
p
u
t
a
t
i
o
n
a
lM
e
t
h
o
d
s
h
r
s
c
i
e
n
c
eandE
n
g
i
n
e
e
r
i
n
g
N
J
.
1
9
91
.

0
.
5

alpha(2)
1.6069
2.1720

REFERENCES

s
o
l
u
t
l
o
n

alpha(l)
0.9931
3.0624

1
G
.N
.
.Nu
O
p
l
i
l

1
1
0
1
1T
e
c
l
m
i
q

f
o
rE
n
g
i
n
e
e
r
i
n
gD
e
s
i
g
l
l

M
c
G
r
a
w
H
i
l
l
NewY
o
r
k
.1
9
8
4
.
4
.H
u
n
t
l
c
y
.H
.E
.
T
h
eD
l
v
I
l
l
eP
r
o
p
o
r
l
i
o
n
:A S
l
u
d
yi
nM
u
t
h
e
m
a
l
i
:
r
a
lB
e
a
u
t
y
.D
o
v
e
r
Pu
b
l
i
c
a
u
o
n
s
NewY
o
r
k
.1
9
7
0
.
5
.S
c
h
l
i
c
h
u
n
g
.H
.
BoundaLa
y
e
rThe
o
r
y

McGrawH
i
U
.NcwYo
1
9
7
9
.

F
l
g
u
r
e5
.
5E
x
a
m
p
l
5
.
4
.
PROBLEMS
o
ft
h
eu
p
p
e
rbound(
b
r
a
c
k
e
t
i
n
gt
h
eminimum)r
a
e
rt
h
a
nw
i
t
h g
o
l
d
e
ns
e
c
t
i
o
n
g
o
l
d
e
ns
e
c
t
i
o
nme
o
d
m
e
t
h
o
d
.C
h
a
n
g
i
n
gt
h
el
o
w
e
rboundt
o1andr
e
r
u
n
n
i
n
gt
h
p
r
o
v
i
d
e
sas
u
c
c
e
s
s
f
u
l s
u
l
twhichm
a
t
c
h
e
seg
r
a
p
h
i
c
a
ls
o
l
u
t
i
o
n
.Whiles
u
c
c
e
s
swas
r
e
a
l
i
z
e
d
.s
u
c
hap
r
e
d
u
r
en
e
e
d
su
s
e
ri
n
t
c
r
v
e
n
t
i
o
nwella
ssomeknowledgea
b
o
u
t
p
roblcm-int
h
i
sc
a
s
et
h
e
r
ewaseg
r
a
p
h
i
c
a
ls
o
l
u
t
i
o
n
_Thesea
r
en
o
ta
c
c
e
p
t
a
b
l
e
t
h
l
u
r
e
.
q
u
a
l
i
t
i

f
o
ra
na
u
t
o
m
a
t
i
cp
r
o
c

5
.
1 E
x
t
e
n
dExample5.
1t
oi
n
c
l
u
d
ctwoi
n
e
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
t
sf
o
rw
h
i
c
has
o
l
u
t
i
o
n

Remedy2
:S
i
o
c
et
h
ep
o
s
i
t
i
v
es
l
o
p
ea
tt
h
el
o
w
e
rbouodi
si
n
d
i
c
a
t
i
v
eo
ft
h
e
o
s
i
t
i
v
es
i
g
n01t
h
es
l
o
p
ea
ndmovet
h
e
p
r
o
b
l
e
m
.i
ts
h
o
u
l
db
ep
o
s
s
i
b
l
et
os
e
n
s
et
h
i
sp
sn
e
g
a
t
i
v
e
. Such ap
o
i
n
tc
a
nbe
l
o
w
e
rboundt
ot
h
ep
o
i
n
twheret
h
eslope i
d
i
s
c
o
v
e
r
e
dbyt
h
esam
s
c
a
n
n
i
n
gp
r
o
c
e
d
u
r
e
.Thcs
l
o
p
ec
o
m
p
u
t
a
t
i
o
ni
sn
u
m
e
r
i
c
a
l
andi
sb
a
s
e
donf
o
r
w
a
r
df
i
r
s
td
i
f
f
e
r
e
n
c
e
.Onccmorci
schangei
nt
h
el
o
w
e
rbound
i
so
n
l
yn
e
e
d
e
df
o
rt
h
eu
p
p
c
rboundc
a
l
c
u
l
a
t
i
o
n
.UpperBound_m1Var.mi
st
h
e
ot
h
ep
o
i
n
t
m
o
d
i
f
i
e
du
p
p
e
rboundc
a
l
c
u
l
a
t
i
o
nwhichr
e
l
o
c
a
t
e
st
h
elow~r boundt
where t
h
es
l
o
p
ei
sn
e
g
a
t
i
v
e
. The u
s
a
g
e and i
n
f
o
r
m
a
t
i
o
n from r
u
n
n
i
n
gt
h
e
GoldSectlon_1Var.mi
n
c
o
r
p
o
r
a
t
i
n
gt
h
echaogedu
p
p
e
rboundc
a
l
c
u
l
a
t
i
o
oa
r
e
shownb
e
l
o
w
.Thesamcu
s
a
g
ew
i
t
ht
h
co
l
dUpperBound_1Var.md
o
e
so
o
t
p
r
o
d
u
c
ea
n
yu
s
e
f
u
lr
c
s
u
l
t
s
.

5
4

e
x
i
s
t
s
D
i
s
p
l
a
y f

ul(
5
.
2 E
x
t
e
n
dExample
5

.
1t
oi
nc
Iu
detwo
i
n u
a

l
i
t
yc
o
n
s
t
r
a
i
n
t
sf
o
rw
h
i
c
h
t
h
e
r
l
1

sn
o
s
o
l
u
t
i
o
n
.D
i
s
p
l
a
yt
h
ef
u
n
c
t
i
o
ng
r
a
p
h
i
c
a
1
1
y
.
5.
3 Reprogramec
o
d
eSec5_2_2.mobtainev
a
l
u
e
so
f

f
andc
t
't
h
r
o
u
g
h
M
A
'
1.ABf
u
n
c
t
i
o
np
r
o
g
r
a
m
s da
p
p
l
yt
oExample5
.1
.

2
2
J
;
;

1
3
2
2
2
:
l
:
:
:
Z
3
4
7
t

5
.
5 I
d
e
n
t
i
f
yt
h
ev
a
l
u
eo
fwhereec
o
n
s
t
r
n
t(
5.
3
)bcomesa
c
t
i
v
eu
s
i
n
g
Ncwton-Raphsoot
e
c
h
n
i
q
u
e
.
u
n
c
t
i
o
no
fo
n
c
5
.
6 I
m
p
l
e
m
c
n
tab
i
s
e
c
t
i
o
np
r
o
c
e
d
u
r
ef
o
rt
h
er
n
i
n
i
m
i
z
a
t
i
o
no
fef
v
a
r
i
a
b
l
e
.Applyi
tt
os
o
l
v
eExample5
.1
.
5
.
7 F
o
rt
h
ep
o
l
y
n
o
r
n
i
a
1a
p
p
r
Ol
d
m
a
t
i
o
ni
m
p
l
e
m
e
n
t
a
t
i
o
n
c
x
t
e
n
dcc
o
d
et
od
i
s
p
l
a
y
eo
r
i
g
i
n
a
lf
u
n
c
o
n dea
p
p
r
o
x
i
m
a
t
i
n
gp
o
l
y
n
o
m
i
a
l
.
5
.
8 F
o
rcg
o
l
d
e
ns
e
c
t
i
o
ni
m
p
l
c
m
e
n
t
a
t
i
o
n

e
x
t
e
n
dt
h
ec
o
d
ct
od
i
s
p
l
a
yt
h
eo
r
i
g
i
n
a
l
f
u
n
c
t
i
o
nandt
h
ca
n
a
le
x
t
e
r
n
a
lb
o
u
n
d
sont
h
ev
a
r
i
a
b
l
e
.

226

NUMERICALTECHNIQUE5THEONE.DIMENSIONALPROBLEM

5
.
9 Verira
teb
o
u
n
d
sont
h
ev
a
r
i
a
b
l
ea
r
ed
e
c
r
e
a
s
i
n
gbyesamer
a
t
i
o
.Wh
yi
s
t
h
i
sb
e
t
t
e
ra
nt
h
eb
i
s
e
c
t
i
o
nmethod?
r
m
i
n
et
h
ez
e
r
oof
5
.
1
0 Howwouldyous
e
tupt
h
eg
o
l
d
e
ns
e
c
t
i
o
nmethodt
od
e
t
f
u
n
c
t
i
o
ni
n
s
t
e
a
do
ft
h
eminimum.

o
l
d
e
ns
e
c
t
i
o
n dep
o
l
y
n
o
m
i
a
la
p
p
r
o
x
i
m
a
t
i
o
nt
of
i
n
dt
h
e
5
.
1
1 Combineeg
u
n
c
t
i
o
n
.
mmimumofef
p
r
o
v
t
h
ea
c
c
u
r
a
c
yo
ft
h
es
o
l
u
t
i
o
nt
ot
h
eB
I i
u
s 1o
v
e
rt
h
a
tp
r
e
s
e
n
t
e
d
5
.
1
2 Im
i
nt
h
et
e
x
.
t
5
.
1
3O
b
t
a
i
n t
h
es
o
l
u
t
i
o
nt
ot
h
eB
l
a
s
s
i
u
s q
u
a
t
l
On u
smg t
h
ep
o
l
y
n
o
m
i
a
l
a
p
p
r
o
x
i
m
a
t
i
o
nm
e
t
h
o
d
.

NUMERICALTECHNIQUES
FORUNCONSTRAINED
OPTIMIZATION

T
h
i
sc
h
a
p
t
e
ri
I
lu
s
a
t
e
smanyn
u
m
e
r
i
ca
1t
e
c
h
n
i
q
u
e
sf
o
rm
u
J
t
i
v
a
r
i
a
b
l
eu
n
c
o
n
s
t
r
a
i
n
e
d
a
t
i
o
ni
sn
o
tacommono
c
c
u
r
r
e
n
c
i
n
o
p
t
i
m
i
z
a
t
i
o
n
.Whileu
n
c
o
n
s
t
r
a
i
n
e
dop
n
e
v
e
r
t
h
e
l
e
s
st
h
en
u
m
e
r
i
c
a
1t
e
c
h
n
i
q
u
e
si
n
c
I
udedh
e
r
ed
e
m
o
n
s
t
r
a
t
e
e
n
g
i
n
e
e
r
i
n
gd
e
s
i
g
n
i
n
t
e
r
e
s
i
d
e
a
sanda
l
s
oc
a
p
t
u
r
esomeo
ft
h
ee
a
r
l
yi
n
t
e
n
s
i
v
eworki
nea
r
e
ao
fd
e
s
i
g
n
i
o
n[
1
]
.Th
e
ya
I
s
op
r
o
v
i
d
et
h
emeanst
os
o
l
v
ec
o
n
s
t
r
a
i
n
e
dp
r
o
b
l
e
m
sa
f
t
e
rt
h
e
y
o
p
t
i
m
i t
h
a
v
eb
e
e
nt
r
a
n
s
f
o
r
m
e
di
n
t
oa
nu
n
c
o
n
s
t
r
a
i
n
e
done(
i
n
d
i
r
e
c
tmo
d
s
C
h
a
p
t
e
r
7
)
.
T
h
i
sbooki
su
n
i
q
u
ef
o
rt
h
ep
r
e
s
e
n
t
a
t
i
o
no
fn
o
n
g
r
a
d
i
e
n
tt
e
c
h
n
i
q
u
e
s(
S
e
c
t
i
o
n6
.
2
)
which c
a
nb
eempowered by t
h
eu
b
i
q
u
i
t
o
u
sa
v
a
i
l
a
b
i
l
i
t
yo
fi
n
c
r
d
i
b
l
ed
e
s
k
t
o
p
c
o
m
p
u
t
i
n
gp
o
w
e
r
.U
n
t
i
lv
e
r
yr
e
c
e
n
t
l
y
.a
p
p
l
i
c
a
t
i
o
no
fd
e
s
i
g
no
p
t
i
m
i
z
a
t
i
o
nr
e
q
u
i
r
e
d
m
a
i
n
f
r
a
m
ec
o
m
p
u
t
e
r
st
h
a
tr
e
l
i
e
donl
a
r
g
ep
r
o
g
r
a
m
sw
r
i
t
t
e
ni
nFORTRAN.T
o
d
a
y
'
s
t
r
e
q
r
eo
n
l
yI
im
i
t
e
dprogrammings
k
i
l
l
s
.Thed
e
s
i
r
ef
o
r
d
e
s
k
t
o
pc
a
nr
u
np
r
o
g
r
a
m
s a
a
Io
p
t
i
m
i
z
a
t
i
o
nh
a
sb
r
o
u
g
h
ti
n
t
of
o
c
u
sn
u
m
e
r
i
ca
It
e
c
h
n
i
q
u
e
sa
ta
r
el
a
r
g
e
l
y
g
l
o
b
h
a
v
el
i
m
i
t
e
dn
e
e
df
o
rs
o
p
h
i
s
t
i
c
a
t
e
dg
r
a
d
i
e
n
t
b
a
s
e
dm
e
t
h
o
d
s
r
e
q
u
i
r
el
i
m
i
t
e
d
h
e
u
r
i
s
t
i
c
programming s
k
i
H
s
. and r
e
q
u
i
r
el
i
m
i
t
e
d programming r
e
s
o
u
r
c
e
s
.U
s
u
a
l
l
ys
u
c
h
h
es
o
l
u
t
i
o
ns
p
a
c
eb
es
c
a
n
n
e
du
s
i
n
gal
a
r
g
enumbero
f
t
e
c
h
n
i
q
u
e
sr
e
q
u
i
r
e t
Id
e
s
k
t
o
p
sf
o
ru
n
l
i
m
i
t
e
dt
i
m
e
.
i
t
e
r
a
t
i
o
n
s
.Th
e
s
em
e
t
h
o
d
sc
a
ni
d
e
a
l
l
yb
er
u
nonp
e
r
s
o
na
Th
ep
r
e
s
s
o
fl
i
m
i
t
e
dc
o
m
p
u
t
i
n
gr
e
s
o
u
r
c
e
si
snol
o
n
g
e
ras
i
g
n
i
f
i
c
a
n
tc
o
n
s
t
r
a
i
n
ton
i
ti
s
t
h
e
s
et
e
c
h
n
i
q
u
e
s
.Whilet
h
e
ymaycompromiseone
l
e
g
a
n
c
eands
o
p
h
i
s
t
i
c
a
t
i
o
n
i
m
p
o
r
t
a
n
tt
or
e
c
o
g
n
i
z
e o
p
p
o
r
t
u
n
i
t
yt
oe
n
g
a
g
ed
e
s
k
t
o
pr
e
s
o
u
r
c
e
sf
o
rg
r
e
a
t
e
r
u
s
e
f
u
l
n
e
s
s
.

6
.
1 PROBLEMDEF1NITION
Theu
n
c
o
n
s
a
i
n
e
do
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
e
mr
e
q
u
i
r
e
so
n
l
yt
h
eo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
n
.The
bookh
a
sc
h
o
s
e
nt
oe
m
p
h
a
s
i
z
ea
na
c
c
o
m
p
a
n
y
i
n
gs
e
to
fs
i
d
ec
o
n
s
t
r
a
i
n
t
st
or
e
s
t
r
i
c
t
227

228

NUMERIC
AlTECHNIQUESFOAUNCONSTRAINEDOPTIMIZATION

6
.
1 PAOBLEMOENI
T
lON 2
2
9

s
o
l
u
t
i
o
nt
o a
c
c
e
p
t
a
b
l
e i
a
c
e
l
g
i
o
n
.Mostn
u
m
e
r
i
c
a
l i
q
u
i
n S
c
h
a
p
t
E
ra
g
p
o
e
?ns intsinI ngout method.I
t
i
su
s
u

d
I

y
e
P
o
n
s
i

bJ
i

t
y
0f
t

h
1

e
d
e
t

ov
e

r
t
h
e

d
cc

n
lIIn
t
sa
sp
a
r
to
fh
i
se
x
p
l
o
r
a
t
i
o
n
o
f o
p
t
i
m
u
m
. i
sc
h
e
c
k
i
n
go
fc
s
i
d
en
s
n
tc
a
nbee
a
s
i
l
yp
r
o
g
r
a
m
m
e
di
nt
h
e
c
o
d
e
.Thu
st
h
ep
r
o
b
l
e
mf
o
ri
sc
h
a
p
t
e
rc
a
nb
ed
e
f
i
n
e
da
s

(
6
.
1
)

44whHMl'I

i
=1

2

n

(
6
.
2
)

t
a
-

x
f$;XI$;x
r
;

'lli
JTIl!ll

S
u
b
j
e
c
t
t
o
:

]"

f(X);

'j

M
i
n
i
m
i
z
e

6
.
1
.
1 Examp'e6
.
1
r
p
o
s
s
i
b
l
e h
i
ca
Jd
e
s
c
r
i
p
t
i
o
r
;;2312::zz;222Z2;2JZem

M
i
n
i
m
i
z
e f(X)=f(xl X2)=3+(XI- 1
2+(X2_ 2
.5
x
2
l
)
2

(
6.
3
)

S
u
b
j
e
c
tt
o
: OSX(S5; OSx2S5

(
6
.
4
)

F
i
g
u
r
e6
.
1p
r
o
v
i
d
e
st
h
ec
o
n
t
o
u
r
so
fep
r
o
b
l
e
ma
n
d

f
J

l
u
t
i
o
nc
a
nb
el
o
c
a
t
e
da
t

e
)

adratic

iZL2;arvalueisf=3 n

h
es
o
l
u
t
i
o
n

0
.
5

1
.
5

2
.
5

3
.
5

4
.
5

X
1

F
l
g
u
r
o6
.
1E
x
a
m
p
l
e6
.
1
.

6
.
1
.
2 NecessaryandS
u
f
f
l
c
l
e
n
tCondltlons

2
;
t
r
z
z
;
:
:
2
j
r
2
z

;
t
t
:
!
s
Z
;
:
?

c
o
n
d
i
t
i
o
n
s
)a
r
eeg
r
a
d
l
e
n
t
sm
u
s
tv
a
n
i
s
ha
tt
h
es
o
l
u
t
i
o
n

21=2(Xt-15
)=0

1)+2
(
X
2-2
)=
0

TheH
e
s
s
i
a
ni
sp
o
s
i
t
i
v
ed
e
f
i
n
i
t
ec
o
n
s
i
d
e
r
i
n
gt
h
ed
e
t
e
n
n
a
n
t
sa
l
o
n
e
.Sec6_1_1.m'
p
r
o
v
i
d
e
sEd
ef
o
rt
h
es
y
m
b
o
l
i
cc
a
l
c
u
l
a
t
i
o
na
n
da
l
s
od
i
s
p
l
a
y
s gp
h
i
l
d
c
r
i
p
t
i
o
n
thep
r
o
b
l
e
m
.Ther
e
s
u
l
t
so
fec
a
l
c
u
l
a
t
i
o
na
r
cd
i
s
p
l
a
y
e
di
nt
h
ecommandw
i
n
d
o
w
.l
e
s
a
t
i
s
f
a
c
t
i
o
no
ft
h
eSOCi
d
e
n
t
i
f
i
e
s s
o
l
u
t
i
o
nf
o
rt
h
ed
e
s
i
g
ni
sam
i
n
i
m
u
m
.
6
.
1
.
3 Elementso
faNumerlcalTechnlque

(
6
.
5
a
)
(
6
.
5
b
)

Equ
ons(6.5)c bevedt
oo
b
t
a
i
nx
j=3x
i=2
.Thi
sm
a
t
c
h
e
seg
r
a
p
h

2
i
l
j
i
t
M
2
2
3
i
C
:
t

;sz

'
h
ee
l
e
m
e
n
t
so
fat
y
p
i
1
num
1
t
e
c
h
n
i
q
u
co
ra
l
g
o
r
i
t
h
m

c
a
nbca
s
s
o
c
i
a
t
e
dw
i
t
ht
h
e
g
e
n
e
r
i
ca
o
r
i
ni
n
d
u
c
e
dd
u
r
i
n
g e d
i
s
c
u
s
s
i
o
no
ft
h
er
e
l
e
v
a
n
c
eo
ft
h
e
o
n
e
d
i
m
e
n
s
i
o
n
a
1op
n
i
z
a
t
i
o
nf
o
rm
u
l
t
i
v
a
r
i
a
b
l
ep
r
o
b
l
e
m
s(
C
h
a
p
t
5
)
.usa
l
g
o
r
i
t
h
mi
s
i
t
e
r
a
t
i
v
ea
n
di
sa
l
s
or
e
f
e
r
r
e
dt
oa
sas
e
a
r
c
ha
l
g
o
r
i
t
l
v

na
st
h
ei
t
e
r
a
t
i
o
n
st
a
k
ep
l
a
c
eb
ym
o
v
i
n
g
a
l
o
n
gas
e
a
r
c
hd
i
r
e
c
t
i
o
no
ras
e
a
r
c
hv
e
c
t
o
r
d
u
r
i
n
ga
ni
t
e
r
a
t
i
o
n
.Th
es
h
d
i
r
e
c
t
i
o
n
sc
a
n
bedetem
.
e
di
nmanyw
a
y
s
.Th
ed
i
f
f
e
r
e
n
t h
n
i
q
u
e
sa
ta
r
ep
s
e
n
t
e
di
nt
h
i
sc
h
a
p
t
e
r
p
r
i
m
a
r
i
l
yd
i
f
f
e
r howt
h
cs
e
a
r
c
hd
i
r
t
i
o
ni
s

t
a
b
l
i
s
h
e
d
. l
e m
a
i
n
i
n
ge
l
e
m
e
n
t
so
ft
h
e
a
l
g
o
r
i
m
e
x
p
tf
o
r c
o
n
v
e
r
g
e
n
c
c
J
s
t
o
p
p
i
n
gc
r
i
t
e
r
i
a

a
r
et
h
es
a
m
ef
o
ra
l
m
o
s
ta
l
lo
f
l s

t
h
em
e
t
h
o
d
s
.Thea
Jg
o
r
i
t
h
mc
a
nb
ee
x
p
r
c
da
sf
o
l
l
o
w
s
.

(
6
.
6
)

I
F
i
l
e
s
l
Obed
o
w
n
l
o
n
d
e
dr
r
o
mI
h
cw
e
bs
i n
r
ei
n
d
i
c
a
l
e
db
yb
o
l
d
r
a
c
es
o
n
ss
e
r
i
rI
y
p
e
.

Hil

6
2

N
U
M
E
R
I
C
A
I
.T
E
C
H
N
I
O
U
E
-N
O
N
G
A
A
O
I
E
N
TMETHOOS 2
3
1

230 N
U
M
E
R
I
C
AlT
E
C
H
N
I
O
U
E
SFO
U
N
C
O
N
S
T
R
A
I
N
E
OO
P
T
I
M
I
Z
A
T
I
O
N

Ao
r
l
t
h
m
:RandomWalk(
A
6
.
2
)
S
t
e
p1
.ChooseXoN(numbero
fi
t
e
r
a
t
l
o
n
s
)

GeneralAlgorithma

S
t
e
p1
. Ch
seXo
S
t
c
p2
.F
o
re
a
c
hi
t
e
r
a
t
i
o
ni

=1
S
e
ti
S
t
e
p2
.F
o
re
a
c
hi
t
e
r
a
t
i
o
nI
e
c
t
o
r
81 Randomv

D
e
t
e
r
m
i
n
es
e
a
r
c
hd
i
r
e
c
t
i
o
nv
e
c
t
o
r8j
(
l
tw
ouldb
en
i
c
ei
ft
h
eo
b
j
e
c
t
i
v
ed
e
c
r
e
a
s
e
da
l
o
n
gt
h
i
sd
i
r
e
c
t
i
o
n
)

S
t
e
p3
.Xj+
=Xj
+
S
j
1
'
ji
sd
e
t
e
r
m
i
n
e
dbym
i
n
i
m
i
z
i
n
gj
{
Xi+I
)

S
t
c
p3
.C
a
l
c
u
l
a
t
es
X
=si
N
o
t
e
: aX
snowaf
u
n
c
t
i
o
no
ft
h
es
c
a
Ia
r<
<
ja
sSi
sknownfromS
t
e
p2
ji
i
sc
a
l
1
e
des
t
e
ps
i
z
ea
si
t

t
a
b
l
i
s
h
e
sel
e
n
g
t
hofAX
j

+1
i
i
l
fi
<N00ToS
t
e
p2
e
l
s
cSP(
i

r
a
t
i
o
n
se
x
c
e
e
d
e
d
)

ji
sdt
e
r
m
i
n
e
dbyM
i
n
i
m
i
z
i
n
g
f
( whcreX
'
+
I Xj+X1
h
ec
o
n
v
e
r
g
c
n
c
ec
r
i
t
e
r
i
a(FOC/SOC)
andt
h
es
t
o
p
p
i
n
gc
r
i
t
c
r
i
a(
Is
H
e
r
et
d
e
s
i
g
nc
h
a
n
g
i
n
g
?E
x
c
e
e
d
i
n
gi
t
e
r
a
t
i
o
nc
o
u
n
t
?e
t
c
.
)mustbec
h
e
c
k
e
d
.

E
Z
2
2
2
2
z
t
e
s
z
z
:
;
1
2
r
z
:
:
;
;i

ii
+1
;

OoToS
t
e
p2

l
nt
h
i
sc
h
a
p
t
e
r
t
h
i
sa
l
g
o
r
i
t
h
m
i
cs
t
r
u
c
t
u
r
ew
i
l
lb
eu
s
c
df
o
ra
Ilo
ft
h
em
e
t
h
o
d
s
.
'
h
e
l
1beimplementedu
s
i
n
gt
h
eg
o
l
d
e
ns
e
c
t
i
o
nm
e
t
b
o
d
.
o
n
e
d
i
m
e
n
s
i
o
n
a
lo
p
t
i
m
i
z
a
t
i
o
nwi
Example6
.
1wi
l
beu
s
e
dt
od
c
v
e
l
o
pt
h
cd
i
f
f
e
r
c
n
tn
u
m
e
r
i
c
a
lm
e
t
h
o
d
s
.

l
;
f

:
;

s
i
g
nb
a
s
e
donat
e
s
to
frandomn
u
m
b
e
r
.

RandomWalk
m:Tlmrandomwalka
l
g
o
r
i
t
h
mtA62)is i
m
p
i
e
m
e
n
t
c
di
nt
h
i
s

s-N
ONGRAOIENTMETHOOS
6
.
2 NUMERICALTECHNIQUE
'
h
e
s
emethodsa
r
ea
Js
oc
a
Jl
e
dz
e
r
o
o
r
d
e
rm
e
t
h
o
d
s
.
'
h
eo
r
d
e
rr
e
f
c
r
st
oeo
r
d
c
ro
ft
h
c
e
c
d
e
dt
oe
s
t
a
b
l
i
s
ht
h
es
e
a
r
c
hd
i
r
e
c
t
i
o
nd
u
r
i
n
g
d
e
r
i
v
a
t
i
v
eo
ft
h
eo
b
j
c
c
t
i
v
ef
u
n
c
t
i
o
nn
a
n
yi
t
e
r
a
t
i
o
n
.Ze
r
oo
r
d
e
rs
i
g
n
i
f
i
e
st
h
a
tnod
e
r
i
v
a
t
i
v
e
s u
s
e
d[
2
]whichi
ntum
n
l
yf
u
n
c
t
i
o
nv
a
l
u
e
sa
r
eu
s
e
dt
oS
b
l
i
s
ht
b
es
e
a
r
c
hv
e
c
t
o
r
.T
h
e
r
ei
s
i
m
p
l
i
e
s o
a
n
o
t
h
e
r i
m
p
o
r
t
a
n
t s
i
g
n
i
f
i
c
a
n
c
e t
h
a
t a
c
c
o
m
p
a
n
i
e
s t
b
e m
i
s
s
i
n
g d
e
r
i
v
a
t
i
v
e
computation-the FOCs
h
o
u
l
dn
o
tbea
p
p
l
i
e
di
nt
h
e
s
em
e
t
h
o
d
s
.I
fwc were
udet
h
i
si
n
f
o
r
m
a
t
i
o
nf
o
re
s
t
a
b
l
i
s
h
i
n
gt
h
e
c
o
m
p
u
t
i
n
gd
e
r
i
v
a
t
i
v
c
swem
i
g
h
ta
sw
e
l
li
nc
b
j
e
c
t
i
v
cf
u
n
c
t
i
o
no
rt
h
ed
e
s
i
g
nv
a
r
i
a
b
l
e
sc
a
n
s
e
a
r
c
hv
e
c
t
o
r
.Onlyt
h
ec
h
a
n
g
e
si
n eo
o
n
v
e
r
g
e
n
c
ed/o
rs
t
o
p
p
i
n
gc
r
i
t
e
r
i
a
.
p
r
o
v
i
d c
et
c
c
h
n
i
q
u
e
sa
r
ei
n
c
l
u
d
e
di
nt
h
i
ss
e
c
t
i
o
n
.Thef
i
r
s
ti
sah
e
u
r
i
s
t
i
coneb
a
s
e
don
c
c
o
n
donec
y
c
l
e
st
h
r
o
u
g
has
e
a
r
c
hd
i
r
c
c
t
i
o
nb
a
s
e
don
randoms
e
a
r
c
hd
i
r
e
c
t
i
o
n
s
.Thes
el
a
s
ti
sP
o
w
c
l
l
'
smethodw
h
i
c
hh
a
st
b
cp
r
o
p
e
r
t
yo
fq
u
a
d
r
a
t
i
c
e
a
c
hv
a
r
l
a
b
l
c
.l
c
o
n
v
e
r
g
e
n
c
e
.
6
.
2
.
1 RandomWalk

Th
es
e
a
r
c
hd
i
r
e
c
t
i
o
nd
u
r
i
n
ge
a
c
hi
t
e
r
a
t
i
o
ni
sarandomd
i
r
e
c
t
i
o
n
.Randomnumbersand
s
e
t
so
frandomnumbersa
r
eu
s
u
a
l
avai
1a
b
l
et
h
r
o
u
g
hs
o
f
t
w
a
r
e
.MATLABi
n
c
l
u
d
e
st
h
e
a
b
t
yt
og
e
n
e
r
a
t
es
e
v
e
r
a
Jt
y
p
c
so
f random numbers and m
a
t
r
i
c
e
s
. Most
c
o
m
p
u
t
e
r
g
e
n
e
r
a
t
e
drandomnumbersa
r
c
a
l
l
e
dp
seudorandomnumbersa
st
h
e
yc
y
c
l
e
b
e
e
nc
r
e
a
t
e
d
.Theone
d
i
m
e
n
s
i
o
n
a
ls
t
c
p
s
i
z
e
a
f
t
e
ras
u
f
f
j
c
i
e
n
tnumbero
fthemhav
c
t
i
o
n(
C
h
a
p
t
e
r5
)
.
c
o
m
p
u
t
a
t
i
o
ni
sdoncbyt
h
cGoldSectlon_nVar.m n

1
'
I~

OP6miadon-TMcodeoft
h
erandomwalkf
u
n
c
t
i
o
na
l
s
oi
n
c
l
u
d
e
st
h
Ef
o
l
l
o
w
i
n
g
ec
o
n
t
o
u
ro
ft
h
co
b
j
e
c
t
i
v
ei
sd
r
a
w
n
.
o
-v
a
r
i
a
b
l
ep
r
o
b
l
e
m
stn
Fortw
Fortwov
a
r
i
a
b
l
E
p
r
o
b
l
e
m
st
h
es
u
c
c
e
s
s
f
u
li
t
e
r
a
t
i
o
n
sMet
r
a
c
k
e
dont
h
ec
o
n
t
o
u
r
dt
h
ei
t
e
r
a
t
i
o
n
"numberi
d
e
n
t
i
f
i
e
d
.
.2;:u16vm
d1
0

C df
o
r
t
h
eo
n
e
d
i
m
e
n
s
i
o
n
a
ls
t
e
p
s
i
z
ec
o
m
p
u
t
a
t
i
o
n
.
. Th
em
u
l
t
i
v
a
r
i
a
b
l
eup
rboundc
a
l
c
u
l
a
t
i
o
n(
d
e
v
e
l
o
p
e
di
nc
h
a
p
t
c
r
5
)
i
su
s
e
df
o
r

~t

:
Ez
r
1
;
z
;
;

1
i

.;

:
:
;

:
2
:

I
n
1

.
:

:
:

;
:

i::

t
h
ci
t
c
r
a
t
i
o
n
sa
r
cp
r
i
n
t
e
d
.

h
e
r
e
0
hd

d
i

i
I
1

cr

e L i
g
nv
a
r
i
a
b
l
e
sand

c
t
i
o
n
sf
o
ra
l
lo
f

e
:
U g

RandomWalk('Example6_1' [0.5 0.5]


20
0.001
0
1
20) (
6
.
7
)
I
n
p
u
t
:(
T
h
e
ya
r
et
h
i
t
c
m
se
n
c
l
o
s
e
dw
i
t
h
i
nt
h
eo
u
t
e
r
m
o
s
tp
a
r
e
n
t
h
e
s
i
si
ne
x
p
r
e
s
s
i
o
n
(
6
.
7
)
.
)
'
E
x
a
m
p
l
e
6
_
1
:
[
0
.
5
0.
5
]
20
0
.
0
0
1

0
20

l
Cf
i
l
eEx
ampl
e6
_1.mw
h
e
r
el
J
leo
b
j
e
c
t
i
v
ei
sd
e
s
c
r
i
b
e
d
r
O
)
s
t
a
r
t
i
n
gd
c
s
i
g
nv
e
c
l
o
r(
d
va
numbero
fi
t
e
r
a
t
i
o
n
so
ft
h
emethod(
n
i
t
e
r
)
o
l
d
e
ns
e
c
t
i
o
np
r
o
c
e
d
e(
t
o
l
)
t
o
l
e
r
a
n
c
ef
o
reg
i
n
i
t
i
a
ls
t
e
p
s
i
z
ef
o
rs
c
a
n
n
i
n
gu
p
p
e
rbound(
Io
wbound)
s
l
e
pi
n
t
e
r
v
a
lf
o
rs
c
a
n
n
i
n
g(
i
n
t
v
I
)
I
s
)
numbero
fs
c
a
n
n
i
n
gs
t
e
p
s(
n a

Z
.~

3.0804
3.0804
3.0804
3.0804
ans
3.0804

2.0350
2.0350
2.0350
2.0350

3.0020
3.0020
3.0020
3.0020

2.0350

3.0020

Useo
fp
r
e
v
i
o
u
sc
o
d
ei
I
Iu
s
a
t
e
sc
o
d
er
e
u
s
eandm
o
d
u
l
a
rp
r
o
g
r
a
m
m
i
n
g
.Oncemore
r
d
e
r
sa
r
s
t
r
o
n
g
l
ye
n
c
o
u
r
a
g
e
d1
0v
i
s
i
tt
h
ec
o
d
et
ou
n
d
e
r
s
t
a
n
dt
h
a
n
s
l
a
t
i
o
no
ft
h
e
a
Jg
o
r
il
J
lma
n
ds
u
b
s
e
q
u
e
n
ta
p
p
l
i
c
a
t
i
o
n
.

20 iterations

a
Iuto e a
c
t
u
a
l u
t
i
o
no
fx
i=3
xi=2
j
*=
3
.T
si
s
Compare v
r
e
m
a
r
k
h
l
ef
o
ramethoda
th
a
snos
e
n
s
eo
fp
r
e
f
e
r
r
e
dd
i
r
e
c
t
i
o
n
.I
ft
h
ee
x
e
c
u
t
i
o
n15
r
e
s
a
t
e
dw
i
t
hM
e
n
u
c
a
l
i
n
p
u
t

me
ntimnews
o
l
u
t
i
o
nw
i
l
lb
e
o
b
t
a
i
n
e
d
.
F
i
g
u
r
e
6
2
t
r
a
c
k
s
g
e
so
f
t
h
ed nv
a
r
i
so
nt
h
ec
o
lfp
l
ot
.Themovem
n
to
fi
l
i
e s
i
g
n
t
h
; I
v
a
r
i
a
b
l
e
sa
r
ep
a
i
r
e
dw
i
t
ht
h
ec
o
r
z
e
s
p
o
n
d
i
n
gi
t
e
r
a
t
i
o
nnumber.Therandomw
a
l
k
f
u
n
c
t
i
o
nc
a
nb
E
E
x
E
C
U
t
e
df
o
rs
e
v
e
r
a
ld
i
f
f
E
r
e
n
ts
t
a
r
t
i
n
gs
e
t
so
fd
e
s
i
g
nv
a
r
i
a
b
l
e
s
.
I
t
.Th
eGoldSection_nVar.mi
ses
a
r
n
eonea
si
nC
h
a
p
l
e
r5
s
h
o
u
l
dn
o
td
i
s
a
p
p
o
i
nt
e
x
c
e
p
t i
l
i
ep
r
i
n
t
i
n
gh
a
sb
e
e
ns
u
p
p
r
e
s
s
e
d
.
h
ep
r
i
n
t
i
n
go
f i
t
e
r
a
t
i
o
nmb?
Whilee
x
e
c
u
t
i
n
geRandomWalk.mt
c
m
a
t
e
samessa
syoua
p
p
r
o
a
c
ht
h
es
o
l
u
t
i
o
n
.
Y
o
uc
a
ns
w
i
t
c
hi
to
f
f
b
ycommenttqgme
( )
s
t
a
t
e
m
e
n
t
T
h
g
e
o
d
e
i
nRandomwalk-ma
l
s
os
e
I
v
e
sa
sat
e
m
p
l
a
t
e
f
o
r
o
t
h
e
r
a
l
g
o
r
i
i
l
i
m
s
.l
ti
sw
o
r
t
he
x
p
l
o
r
i
n
gi
nd
e
t
a
i
l
.

O
u
t
p
u
t
:
Av
e
c
t
o
ro
fn+1v
a
Ju
e
so
fd
e
s
i
g
nv
a
r
i
a
b
l
e
sa
tt
h
el
a
s
ti
t
e
r
a
t
i
o
na
n
dt
h
ev
a
Ju
eo
ft
h
e
o
b
j
e
c
t
i
v
cf
u
n
c
t
i
o
n
.
J
le Command w
indow p
r
o
d
u
c
e
st
h
ef
o
l
l
o
w
i
n
g
E
x
e
c
u
t
i
n
gf
u
n
c
t
i
o
n(
6
.
7
)i
nl
i
n
f
o
r
m
a
t
i
o
n(
ecommentsa
r
ea
d
d
e
dl
a
t
e
r
)
:
k

- The design vector and function during the iterations


0011112288444444
0099998803000000

nunununununn3nEnononunDnD
u

=RJRd 'uu Jnunvnununununu

m0000000023333333

....
0.5000

0.5000
1.0440
1.0440
1.0440
1.0440
0.9756
0.9756
2.1093
2.0481
2.0350
2.0350
2.0350
2.0350
2.0350
2.0350

5.3125
5.3125
4.6481
4.6481
4.6481
4.6481
4.6349
4.6349
3.0663
3.0025
3.0020
3.0020
3.0020
3.0020
3.0020
3.0020

start
did not move
decreased function
didnot move
didnot move
didnot move
d
ecreased function
etc

0
.
5

i1j
Lii
1J
ll
l i--

6
.
2N
U
M
E
R
I
C
A
LT
E
C
H
N
I
O
U
ES-N
O
N
G
R
A
D
I
E
N
TMETHODS 233

232 N
U
M
E
R
I
C
AlT
E
C
H
N
I
O
U
E
SF
O
RUNCONST
RAlN
EDo
p
n
M
I
Z
A
T
I
O
N

1
.
5

F
l
g
u
r
e6
.
2 R
andomw
a
l
k
:Exam
p
l
e6
.
1
.

234 N
U
M
E
A
I
C
AlT
E
C
H
N
I
O
U
E
SF
O
U
NCONSTA
A
lN
EDO
P
T
I
M
I
Z
A
T
I
O
N

6
.
2N
U
M
E
A
I
C
AlT
E
C
H
N
I
O
U
E
S
O
N
G
A
A
O
I
E
N
TMETHODS 235

6
.
2
.
2 P
a
t
t
e
r
nSearch

j j+l

TheP
a
t
t
e
mS
e
a
r
c
hmethodi
sam
i
n
o
rm
o
d
i
f
i
c
a
t
i
o
n1
0l
h
eU
t
l
i
v
a
r
i
a
t
emethodw
i
t
ha
m
a
J
o
rI
m
p
a
c.
tAst
h
eP
a
t
t
e
mS
e
a
r
c
hu
s
e
st
h
eU
n
i
v
a
r
i
a
t
emethodt
h
el
a
t
t
e
ri
sn
o
t
a
1
s
oknown t
h
eC
y
c
l
i
cC
o
o
r
d
i
n
a
l
e
d
i
s
c
u
s
s
e
ds
e
p
a
r
a
t
e
l
y
.I
nt
h
eU
n
i
v
a
r
i
a
t
emethod
a
c
hd
e
s
i
g
nv
a
r
i
a
b
l
e(
c
o
n
s
i
d
e
r
e
dac
o
o
r
d
i
n
a
)p
r
o
v
i
d
e
sas
e
a
r
c
h
D
e
s
c
e
n
lmethode
d
i
r
e
c
t
i
o
n
.T
h
i
si
sa
1s
or
e
f
e
e
dt
oa
sac
o
o
r
d
i
n
a
t
ed
i
r
e
c
t
i
o
n di
se
a
s
i
l
ye
x
p
r
e
s
s
e
d
t
h
r
o
u
g
h
t
h

1
e
u

n
1
n
u
mb

r
o
fv

a
r
i
a

b
l
e
s

s
d
e

s
i

g
nv
a
r
i
a
b
l
e
s
)
i
t
e
r
a
t
i

on
so
n
e
f
o
ra
c
h
oft

hens
e
a
r
c
hd
i
r
e
c
t
i
o
n
s
.I
tcanbeshownbv
app
I
ic
a
t
i
o
n tf
o
rp
r
o
b
l
e
m
sw
i
t
hc
o
n
s
i
d
e
r
a
b
l
en
o
n
l
i
n
e
a
t
h
eU
n
i
v
a
d
a
t
e
m
e
t
1
3
0
J
t
e
n
d
st
og
e
tl
o
c
k
i
n
t
o
a
z
i
g

z
ag
p
a
t

emofs
m
m

a
s
o
l
u

t
1
1

0n
.

l
l
e
P
a
t
m
Searchprc
d
u
r
ea
t
t
e
m
p
t
st
od
i
s
r
u
p
tt
h
i
sz
i
g
z
a
gb
e
b
a
v
i
o
rbye
x
e
c
u
t
i
n
go
n
e
a
d
d
i
t
i
o
n
a
li
t
e
r
a
t
i
o
nf
o
re
a
c
hc
y
c
l
e
.I
ne
a
c
hc
y
c
l
e

a
tee
n
do
f1
1U
n
i
v
a
r
i
a
t
ed
i
r
e
c
t
i
o
n
st
h
e
n+lsea
hd
i
r
e
c
t
i
o
ni
sa
s
s
e
m
b
l
1
a
sa
I

I
I
n

a
r

m
1

on
s
d

eoptimum u
e
o
f
t
h

l
e
p
z

ef

o a
td
i
r
e
c
t
i
o
n
. o
n
e
d
i
m
i
o
n
a
l
d
nm
p
u
t
e
d dt
h
en
e
x
tc
y
c
l
eo
fi
t
e
r
a
t
i
o
nbe
g
I
l
1
s
.
o
p
t
i
m
a
ls
t
e
p
s
i
z
ei
se

GoToStep2
T
h
e
P
a
t
t
e
r
n
S
e
a
r
c
h
m
e
t
h
o
d
h
a
s
a
f
c
w
a
d
d
i
t
i
o
n
a
l
s
t
o
p
p
i
n
g
c
r
i
t
c
r
i
a
.
T
h
c
s
c
a
r
e
b
a
s
e
d
o
n
howmucht
h
e
f
u
n
c
t
i
o
ni
s
d
e
c
r
e
a
s
i
n
gf
o
re
a
c
hc
y
c
l
e
a
n
dhowmuchchangei
st
a
k
i
n
g
l
1f
o
r
m
a
t
i
o
ni
s
b
a
s
e
dont
h
e
l
e
n
g
t
h
p
l
a
c
e
i
nt
h
c
d
e
s
i
g
nv
a
ri
a
b
l
e
i
t
s
e
l
f
.Fort
h
e
l
a
t
t
e
rt
h
i
sI
o
f
t
h
e
c
h
a
n
g
e
i
nt
h
e
d
e
s
i
g
nv
e
c
t
o
rf
o
rt
h
e
c
y
c
l
e
.Onceagaintheapproachi
s
s
i
m
p
l
e
a
n
d
t
h
ei
m
p
l
e
m
e
n
t
a
t
i
o
ni
sd
i
r
e
ct
.Programmings
u
c
hat
e
c
h
n
i
q
u
ew
i
l
ln
o
tbct
o
o
f
o
r
b
i
d
d
i
n
g
.
In
ueee
x
p
o
s
u
r
e1
0MA
Tu.Bp
rogrammingi
s
m
e
t
h
o
d
PatternSearch.m: Toc
o
nl
i
si
m
p
l
e
m
e
n
t
e
dv
e
r
yd
i
f
f
e
r
e
n
1
tyfromt
h
eo
t
h
e
ra
l
g
o
r
i
t
h
m
st
h
a
th
a
v
eb
c
e
nt
r
a
n
s
l
a
t
e
d
.I
ti
sas
t
a
n
d
a
1onep
r
o
g
r
a
m
.T
h
e
r
ca
r
ep
r
o
m
p
t
sf
o
r e
ri
n
p
u
t
sw
i
i
n
t
oc
o
d
es
of
i
a
p
p
r
o
p
r
i
a
t
ed
e
f
a
u
l
tv
a
l
u
e
si
nc
a
s
t
h
eu
s
e
rd
e
c
i
d
e
sn
o
tt
oe
n
t
e
ra v
a
l
u
e
.The
o
n
e
d
i
m
e
n
s
i
o
n
a
ls
t
e
p
s
i
z
ec
o
m
p
u
t
a
t
i
o
ni
si
m
p
l
e
m
e
n
t
e
dbyem
u
l
t
i
v
a
r
i
a
b
l
eg
o
l
d
e
n
s
a
r
ea
sf
ol
1o
w
s
.
s
e
c
t
i
o
nm
c
t
h
o
d
.Thei
n
t
e
r
e
s
t
i
n
gf
e
a
t
u

.I
ti
saf
r
e
e
s
t
a
n
d
i
n
gp
r
o
g
r
a
m
.
l
ef
i
l
ec
o
nt
a
I
l
1i
n
gt
h
eo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
ni
ss
e
l
e
c
t
e
dt
h
r
o
u
g
hal
i
s
tb
o
x
.

Ao
r
i
t
h
m
:P
a
t
t
e
r
nSearch6
.
3
)
S
t
e
p1
.ChooseX1

Nc(
n
u
m
b
e
ro
fc
y
c
Ie
s
)
j(
1
)=j(X Xc(
1
)=X

S
c
v
e
r
a
lp
r
o
g
r
a
mc
o
n
t
r
o
lp
a
r
n
e
t
e
r
sa
r
ee
n
t
e
r
e
df
r
o
mt
h
ek
e
y
b
o
a
r
da
f
t
e
ra
s
u
i
t
a
b
l
eprompti
n
d
i
c
a
t
i
n
gd
e
f
l
l
u
l
tv
a
l
u
e
s
.

E
2
:t
o
l
e
r
a
n
c
ef
o
rs
t
o
p
p
i
n
gc
r
i
t
e
r
i
a
S
e
t
j=I(
i
n
i
t
i
a
l
i
z
ec
y
c
l
ec
o
u
n
t
)
S
t
e
p2
.F
o
re
a
c
hc
y
c
l
ej

D
c
f
a
u
l
tv
a
l
u
e
sa
r
ci
n
i
t
i
a
l
i
z
e
di
ft
h
eu
s
e
rc
h
o
o
s
c
sn
o
t1
0e
n
t
c
rt
h
ev
a
l
u
e
s
.
Thcd
e
s
i
g
nv
c
c
t
o
ri
st
r
a
c
k
e
da
n
dp
r
i
n
t
c
da
f
t
c
rt
h
cp
r
o
g
r
a
mi
ss
t
o
p
p
e
d
.

;
)

1
1
Fori=1
S;=(
u
n
i
v
a
r
i
a
t
e p
)
X
;
+ =X;+

/i
sd
e
t
e
r
m
i
n
e
db
ym
i
n
i
m
i
z
i
n
g
j
(
X
;
+
)
(
s
t
o
r
ev
a
l
u
e
so
f

.
ia
n
dS
/
)
cndo
fF
o
rl
o
o
p

=:;S;aX

-X (Pattcm

Xj=X
+ +C

i
S
j
n
X+1
)Xs J
"

+1
)=j(
)(
s
t
o
r
c
y
c
l
ev
a
l
u
e
s
)
1
r
U+1)-1
r
AX=X
I)-X
S
t
e
p
3
. 6j=
c

.
1
j
lSE"s
t
o
p
I
fl
T.
IfAX
1
xS~. s
t
o
p

.
u
+

I
f
j=N s
t
o
p
X ~

F
o
rt
w
o
v
a
r
i
a
b
l
ep
r
o
b
l
e
mt
h
eo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
nc
o
n
t
o
u
r
sa
r
ep
l
o
t
t
e
d
.
. Thef
i
r
s
tn
i
n
eo
rI
c
s
si
t
e
r
a
t
i
o
n
sa
r
et
r
a
c
e
don c
o
n
t
o
u
r
p
l
o
.
tThcU
n
i
v
a
r
i
a
t
e
a
n
d
t
h
ep
a
t
t
e
ms
t
e
p
sa
r
ec
o
)
o
rc
o
d
c
d
.

e
: PatternSearch
U g

"

S
i
n
c
et
h
eu
p
p
e
rboundc
a
l
c
u
l
a
t
i
o
n
sa
r
cb
a
s
e
donp
o
s
i
t
i
v
cv
a
l
u
c
so
fs
t
e
p
s
i
z
e

t
h
c
s
e
a
r
c
hd
i
r
e
c
t
i
o
ni
s v
c
r
s
i
f tumeds
t
e
p
s
i
z
ei
sesamea
st
h
el
o
w
c
r
b
o
u
n
d
.
I
t
e
r
a
t
i
o
nc
o
u
n
t
sa
r
er
e
c
o
r
d
e
d
.

(
6
.
8
)

Thep
r
o
b
)
e
mi
sExample6.
1
.
l
ef
i
]
ei
sEx
ample6_1.m.
Output(MAT
l
.
A
BCommandw
i
n
d
o
w
)
:Withed
e
f
a
u
l
tv
a
)
u
e
sandt
h
es
t
a
r
t
i
n
g
v
e
c
t
o
r[
43
]t
h
ef
o
l
l
o
w
i
n
ga
r
ew
r
i
t
t
e
nt
ot
h
eCommandwindow.
>> PatternSearch
The function for which the minimum is sought must be a
MATwa functionM - File. Given a vectordependentvariable
it must return a scalar value. This is the function to be
MINIMIZED. Please select function name in the dialog box
and hit return:

236 N
U
M
E
A
I
C
A
LT
E
C
H
N
I
O
U
E
SF
O
AU
N
C
O
N
S
T
A
A
I
N
E
DO
P
T
I
M
I
Z
A
T
I
O
N

N
G
A
A
D
I
E
N
TMETHODS 237
6
.
2N
U
M
E
A
I
C
A
LT
E
C
H
N
I
O
U
E9-NO
2.9998
2.9998
2.9998
2.9998
2.9998

Input the starting design vector. This is mandatorv as


there is no default vector setup. The length of our
vector indicates the number of unknowns. Please enter it
now and hit return:

1.9999
1.9999
1.9999
1.9999
1.9999

3.0000
3.0000
3.0000
3.0000
3.0000

0.0000
0.3398
0.0000
0.0000
0.2020

Tov
e
r
i
f
y eprogrami
se
x
e
c
u
t
i
n
gt
h
ea
1g
o
r
i
t
hm
s
c t
h
ed
e
s
i
g
nv
e
c
l
o
r
s E
f
i
r
s
tlWoc
o
l
u
m
n
s
.Theu
n
i
v
a
r
i
a
l
es
t
e
p
sc
a
nbr
e
c
o
g
n
i
z
e
dwh
no
necomponento
ft
h
e
v
e
c
1
0
rc
h
a
n
g
e
sw
h
i
l
e o
t
h
e
rm
a
i
n
t
a
i
n
si
t
sv
a
l
u
e
.Thep
a
t
t
e
r
nd
i
r
e
c
t
i
o
n
sc
a
nb
e
[
4 3]
;
1
. r
e
c
o
g
n
i
z
e
dwhenb
o
t
hc
o
m
p
o
n
e
n
t
sc
h
a
n
g
e
.
The initial design vector [4.00 3.00]
~:
Thenumbero
fi
t
e
r
a
t
i
o
n
si
sas
i
g
n
i
f
i
c
a
n
tamoun.
tOner
e
a
s
o
ni
st
h
a
tt
h
et
o
l
e
r
a
n
c
e
s
Convergence in f
: 3.999E 009 reached in 10 iterations
i
f
f
e
r
e
n
c
eb
e
l
w
e
e
nlwoi
l
e
r
a
t
i
o
n
sch
g
e
s
a
r
emuchs
m
a
l
l
e
r
.I
tc
a
nben
o
t
e
dt
h
a
t d
Number of useful calls to the Golden Section Search
i
I
n
e
a
r
l
y
.Suchm
e
t
h
o
d
sa
r
ee
x
p
e
c
1
e
dt
or
u
nf
o
ral
a
r
g
enumbero
fi
t
e
r
a
t
i
o
n
s
.The
Method: 31
:
1 p
r
o
g
r
a
mcon
0
1p
a
r
t
e
r
sc
o
n
t
r
o
lt
h
ea
c
c
c
yo
fer
e
s
u
l
t
sa
sw
e
l
la
senumber
Total number of calls to the Golden Section Search
t
h
a
nt
h
eRandomWa
1
kme
o
d
.
o
fi
t
e
r
a
t
i
o
n
s
.Thes
o
l
u
t
i
o
ni
smorea
c
c
u
r
a
t
Method: 40
F
i
g
u
r
e6
.
3i
I
Iu
s
t
r
a
l
e
sea
1gorithmg
r
a
p
h
i
c
a
l
l
y
.Theu
n
i
v
a
r
i
a
t
ea
n
dt
h
ep
a
U
e
r
n
d
i
r
e
c
t
i
o
n
s
a
r
e
q
u
i
t
e
d
i
s
t
i
n
g
u
i
s
h
a
b
l
e
.
Wh
e
n
e
x
e
c
u
t
e
d
o
n
y
o
u
r
m
a
c
h
i
n
e
t
h
s
t
e
p
s
a
r
e
The values for x and f are
e
x
p
l
i
c
i
t
l
yc
o
l
o
rc
o
d
e
d
.R
e
a
d
e
r
sa
r
ee
n
c
o
u
r
a
g
e
d1
0goo
v
e
rt
h
ec
o
d
e
.andmodifyi
1t
o
4.0000
3.0000
4.2500 0
4.4992
3.0000
4.0000 0.4996
4.4992
2.6924
3.6916
0.3081
E
4.3868
2.7616
3.6398
0.2254
4.1432
2.7616
3.5801 0.2446
4
.
5
4.1432
2.5287
3.4022
0.2339
3.8229
2.2224
3.2890 1.3148
3.3338
2.2224
3.0494
0.4896
3.3338
2.1543
3.0343
0.0681
3.1750
2.1321
3.0180 0.3248
3.1980
2.1321
3.0175 0.0234
3.1980
2.0916
3.0121
0.0405
3.1944
2.0980
3.0118
0.1579
3.1473
2.0980
3.0096 0.0471
3.1473
2.0682
3.0067
0.0298
3.0346
1.9969
3.0016 2.3916
2.9958
1.9969
3.0000 0.0391
2.9958
1.9981
3.0000 0.0013
2.9971
1.9981
3.0000
0.0322
2.9972
1.9981
3.0000 0.0001
2.9972
1.9986
3.0000
0.0005
2.9972
1.9988
3.0000
0.4563
2.9977
1.9988
3.0000 0.0005
2.9977
1.9990
3.0000
0.0002
3
0
.
5
1
.
5
2.9999
1.9999
3.0000 4.9305
2.9998
1.9999
3.0000 0.0002

Rgure6
.
3P
a
l
l
e
ms
e
a
r
c
h
:Exam
p
l
e6
.
1
.

itill--

The function you have chosen is ::Example6_1


maximum number of cycles [1000]:
convergence tolerance for difference in f[le8
]:
convergence tolerance on change in design x [le-8]:

N
U
M
E
A
I
C
A
LT
E
C
H
N
I
Q
U
E
O
N
G
R
A
D
I
E
N
TMETHODS 239

Xf=X
++
j
n

t
t
h
i
rp
r
e
f
e
r
e
n
c
e
.Thenumbero
fl
i
n
e
si
sa
r
o
u
n
d2
5
0
.Wea
t
r
t
i
n
gt
og
e
ti
n
t
o
r
e
f
l
s
e
n
o
u
sprogrammmg.

Xc
U
+1
)Xf; fcU+ f(
X
j
)(
s

c
yc
I
e

S
t
e
p3
.!
:
i
f
=fcU+1
)

U
)
;d1{ XcU+1
)-X
c
(
j
)
6
.
2
.
3 Powell'sM
e
'
od

l
ft
h
e
r
ew
e
r
eo
n
l
y one z
e
r
o
o
r
d
e
r method must be p
r
o
g
r
a
n
u
n
e
dt
h
e
overwhelmingc
h
o
i
c
ewouldbePowe
l
1'
smethod[
3
]
.Thep
r
i
n
c
i
p
a
lr
e
a
s
o
nf
o
re
d
e
c
i
s
i
o
nwouldb
et
h
a
ti
th
a
st
h
ep
r
o
p
e
r
t
yo
fq
l
l
a
d
r
a
t
i
cc
o
n
v
e
r
g
e
n
c
e

namely
f
o
ra
q
u
a
d
r
a
t
i
cp
r
o
b
l
e
mw
i
t
hnv
a
r
i
a
b
l
e
sc
o
n
v
e
r
g
e
n
c
w
i
l
lb
ea
c
h
i
e
v
e
di
nl
e
s
st
h
a
no
re
q
u
a
l
l
1cy
c
I
e
s
.A q
u
a
d
r
a
t
i
cp
r
o
b
l
e
mi
sa
nu
n
c
o
n
s
i
n
e
dm
i
n
i
m
i
z
a
t
i
o
no
fa
t
onPowe
f
u
n
c
t
i
o
nt
h
a
ti
se
x
p
r
e
s
s
e
da
saq
l
.
d
r
a
t
i
cpolynomial-ap
o
l
y
n
o
m
i
a
lw
i
t
hnot
e
r
m
h
a
v
i
n
g ad
e
g
r
e
eg
r
e
a
t
e
rt
h
a
nt
w
o
. Example6
.
1i
sa
n example o
faq
u
a
d
r
a
t
i
c
p
o
l
y
n
o
m
i
a
li
ntwov
a
r
i
a
b
l
e
s
.E
n
g
i
n
e
e
r
i
n
gd
e
s
i
g
no
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
e
m
sa
r
er

r
e
l
y
d
e
s
c
r
i
b
e
dbyaq
u
a
d
r
a
t
i
cp
o
l
y
n
o
m
i
a
l
.

sd
o
e
sn
o
ti
m
p
l
yt
h
a
tyouc
a
n
n
o
tu
s
ePowe
l
1
's
m
e
t
h
o
d
.Wh
a
tt
h
i
smeansi
st
h
a
tt
h
es
o
l
u
t
i
o
ns
h
o
u
l
dn
o
tbee
x
p
e
c
t
e
dt
oc
o
n
v
e
r
g
e
q
u
a
d
r
a
t
i
c
a
l
l
y
.F
o
rn
o
n
q
u
a
d
r
a
t
i
cp
r
o
b
l
e
m
s
a
st
h
es
o
l
u
t
i
o
ni
sa
p
p
r
o
a
c
h
e
di
t
e
r
a
t
i
v
e
l
y
eo
b
j
e
c
t
i
v
ec
a
nbea
p
p
r
o
x
i
m
a
t
e
dv
e
r
yw
e
l
lbyaq
u
a
d
r
a
t
i
cf
u
n
c
t
i
o
n
.I
ti
sa
ti
ss
t
a
g
e
a
tt
h
e
q
u
a
d
r
cc
o
n
v
e
r
g
e
n
c
p
r
o
p
e
r
t
yi
sr
e
a
Ji
z
e
di
nt
h
ec
o
m
p
u
t
a
t
i
o
n
s
.
Thea
c
t
u
a
la
Jg
o
r
i
t
h
m(A6.
4
)i
sas
i
m
p
l
e(
t
h
ewordi
sn
o
tb
e
i
n
gu
s
e
dI
ig
h
t
l
y
)
m
o
d
i
f
i
c
a
t
i
o
nt
ot
h
ePattemS
e
a
r
c
ha
l
g
o
r
i
t
h
m
.l
ne
a
c
hc
yc
Je(
a
e
ref
i
r
s
t
)
i
n
s
t
e
a
d
t
e
r
a
t
i
o
n
s
.t
h
es
e
a
r
c
hd
i
r
e
c
t
i
o
n
sa
r
e
o
fu
s
i
n
gu
n
i
v
a
r
i
a
t
ed
i
r
e
c
t
i
o
n
si
nt
h
ef
i
r
s
tni
o
b
t
a
i
n
e
dfromt
h
ep
r
e
v
i
o
u
sc
yc
Je
.Thenews
e
a
r
c
hd
i
r
e
c
t
i
o
n
sa
r
eo
b
t
a
i
n
e
dbyI

s
h
i
f
t
i
n
gt
h
ed
i
r
e
c
t
i
o
n
so
ft
h
ep
r
e
v
i
o
u
scy
c
Je
.I
ni
swayah
i
s
t
o
r
yo
ft
h
ep
r
e
v
i
o
u
s
s
e
a
r
c
hd
i
r
e
c
t
i
o
n
si
su
s
e
dt
oe
s
t
a
b
l
i
s
ht
h
en
e
x.
tWitht
h
i
sc
h
a
n
g
e
.P
o
w
e
l
l
'
smethod
Je
sf
o
rExamplc6
.
1r
a
t
h
e
rt
h
a
nt
h
es
e
v
e
r
a
l
s
h
o
u
l
dc
o
n
v
e
r
g
ea
tmosti
nt
h
r
e
ec
yc
l
i
s
t
e
di
nt
h
ep
r
e
v
i
o
u
ss
e
c
t
i
o
n
.
Ao
r
i
t
h
m
:P
o
w
e
l
l
'
sMethod 6
.
4
)
S
t
e
p1
.ChooseX" Nc(numberofc
y
c
l
e
s
)
;Xc
J
tl
)=j
{X1)
{
l
)=X
E

E
2
:t
o
l
e
r
a
n
c
ef
o
rs
t
o
p
p
i
n
gc
r
i
t
e
r
i
a
S
e
t
j=1(
i
n
i
t
i
a
l
i
z
eP
o
w
e
l
lc
y
c
l
ec
o
u
n
t
)
Fori=1

n
S
;=~; (
u
n
i
v
a
r
i
a
t
es
t
e
p
)
S
t
e
p2
.F
o
re
a
c
hy
c
l
ej
Fori=l.n
I
fj2
S
;S
;
+
X
;
+ =X;+ISI
t
Xi
+
'
)

;i
sd
e
t
e
n
n
i
n
e
dbym
i
n
i
m
i
z
i
n
gJ
e
n
do
fForl
o
o
p

c
.

If j=Nc sp

X
XJ

jj+l
GoToStep2

APPMeattonofpoweII'sMethod:Thet
r
a
n
s
l
a
t
i
o
no
ft
h
ea
l
g
o
r
i
t
h
mi
n
t
oMATLAB
C
O
d
e
i
s
l
e
f
ta
sm e
x
e
r
c
i
s
ef
o
rt
h
es
t
u
d
e
n
t
.
S
t
e
p
b
y
s
t
e
pa
p
p
l
i
c
a
G
o
ntoExample6
.
1i
s
shownh
e
r
e
.Example6
.
1i
sr
e
s
d
as

Exam
p
l
e
6
.
1

1
.

Minimize f
(
X
)=f(x"X2)=3+(
x

Sj=S;+=I;
S
;=Xn
+ XJ

(
6.
3
)

=[g:~lfl 5.3125; E
=

S =[6];

=[~]

S
2

j=l
S
t
e
p2
.j=1

X2431+l

=??l
2
.

i 2
:

444615;X4zj;)=4.5576
i=3:

+)[~]=[5]

S
)=0
.
2
5[
6
]

i
5
]

=1
/
4
; X2=

X3 =+[o.~1

X4=[03251
(p
a
t
t
e
ms
t
e
p
)

5X
x
2
2
)
2+(

(
6.
4
)

l
fl
d
f
lSE"s
t
o
p
l
fd1{TdXSE
2
.s
t
o
p

111 f H {

62

238 N
U
M
E
A
I
C
A
LT
E
C
H
N
I
Q
U
E
SFOAU
N
C
O
N
S
T
A
A
I
N
E
DO
P
T
I
M
I
Z
A
T
I
O
N

2
4
0 N
U
M
E
R
I
C
A
LT
E
C
H
N
I
Q
U
E
SF
O
RU
N
C
O
N
S
T
R
A
I
N
E
DOP
T
lM
I
Z
A
T
l
ON

6
.
3

L
=
1
1
:
5
7
0
;

;=0.
42
3
5
;

f(X
.
4843
4)=4

s-GRADIENT-BASEDMETHODS
6
.
3 NUMERICALTECHNIQUE
I
nl
i
g
h
to
ft
h
ep
r
e
v
i
o
u
sd
e
f
i
n
i
t
i
o
n
.t
h
e
s
ew
o
u
l
db
er
e
f
e
r
r
e
dt
oa
s
f
i
r
s
t
.
o
r
d
e
rmo
d
s
.
Thes
e
a
r
c
hd
i
r
e
c
t
i
o
n
sw
i
l
lb
ec
o
n
s
t
r
u
c
t
e
du
s
i
n
gt
h
eg
r
a
d
i
e
n
to
ft
h
eo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
n
.
S
i
n
c
eg
r
a
d
i
e
n
t
sa
r
eb
e
i
n
gc
o
m
p
u
t
e
d
. eK
u
h
n
T
u
c
k
e
rc
o
n
d
i
t
i
o
n
s (F
OC) f
o
r
u
n
c
o
n
s
t
n
e
dp
r
o
b
l
e
m
s
.V
f=O
.c
a
nb
eu
s
e
dt
oc
h
e
c
kf
o
rc
o
n
v
e
r
g
e
n
c
e
.TheSOCa
r
e

gu

S
t
e
p3
.l
.
d
fl
=1
.
353>E
I
'c
o
n
t
i
n
u
e
T
l
dX dXl
=1
.
8
9
9
2>E
z
.c
o
n
t
i
n
u
e
}=2
:Onec
y
c
l
eo
v
e
r
GoToStep2
S
t
e
p2:}=2
xJ0.855

J=
1
1
1c'7n:
1
1
.
l
5
7
0'
l
1

5
5
81

=-0
.
1
4
1
1
1
'

'Y
.8
1- ..l ~OO; 1

.2
=1
1
.0
1
0
4
1
;f(X)=4.
41
4
5
2

1
1
1
i
Z
]
;

S2=[0 25J =

:
_
i=3

r
0.
10

1
1

r
'
ll

U
.1966J~3 =.0396/:3=20~ =/
2
/
;f(X =3

4)

:
:
i;;32!emd

C
u
t
a
t
i
o
n
si
ni e
r
c
i
s
e s
t
o
p
p
e
d

;:Jr:atJ;;::::wouldit

itE:27

h
e
r
e
?You a
g
e
dt
o

-lem
e
t
h
o
do
n
l
yt
o
o
kt
w
oc
y
c
l
e
st
oc
o
n
v
e
r
g
e(
a
se
x
p
e
c
t
e
d
)

Thee
x
t
e
n
to
fcompu o
ni
s s
a
m
ea
st
h
eP
a
t
t
e
mS c
hb
l
enumberof

i
t
e
r
a
t
i
o
n
si
ss
i
g
n
i
f
i
c
a
n
t
l
ysma1
l
e
r
.

2:;;212S22;:rsecondmleiS

t
i
v
e
.H
E
n

Thes
t
e
p
s
i
z
ef
o
rt
h
el
a
s
ti
t
c
r
a
t
i
o
ni
s20
.Thes
u
p
p
o
r
t
i
n
ga
l
g
o
r
i
t
h
mf
o
r
pt
h
eminimum
a
t el
a
r
;
;
:
:
:
!
1
g
a
n
dg
o
l
d
e
ns
e
c
u
o
nmustbea
b
l
et
O t

:
:

:
L

:
:
z
2
1;
f
:

Wh
i

i
l
e

n
o
t
i
l

u
I

s
l

m
t

dh

e
r

E
.
t

e
l
s
a
x
i

m
3

nm
mIl
um

b
?

r
'

0
f
c
y

c
1
e

st

o
c

o
n
v
e
r
:

g
e

n
1

c
e
f
l

o
r

:f gp
o
i
n
.
tY

2
j
z
:
3
2
i
f
i
:
5
1
3
2
t
i
:
;
m
a
t
r
i
x
.l
fs
e
a
d
i
r
e
c
t
i
o
n
s
.
Sj
a
n
dao
n~u
gate d
i

e
c

o
n
E

c
ou
g
a
c
ya
r
c
n
.

ti
tw
o
u
l
di
n
v
o
l
v
et
h
ec
o
m
p
u
t
a
t
i
o
no
fa
n
h
a
r
d
l
ye
v
e
ra
p
p
l
i
e
d
.Oneo
ft
h
er
e
a
s
o
n
si
sa
nX1s
e
c
o
n
dd
e
r
i
v
a
t
i
v
em
a
t
r
i
xw
h
i
c
hi
sc
o
n
s
i
d
r
e
dcompu
t
i
o
n
a
l
l
ye
x
p
e
n
s
i
v
e
.
v
a
l
u
a
t
i
o
no
ft
h
e0
t
i
v
ef
u
n
c
t
i
o
n q
u
i
r
e
sac
a
l
lt
oaf
i
n
i
t
ee
l
e
m
e
n
t
p
a
r
t
i
c
u
l
a
r
l
yi
fee
m
e
t
h
o
df
o
rg
e
n
e
r
a
t
i
n
gr
e
q
u
i
r
i
n
f
o
r
m
a
l
i
o
n
.A
n
o
t
h
e
rr
e
a
s
o
nf
o
rn
o
tc
a
l
c
u
l
a
t
i
n
gt
h
e
H
e
s
s
i
a
ni
sa
tt
h
ee
x
i
s
l
e
n
c
eo
fes
e
c
o
n
dd
e
r
i
v
a
t
i
v
ei
nar
e
a
ld
e
s
i
g
np
r
o
b
l
e
mi
sn
o
t
c
e
r
t
a
i
ne
v
e
nt
h
o
u
g
hi
ti
sc
o
m
p
u
t
a
t
i
o
n
a
l
l
yp
o
s
s
i
b
l
eo
rf
e
a
s
i
b
l
e
.F
o
rp
r
o
b
l
e
m
st
h
a
tc
a
n
11.ABs
h
o
u
l
db
ea
b
l
et
oh
a
n
d
l
ec
o
m
p
u
t
a
t
i
o
n
b
ed
e
s
c
r
i
b
e
db
ys
y
m
b
o
l
i
cc
a
l
c
u
l
a
t
i
o
n
s
.MA
o
fs
e
c
o
n
dd
e
r
i
v
a
t
i
v
a
tt
h
ep
o
s
s
i
b
l
es
o
l
u
t
i
o
na
n
di
e
i
g
e
n
v
a
l
u
e
s
.
W
i
t
h
o
u
tSOCt
h
e
s
em
e
t
h
o
d
sr
e
q
u
i
r
eu
s
e
r
'
sv
i
g
i
l
a
n
c
et
oe
n
s
u
r
e t
h
es
o
l
u
t
i
o
n
o
b
t
a
i
n
e
di
saminimumr
a
t
h
e
ra
namaximumo
ras
a
d
d
l
ep
o
i
n
t
.As
i
m
p
l way1
0
v
e
r
i
f
yi
si
st
op
e
r
t
u
r
bt
h
eo
b
j
c
t
i
v
ef
u
n
c
t
i
o
nt
h
r
o
u
g
hp
e
u
r
b
a
t
i
o
ni
nt
h
d
e
s
i
g
n
v
a
r
i
a
b
l
e
sa
tt
h
es
o
l
u
t
i
o
na
n
dv
e
r
i
f
yi
ti
sal
o
c
a
lm
i
n
i
m
u
m
.T
h
i
sb
r
i
n
g
su
pa
ni
m
p
o
r
t
a
n
t
o
f
t
h
e
s
e
m
e
ods-theyo
n
1
yf
i
n
dl
o
c
a
lo
p
t
i
m
u
m
s
.U
s
u
a
l
l
yt
h
i
sw
i
l
lb
ec
10
s
e
p
r
o
p
e
t
ot
h
d
e
s
i
g
nw
h
e
r
et
h
ei
t
e
r
a
t
i
o
n
sa
r
eb
e
g
u
n
.B
e
f
o
r
ec
o
n
c
l
u
d
i
n
gt
h
ed
e
s
i
g
n
Yt
oe
x
e
c
u
t
et
h
em
e
t
h
o
df
r
o
ms
e
v
e
r
a
ls
t
a
r
t
i
n
gp
o
i
n
t
st
o
e
x
p
l
o
r
a
t
i
o
n
.i
ti
sn
e
c
e
s
s
d
i
s
c
o
v
e
ri
fo
t
h
e
rminimumsx
i
s
ta
n
ds
e
l
e
c
tt
h
eb
e
s
to
n
eb
yh
e
a
dt
oh
e
a
d
c
o
m
p
a
r
i
s
o
n
.Theb
u
l
ko
fe
x
i
s
t
i
n
gu
n
c
o
n
s
t
r
a
i
n
e
da
n
dc
o
n
s
t
r
a
i
n
e
do
p
t
i
m
i
z
a
t
i
o
n
m
e
t
h
o
d
sb
e
l
o
n
gt
oi
sc
a
t
e
g
o
r
y
.
F
o
u
rm
c
t
h
o
d
sa
r
ep
r
e
s
e
n
t
e
d
.Thef
i
r
s
ti
st
h
cS
t
e
e
p
e
s
tD
e
s
c
e
n
tm
e
t
h
o
d
.W
h
i
l t
h
i
s
m
e
t
h
o
di
sn
o
tu
s
e
di
np
r
a
c
t
i
c
e
.i
tp
r
o
v
i
d
e
sa
ne
x
c
e
l
l
e
n
te
x
a
m
p
l
ef
o
ru
n
d
e
r
s
t
a
n
d
i
n
gt
h
c
r
a
d
i
e
n
t
b
a
s
e
dt
e
c
h
n
i
q
u
e
s
.Thes
e
c
o
n
di
st
h
eConjugate
a
l
g
o
r
i
t
h
m
i
cp
r
i
n
c
i
p
l
e
sf
o
r g
e
c
h
n
i
q
u
ew
h
i
c
hi
sac
l
a
s
s
i
c
a
lw
o
r
k
h
o
r
s
epa
c
u
l
a
r
l
yi
ni
n
d
u
s
u
s
a
g
e
.The
G
r
a
d
i
e
n
tt
t
h
i
r
da
n
df
o
u
r
t
hb
e
l
o
n
gt
ot
h
ec
a
t
e
g
o
r
yo
fV
a
r
i
a
b
l
eM
e
t
r
i
cm
e
t
h
o
d
s
.o
rQuasi-Newton
me
o
dstheya
r
ea
l
s
oc
a
l
l
e
d
.T
h
e
s
em
e
t
h
o
d
sh
a
v
cb
e
e
np
o
p
f
o
rsomet
i
m
e
. d
w
i
l
lp
o
s
s
i
b
l
ys
t
a
yt
h
a
twayf
o
ral
o
n
gt
i
m
et
oc
o
m
e
.
Thcg
c
n
e
r
a
lp
r
o
b
l
e
ma
n
ds
p
e
c
i
f
i
ce
x
a
m
p
l
ea
r
p
r
o
d
u
c
e
df
o
rc
o
n
v
e
n
i
e
n
c
e
M
i
n
i
m
i
z
e j
(
X
)
;
S
u
b
j
e
c
t

[
X
]
"

XI~Xj~xlt; i=I.2.....n

(
6.
1
)

(
6
.
2
)

M
i
n
i
m
i
z
e j
6
.
3
)
(
X
I
.X2)=3+(
X
I- 1
.
5
x
2
)
2+(X2- 2
)
2 (
(
X
)=f
S
u
b
j
e
c
tt
o
: OSXIS
5
:

oSX2S5

(
6
.
4
)

SJ
[U]Sj=0

V
e
r
i i
fi
ti
st
r
u
cf
o
rt
h
ce
x
a
m
p
l
e
.

(
6
.
9
)

6
.
3
.
1 SteepestDescentMethod
Thi
sm
e
t
h
o
dp
r
o
v
i
d
e
san
a
t
u
r
a
le
v
o
l
u
t
i
o
nf
o
rt
h
eg
r
a
d
i
e
n
tb
a
s
e
dt
e
c
h
n
i
q
u
e
s[
4
]
.
Theg
r
a
d
i
e
n
to
faf
u
n
c
t
i
o
na
tap
o
i
n
ti
st
h
ed
i
r
e
c
t
i
o
no
ft
h
em
o
s
tr
a
p
i
di
n
c
r
e
a
s
ei
n

ilillit---

v _
10
.
9
5
6
7
1 ~

N
U
M
E
R
I
C
A
LT
E
C
H
N
I
Q
U
Es-GR
A
D
I
E
N
T
.
B
A
S
E
DM
E
T
H
O
D
S 2
4
1

E
C
H
N
I
Q
U
Es-GR
A
D
I
E
N
T
B
A
S
E
DMETHODS 243
6
.
3N
U
M
E
R
I
CAlT

242 N
U
M
E
R
I
C
A
lT
E
C
H
N
I
Q
U
E
SF
O
RU
N
C
O
N
S
T
R
A
I
N
E
DO
P
T
I
M
I
Z
A
T
I
O
N

i
n
d
o
w
)
:A c
o
u
p
l
eo
fi
t
e
r
a
t
i
o
n
sa
tes
t
a
r
tanda
t
Output(
w
r
i
t
t
e
nt
oeCommandw
t
h
ev
a
l
u
eo
f
t
h
e
f
u
n
c
t
i
o
n
a
tt
l
l
a
tp
o
i
nt
.Thed
e
s
c
e
n
t
d
i
r
c
t
i
o
ncanbeo
b
t
a
i
n
e
dr
e
v
e
r
s
i
nl
!

E
E
n
d
(
o
u
t
o
f
2
0
i
t
e
r
a
t
i
o
n
s
)
M
e
c
o
p
i
e
d
b
l
o
w
.
t
h
eg
r
a
d
i
e
n
t(
o
rm
u
l
t
i
p
l
y
i
n
gi
tby1
)
.Th
en
e
x
ts
t
e
pwouldbet
or
e
g
a
r
dt
l
1ed
e
s
c
e
n
t
v
e
c
l
o
ra
sas
e
a
r
c
hd
i
r
e
c
t
i
o
na
f
t
e
ra
l
lwea
r
eaUempting1
0d
e
c
r
e
a
s
et
l
1ef
u
n
c
t
i
o
n
5
.
2
9
6
9
5
.
3
1
2
5
0
.
5
0
0
0
t
e
e
p
e
s
tD
e
s
c
e
n
t
t
h
r
o
u
g
hs
u
c
c
e
s
s
i
v
ei
t
e
r
a
t
i
o
n
s
.T
h
i
ss
e
r
i
e
so
fs
t
e
p
sg
i
v
er
i
s
et
oeS
0
.
5
0
0
0
2
.
8
4
1
1
4
.
7
8
3
2
0
.
9
4
8
3
a
l
g
o
r
i
t
h
m
0
.
5
9
9
6
3.
1379
4
.
3
7
6
6
0
.
8
4
2
9
1
.0701

A'
o
r
i
t
h
m
:S
t
l
P
t
Descent6
.

S
t
e
pJ
.C
hooseX
J N(
numbero
f
i
t
e
r
a
t
i
o
n
s
)
!
.
(
l
)=j
{
X1
)
;X.(
l
)=X1(
s
t
o
r
ev
a
l
u
e
s
)

JE
2
'E
J
:(
t
o
l
e
r
a
n
c
ef
o
rs
t
o
p
p
i
n
gc
r
i
t
e
r
i
a
)
S
e
ti
=1(
in
i
t
ia
1i
z
ei
t
e
r
a
t
i
o
nc
o
u
n
t
e
r
)

j=V
f
(
X
/
)
(
i
si
scomputedi
nS
t
e
p3
)
S
t
e
p2
.S
X
;
+
l=X;+~;
i
sd
e
t
e
r
m
i
n
e
dbym
i
n
i
m
i
z
i
n
g
f
(
X
;
+
I
)

X
s
(
i+1
)X 1
; (
i+1
)=f(X
+
)(
s
t
o
r
eva
1u
e
s
)
i
l
S
t
e
p
3
.a
J
=
!
.
(
i
+
l
)aX=X
s
(
i+1
)-Xs
(
i
)
I
f
l
a
J
l
~EI; p (
f
u
n
c
t
i
o
nn
o
tc
h
a
n
g
i
n
g
)
T
I
faX aX~ E
2
; s
t
o
p(
d
e
s
i
g
nn
o
tc
h
a
n
g
i
n
g
)
[
f
i+1=N; s
t
o
p

I
fVf
(X;
+
l
l
V
f
(
Xi
+l
)~ E
3
; c
o
n
v
e
r
g
e
d

ii+1

I~

1
.8659

2
.
6
9
0
1
2
.
7
4
5
5
2
.
7
5
8
5

1
.
8
5
0
2
1
.8960

3
.
0
2
9
8
3
.
0
2
3
3
3
.
0
1
8
1

0
.
0
5
0
7
0
.
0
4
6
2
0
.
0
3
1
6

jFigure6.4rep sentsthegraphicalrnotionofthedesignvariablesfor20itErations.

t
e
e
p
e
s
t
Theg
r
a
p
h
i
c
a
lp
i
c
t
u
r
eandt
h
ec
h
a
n
g
e
si
nt
h
ed
e
s
i
g
na
r
En
o
ti
m
P
E
E
S
S
i
V
ETheS
D
e
s
c
e
n
tmEthodi
SW
O
E
f
u
l
l
yMadequatECompazedt
oP
o
w
e
l
l
smethode
v
e
ni
f
t
h
el
a
t
t
e
r
i
saz
e
r
o
o
r
d
e
rme
t
l
1
o
d
.MO[1v
e
r
i
sconclusioni
sdrawnw
i
t
hr

p
e
c
tl
oa
n s
y
p
r
o
b
l
e
m(
q
u
a
d
r
a
t
i
c
)Example6
.1.i
sp
e
r
f
o
r
m
a
n
c
ej
u
s
t
i
f
i
e
st
l
1el
a
c
ko
fs
e
r
i
o
u
s
i
n
t
e
r
e
s
ti
nt
h
eS
t
e
c
p
e
s
tD
c
s
c
e
n
tm
e
t
h
o
d
.

4
.
5

GoToStep2
SteepestDescen
t
.m: Thisi
sa
nm
f
i
l
ea
1e
x
e
c
u
t
a
1g
o
r
i
t
h
mA
6
.
5
.I
tu
s
e
se
g
o
l
d
e
ns
e
c
t
i
o
nandt
l
1eu
p
p
e
rbounds
c
a
n
n
i
n
gPI1e
s
s
.l
eareso
f
t
h
i
sprograma
r
e
a
sf
o
l
l
o
w
s
.
J
Id
rawt
l
1ec
o
n
t
o
u
rp
l
o
.
t
Fortwov
a
r
i
a
b
l
e
si
twi

Fortwov
a
r
i
a
b
l
e
st
h
ed
e
s
i
g
nv
e
c
t
o
rc
h
a
n
g
e
sc
a
nb
es
e
e
ng
r
a
p
h
i
c
a
l
l
yi
ns
l
o
w
l
1s
t
e
p
si
nd
i
f
f
e
r
e
n
tc
o
l
o
r
.
m
o
t
i
o
nwit

ef
u
n
c
t
i
o
nva
1ueandeS
q
U
a
r
o
ft
l
1
el
e
n
gt
l
1of
Thed
e
s
i
g
nv
a
r
i
a
b
l
e
s
g
r
a
d
i
e
n
tv
e
c
t
o
r
(
c
a
11
e
dK
T
v
a
l
u
e
)a
te
a
c
hi
t
e
r
a
t
i
o
na
r
e
d
i
s
p
l
a
y
e
di
nt
l
1eCommand
l
1
enumbero
fi
t
e
r
a
t
i
o
n
s
.
windowa
tc
o
m
p
l
e
t
i
o
no
ft
l
1ef
u
n
c
t
i
o
ni
sn
u
m
e
r
i
c
a
l
l
ycomput
u
s
i
n
gf
i
r
s
tf
o
r
w
a
r
df
is
It
e
Theg
r
a
d
i
e
n
1o
ft
d
i
f
f
e
r
e
n
c
e
.Theg
r
a
d
i
c
n
tc
o
m
p
u
t
a
t
i
o
ni
st
l
1e
r
e
f
o
a
u
t
o
m
a
t
i
c
.

b
a
s
i
cs
t
r
i
n
gh
a
n
d
l
i
n
gt
oc
h
a
n
g
el
i
n
ec
o
l
o
ri
s
Fromaprogrammingp
e
r
s
p
e
c
t
i
v
e
i
n
t
r
o
c
l
u
c
e
d
.
0
.
5

Usage: SteepestDescent('Example6_1'(0.5 0.5]


20
0.0001
0
1
20)
(
6
.
1
0
)

g
u
r
e6
.
4

S
l
e
p
e
s
tdesn
:
tExa
m
p
l
e6
.
1
.

244 N
U
M
E
A
I
C
L
T
E
C
H
N
I
Q
U
E
SF
O
AU
N
C
O
N
S
T
A
A
I
N
E
DOPT1M
I
Z
A
T
l
ON

?
1
2
:
:
J
I
:
2
4
1
2
E
Z

rJ
x
p
e
c
t
e
dt
oovercom
t
h
i
sp
a
t
t
e
m
.I
nt
h
ec
a
s
eo
ft
h
eU
n
i
v
a
r
i
a
t
em odt
h
i
swas
a
c
h
E
e
y
e
dt
h
r
o
u
g
hp
a
t
F
r
nS
e
a
r
c
hmethodi
nt
h
ez
e
r
o
o
r
d
e
rfamily.Ani
t
e
r
a
t
i
o
n
ft
h
ez
i
g
z
a
gp
a m (
o
rp
r
e
v
e
n
t
i
n
gg
e
t
t
i
n
gl
o
c
k
e
di
n
t
oo
I
n
E
Pi
5
b
r
e
a
k
i 0o

:
::rro

:
l
i
z
:

:
:

2
:

e
me
o
d S

uc

6
.
3
.
2 ConJugateGradient(
F
l
e
t
c
h
e
r
R
e
e
v
e
s
)Method

6
.
3N
U
M
E
R
I
C
A
LT
E
C
H
NQUEs-GR
A
D
'
E
N
T
.
B
A
S
E
DMETHODS 245

s
. and th succes5ive adjustment of search direction
c
o
m
p
u
t
a
t
i
o
n
i
n
c
o
r
p
o
r
a
t
i
n
gt
h
i
sv
a
l
u
e
.~ r
e
p
r
e
s
e
n
t
st
h
er
a
t
i
oo
ft
h
es
q
u
a
r
eo
ft
h
ec
u
r
r
e
n
tg
r
a
d
i
e
n
t
t
v
e
c
t
o
rt
ot
h
es
q
u
a
r
eo
ft
h
ep
r
e
v
i
o
u
sg
r
a
d
i
e
n
tv
e
c
t
o
r
.Thef
i
r
s
tt
h
i
n
gt
on
O
l
i
c
ei
sa
ad
e
g
r
e
eo
fr
o
b
u
s
t
n
e
s
si
sb
u
i
l
ti
n
t
ot
h
emethodbyc
a
r
r
y
i
n
gi
n
f
o
r
m
a
t
i
o
nfromt
h
e
ikem
a
i
n
t
a
i
n
i
n
gah
i
s
t
o
r
yo
ft
h
emethoda
l
b
e
i
tf
o
rj
u
s
t
p
r
e
v
i
o
u
si
t
e
r
a
t
i
o
n
.T
h
i
si
sI
onei
t
e
r
a
t
i
o
n
.S
i
n
c
et
h
eFOCi
sb
a
s
e
dont
h
el
e
n
g
t
ho
ft
h
eg
r
a
d
i
e
n
ta
p
p
r
o
a
c
h
i
n
g
sP t
i
c
u
l
a
rfonno
fi
n
c
o
r
p
o
r
a
t
i
o
no
f~ i
si
n
g
e
n
i
o
u
s
.I
ft
h
e
z
e
r
oa
tt
h
es
o
l
u
t
i
o
n
. i
p
r
e
v
i
o
u
si
l
e
r
a
t
i
o
ni
sc
l
o
s
et
ot
h
es
o
l
u
t
i
o
n
.t
h
e
ns
i
sl
a
r
g
eand p
r
e
v
i
o
u
si
t
e
r
a
t
i
o
n
p
l
a
y
sas
i
g
n
i
f
i
c
a
n
tr
o
l
ei
nt
h
ec
u
r
r
e
n
ti
t
e
r
a
t
i
o
n
.Ont
h
eo
t
h
e
rhandi
fsi
sl
a
r
g
e
.
s
u
g
g
e
s
t
i
n
gt
h
es
o
l
u
t
i
o
ni
ss
t
i
l
lf
a
rawayt
h
e
nt
h
ec
u
r
r
e
n
tv
a
l
u
eo
ft
h
eg
r
a
d
i
e
n
t
d
e
t
e
r
m
i
n
e
st
h
enews
e
a
r
c
hd
i
r
e
c
t
i
o
n
.

i
e

;
i

?
:

5
;

2
5
g
:
;
:
2

;
q
i

:
?

:
1
i

i
5
;

TheCou
g
a
t
eG
r
a
d
i
e
n
tmethod
o
r
i
g
i
n
a
l
I
ydue F
1e
t
c
h
e
ra
n
dR
e
e
v
e
s[
5
]
i
sasma
Il

c
onv

e
r
g

en
c
b
e
c
a
us
et
h
es
e
a
r
c
hdl
I
r
.
e
c
t
u
1
o3
n
Esa
M
r
c

0
1

I
j
j
ugate
wi
t

h s

pt
t

ot
h
eH
e
s
s
i
a
n

:
?
xi
:
E

:
:

o
n
.Aq
u

l
a

A'
o
r
/
l
m:C
onJuga
G
r
a
d
f
j
t
a
S
t
e
p1
. ChooseX" N (
n
u
m
b
e
ro
fi
t
e
r
a
t
i
o
n

1
0
.
5
1
.X1=1
S
t
e
p1

o
:
s
l
;

!
s
(
l
)=j(X
)
;X.(
l
)=X
(
s
t
o
r
ev
a
l
u
e
s
)

E
2

E
3
:(
t
o
l
e
r
a
n
c
ef
o
rs
t
o
p
p
i
n
gc
r
i
t
e
r
i
a
)
E
S
e
li
=I(
in
i
l
i
a
l
i
z
ei
t
e
r
a
t
i
o
nc
o
u
n
t
e
r
)
fi=I

Sj=-Vj(X
)
S
t
e
p
2
.I
j

t=2-1.64591
S
t
e
p2
:s
=0
.
5
3
5
1;)=IO~36571

Si=-Vj(
)+s
S
j
_

11
.91351
1
3
1
2= O:83831;i
=1
.
2
5
4
4
;f
(
)=
3
.
0
; X3=121

X
'
+
I=X/+t

X
.
(
i+1
)
X
;
+

;
(i+1)=j
(X
)% (
s
V s
)
.!
J
.
j
=
f
s(
i+l)1); X=X.
(i+1
X
.
(
S 3
I
f
l
!
J
.
j
lsE
I

;
(

u
m

J
n

n
E

t
i

i
o
nn

fVj(X

):
$E
3
; c
o
n
v
e
r
g
d
I
f
V
j
(
X
i
+
i+1

i~

GoToS
t
e
p2

Con)ugateGradlent.m: Thi
sm
f
i
l
ew
i
l
le
x
e
c
u
t
et
h
eC
o
n
j
u
g
a
t
eG
r
a
d
i
e
n
tm
e
t
h
o
d
.
Fortwo a
r
i
a
b
l
ep
r
o
b
l
e
m
st
h
e
r
ew
i
l
lb
eac
o
n
t
o
u
rp
l
o
to
v
e
rw
h
i
c
hei
t
e
r
a
t
i
v
e
p
r
o
g
r
e
s
si
nt
h
ed
e
s
i
s
p
a
c
ei
sr
e
c
o
r
d
e
d
.Thep
r
o
p
e
r
t
yo
fqua
8
t
i
Cc
o
n
v
e
r
g
e
n
c
ec
a
n
b
eo
b
s
e
r
v
e
d t
h
es
o
l
u
t
i
o
ni
so
b
t
a
i
n
e
da
f
t
e
rtwoi
t
e
r
a
t
i
o
n
s
. Th
i
si
sd
e
f
i
n
i
t
e
l
y
i
m
p
r
e
s
s
i
v
ec
o
m
p
a
r
e
dw
i
t
h S
t
p
e
s
tD
e
s
c
e
n
tm o
d
.K pi
nminda
tt
h
e
Ig
o
r
i
t
h
mi
sm
i
n
o
r
.Asev
a
r
i
o
u
sc
o
d
e
sa
r
er
u
ni
no
r
d
e
r
.some
m
o
d
i
f
i
c
a
t
i
o
nt
ot
h
ea
c
h
a
n
g
e
si
nt
h
ef
i
g
u
r
ec
a
nb
eo
b
s
e
r
v
e
d
.The
s
ea
r
el
a
r
g
l
yduet
ot
h
eu
s
eo
fsomes
t
r
i
n
g
ya
r
ei
m
p
l
e
m
e
n
t
e
d
.The
p
r
o
c
e
s
s
i
n
gf
u
n
c
t
i
o
n
s
.Youa
r
ee
n
c
o
u
r
a
g
e
dt
os
e
ehowe
n
g
syouwoulddod
i
f
f
e
r
e
n
t
l
y
.Don
o
th
e
s
i
t
a
t
t
oy
a
s
a
u
t
h
o
ri
sc
e
r
t
a
i
nt
h
e
r
ea
r
ei
i
st
h
eon
1ywayt
ol
e
a
r
nt
ou
s
eMATLABe
f
f

c
t
i
v
e
l
y
.

f!
J
.X
'!
J
.
X:$E2
s
t

o
p(
d
e
s
i
g
nn
o
tc
h
a
n
g
i
n
g
)
Ifi+I=N; 5
t

o
p

f
(
X

)=5
.
3
1
2
5

10
.
5 1 _._
_
_ .
.
.
. ._
_
_
. _
_ 1
0
.
5
9
9
6
1
S
t
e
p2
.S =12~i51;i =0
.
1
9
9
3
;f
(

i
)=4
.
7
8
3
1
; X2=
1O
:
9
4
8
4
1

j
Elses= '(xjfVj(X
)
Vj
(X
j
_
l
)TVj
(Xi
_
'
)

Ii
sd
e
t
e
r
m
i
n
e
dbym
i
n
i
m
i
z
i
n
g
j
(
Xi
+

'
0

A
p
p
l
l
c
a
t
l
o
n ConJugateGradlentMethod:Thi
s method i
si
m
p
o
r
t
a
n
tt
o
e
s
ec
a
I
c
u
l
a
t
i
o
n
s
s
i
m
i
l
a
rt
oP
o
w
e
l
l
'
s
w
a
r
r
a
n
tw
o
r
k
i
n
gt
h
r
o
u
g
ht
h
ei
t
e
r
a
t
i
o
n
s
.Th
h
o
d
c
a
nb
edoneu
s
i
n
gahandc
a
Ic
u
l
a
t
o
r
.u
s
i
n
gas
p
a
d
s
h
t
o
rusingMATLAB
mt
i
t
s
e
lf
.Thea
u
t
h
o
rrecommendst
h
ec
a
Ic
u
l
a
t
o
ra
sep
r
o
c
e
s
si
ss
l
o
we
n
o
u
g
ht
od
e
v
e
l
o
p
r
s
.I
nt
h
ef
o
l
l
o
w
i
n
ga
n
du
n
l
i
k
e
a
ni
n
t
u
i
t
i
o
nw
o
r
k
i
n
gw
i
t
ht
h
ep
a
l
t
e
mo
fnumb
P
o
w
e
l
l
'
smt
h
o
d
o
n
1
yt
h
er
e
l
e
v
a
n
tc
a
I
c
u
1
a
t
i
o
n
sa
r
er
e
c
o
r
d
e
d
.Ther
e
a
d
e
rs
h
o
u
l
d
f
o
l
l
o
wa
Ioogv
e
r
i
f
y
i
n
gt
h
enumbersa
r
ei
o
d
e
e
dc
o
r
r
e
ct
.

;
:
;

;::P22Jt::2JUlz;ittrtt;123tzr:rn;

Example6_1'[0.5 0.5]
U ge:ConjugateGradient(
20
0.0001
0
1
20);

(
6.
11
)

246

N
U
M
E
R
I
C
A
LT
E
C
H
N
I
Q
U
E
SFO
U
N
C
O
N
S
T
R
A
I
N
E
DO
P
T
I
MA
T
I
O
N

6
.
3N
U
M
E
R
I
C
A
LT
E
C
H
N
I
Q
U
E
-GR
A
D
I
E
N
T

BASEDMETHODS 247

Output(Commandwindow)
Thep
r
o
b
l
e
m
:Example6_1
Thed
e
s
i
g
nv
e
c
t
o
r

f
u
n
c
t
i
o
nv
a
l
u
ea
n
dKTv
a
l
u
ed
u
r
i
n
gt
h
ei
t
e
r
a
t
i
o
n
s
0.
5000
5
996
0.
2
.
9
9
9
5

5000
0.
0
.
9
4
8
3
1
.
9
9
8
2

3125
5.
4
.
7
8
3
2
3
.
0
0
0
0

t
h
i
sf
a
r
l
yo
fmethodsgobeyondt
h
a
t
.Ast
h
es
o
l
u
t
i
o
ni
sa
p
p
r
o
a
c
h
e
d

t
h
e
ybehavel
i
k
e
4
)
.Theq
u
a
d
r
a
t
i
cc
o
n
v
e
r
g
e
n
c
eo
fNewton'smethodi
s
Newton'smethod(
S
e
c
t
i
o
n6.
q
u
i
l
ei
m
p
r
e
s
s
i
v
e
i
tw
i
l
ll
o
c
a
t
e s
o
l
u
t
i
o
nt
oaq
u
a
d
r
a
t
i
cproblemi
no
l
l
ei
t
e
r
a
t
i
o
n
.
i
I
keb
e
h
a
v
i
o
ro
ft
h
eVMM
a
st
h
es
o
l
u
t
i
o
ni
sa
p
p
r
o
a
c
h
e
d

h
a
sg
i
v
e
nthem
T
h
i
sNewtont
e
r
m
e
dQ
u
a
s
i
N
e
w
t
o
l
lo
rN
e
w
t
o
n
l
i
k
em
e
t
h
o
d
s
.
anothernam
TheDFPmethodi
sp
r
e
s
e
n
t
e
dh
e
r
ef
o
rh
i
s
t
o
r
i
c
a
li
n
t
e
r
e
s
t db
e
c
a
u
s
ei
ti
sal
i
t
t
l
e
t
h
e
r
s
.I
nf
a
c
tHuang(
7
)documentsag
e
n
e
r
i
cp
r
o
c
e
d
u
r
e
e
a
s
i
e
rt
ou
n
d
e
r
s
t
a
n
dt
h
a
n o
o
p
u
l
a
rmethodsc
a
nbeo
b
t
a
i
n
e
d

andfromwhichyouc
o
u
l
d
fromwhichmosto
fep
a
l
s
od
e
r
i
v
ey
o
u
rown.
Th
eCo
n
j
u
g
a
t
eG
r
a
d
i
e
n
tm
e
t
h
o
d
'si
r
n
p
r
o
v
e
m
e
n
to
v
e
rt
h
eS
t
e
e
p
e
s
lD
e
s
c
e
n
tmethodwas
h
ep
r
e
v
i
o
u
si
t
e
r
a
t
i
o
n
.l
nt
h
e
p
o
s
s
i
b
l
eb
e
c
a
u
s
eo
ft
h
ei
n
c
l
u
s
i
o
no
ft
h
eh
i
s
t
o
r
ym t
q
u
a
s
i
N
e
w
t
o
nm
e
t
h
o
d
st
h
h
i
s
t
o
r
yfrom a
I
lp
r
e
v
i
o
u
si
t
e
r
a
t
i
o
n
si
sa
v
a
i
l
a
b
l
e
.T
h
i
s
oni
sc
o
l
l
e
c
t
e
di
na
nnXnm i
xc
a
l
l
e
dem
e
t
r
i
c
.Them
e
t
r
i
ci
su
p
d
a
t
e
dwi

i
n
f
o
r
m
e
a
c
bi
t
e
r
a
t
i
o
n di
su
s
e
dt
oe
s
t
a
b
l
i
s
ht
h
es
hd
i
r
e
c
t
i
o
n
.Ani
n
i
t
i
a
lc
h
o
i
c
ef
o
rt
h
em
e
t
r
i
c
.I
tmustbea t
e
t
'
cp
o
s
i
t
i
v
ed
e
f
t
n
i
t
em
a
t
r
i
x
.F
o
rt
h
em odt
o
i
sa
l
s
o u d
t
h
emet
l
Icm
usth
o
l
dont
oi
t
sp
o
s
i
t
i
v
ed
e
f
i
n
i
t
ep
r
o
p
e
r
t
yt
h
r
o
u
g
ht
h
ei
t
e
r
a
t
i
o
n
s
.
c
o
n
v
e
r
g
e
heDFPmethodt
h
em
e
t
r
i
ca
p
p
r
o
a
c
h
e
s i
n
v
e
r
s
eo
ft
h
eH
e
s
s
i
a
na
tU
l
es
o
l
u
t
i
o
n
.

0
.
2
4
9
0
2
.
8
4
1
1
0
.
0
0
0
1

Keepi
nmindt
h
es
o
l
u
t
i
o
ni
sd
e
p
e
n
d
e
n
tont
h
et
o
l
e
r
a
n
c
ef
o
rt
h
eo
n
e
d
i
m
e
n
s
i
o
n
a
I
s
t
e
p
s
i
z
es
e
a
r
c
ha
n
dt
h
a
tt
h
ed
e
r
i
v
a
t
i
v
e
sa
r
ecomputedn
u
r
n
e
r
i
c
a
l
l
y
.Thee
x
e
c
u
t
i
o
no
f
ec
odes
h
o
u
l
dd
i
s
p
l
a
yF
i
g
u
r
e6
.
5i
nt
h
ef
i
g
u
r
ewindow.
6
.
3
.
3 Davldon-Fletcher-PowellMethod
a
m
i
l
yo
fV
a
r
i
a
b
l
e
TheD
a
v
i
d
o
n
F
l
e
t
c
h
e
r
P
o
w
e
l
l(DFP)[
6
]methodb
e
l
o
n
g
st
oef
M
e
t
r
i
cMethods(VMM).Itwasn
r
s
ti
n
t
r
o
d
u
c
e
dbyDavidonands
e
V
E
r
a
lyEMBl
a
t
e
r
le
t
c
h
e
r dPowel
I
. Bi
n
gawareo
ft
h
e
wasd
e
v
e
l
o
p
e
di
ni
t
sc
u
r
r
e
n
tf
o
r
mbyF
C
o
n
j
u
g
a
t
eG
r
a
d
i
e
n
tmethod
t
h
e
s
emethodswouldn
o
tme
r
I
ti
n
c
Iu
s
i
o
ni
nt
h
i
sbooki
f
t
h
e
yd
i
dn
o
th
a
v
et
h
ep
r
o
p
e
r
t
yo
fqu
.
a
d
r
a
t
i
cc
o
n
v
e
r
g
e
n
c
ewhicht
h
e
yd
o
.G
e
n
e
r
a
l
l
y

Aorithm:Davidon
F '
c
h
e
r
P
o
w
e
l
l(
A6
.

S
t
e
p1
.Choos

x
.
.[A1l(initialmeic}N

Ehe
2
'e
J
:(
t
o
l
e
r
a
n
c
ef
o
rs
t
o
p
p
i
n
gc
r
i i
a
)
S
e
ti=1(
in
i
t
i
a
l
i
z
i
t
e
r
a
t
i
o
nc
o
u
n
t
e
r
)

[Aj]Vf(Xj
)
S
t
e
p2
.8=

= Xi
+
+1
s
u AX=jS
Xi

ii
sd
e
t
e
r
m
i
n
e
dbym
i
n
i
m
i
z
i
n
g
f
(
X
i
+
l
)
S
t
e
p3
.I
fV
f
(
X
i
+
I
)
TVf(X
+1
3
; c
o
n
v
e
r
g
e
d
)e
i

I
ff
!
(X
+l
(
X
j
)
1:
;e
l
; s
t
o
p (
f
u
n
c
t
i
o
nn
o
tc
h
a
n
g
i
n
g
)
)f
i
T
I
fAX AX~e2 stop (
d
e
s
i
g
nn
o
tc
h
a
n
g
i
n
g
)
t
o
p (
i
t
e
r
a
t
i
o
ni
I
m
i
l
)
I
fi+1=N s

E
l
s

j
Y=Vf(Xj
+
l
)-Vf(X
}

z=[A;]Y
T
AXaX
[
B
]
=
:
:
:
:
:
;
:
:
:

AX'Y

ZZT

[q= --~-

V'Z

3
.
5

F
l
g
u
r
e6
.
5C
o
n
i
u
g
a
t
eg
r
a
d
l
e
n
tm
e
t
h
o
d
:E
x
a
m
p
l
e6
.
1
.

4
.
5

[Ai
+
J
l=[A]
+[
B
]+[
q

i i+1
GoToS
t
e
p2

I
nt
h
ea
b
o
v
e em
a
t
r
i
c
e
sa
r
ee
n
c
l
o
s
e
db
ys
q
u
a
r
eb
r
a
c
k
e
t
s
.Thei
n
i
t
i
a
lc
h
o
i
c
eo
ft
h
e
m
e
t
r
i
ci
sap
o
s
i
t
i
v
ed
e
f
i
n
i
t
em
a
t
r
i
x
.Thei
d
e
n
t
i
t
ym
a
t
r
i
xi
sas
a
f
ec
h
o
i
c
e
.
sm
f
i
l
e
.S
i
m
i
l
a
r1
0t
h
ep
r
e
v
i
o
u
s
DFP.m: TheDFPa
l
g
o
r
i
t
h
m(
A
6
.
7
)i
sc
o
d
e
di
ni
p
r
o
g
r
a
m
sf
o
rtwov
a
r
i
a
b
l
e
sed
e
s
i
g
nc
h
a
n
g
e
sa
r
et
r
a
c
k
e
do
nab
a
c
k
g
r
o
u
n
dc
o
n
t
o
u
r
t F
i
g
u
r
e6
.
6.Th
ed
e
f
a
u
l
ti
n
i
t
i
a
lm
e
t
r
i
ci
sI
h
ei
d
e
n
t
i
t
ym
a
t
r
i
xg
e
n
e
r
a
t
e
du
s
i
n
ga
p
l
o
MATLABb
u
i
l
t
i
nf
u
n
c
t
i
o
n
.
U
s
a
g
e
:DFP('Example6_1'[0.5 0.51
4
0.0001
0
1
20); (
6.
12
)
O
u
t
p
u
t
: The o
u
t
p
u
tf
r
o
mt
h
ea
b
o
v
ei
n
v
o
c
a
t
i
o
n(
c
o
p
i
e
df
r
o
mt
h
e Command
.
tThev
a
r
i
a
b
l
e
window)c
a
nb
eo
b
t
a
i
n
e
db
yr
e
m
o
v
i
n
gt
h
cs
c
m
i
c
o
l
o
na
f
t
e
rt
h
es
t
a
t
e
m
e
n
d
e
f
i
n
i
t
i
o
n
sm
a
t
c
ht
h
o
s
ed
e
f
i
n
e
di
nt
h
ea
l
g
o
r
i
t
h
m
.F
o
rExample6
.
1

iteration number: 1
s
0.4990 2.2467
delx =
0.0996
0.4483
y

-1.1458
2.6153

z
-1.1458
2.6153
B

0.0094 0.0422
0.0422 0.1899

nU

01

0.1610 0.3676
0.3676 -0.8390

A
0.8483 0.4097
0.4097 0.3509
iteration number: 2
s
1.2443 0.5446
The problem: Example6_1
The design vectorfunction value and KT value
during the ite ations
0.5000 0.5000
5.3125 5.2969
0.5996 0.9483
4.7832 2.8411
2.9996 1.9987
3.0000 0.0000

6
.
3
.
4 Broydon-Fletcher-Goldfart
S
hannoMethod

F
J
g
u
r
e6
.
6 DFPM
e
l
h
o
d
:E
x
a
m
p
l
e6
.
1
.

4
.
5

I
fyou weret
oprogramo
n
l
yoneg
r
a
d
i
e
n
t
b
a
s
c
dmt
h
o
d
.t
h
e
nt
h
cBroydonFletcher-Goldfarb-Shanno(BFGS)181method would be t
h
e one.It i
sa
V
a
r
i
a
b
l
eM
c
t
r
i
c
q
u
a
s
i
N
c
w
t
o
nmethodandc
u
r
r
e
n
t
l
yi
st
h
emostp
o
p
u
l
a
ro
ft
h
te
n
j
o
y
sl
h
ep
r
o
p
c
r
l
yo
fq
u
u
d
r
a
t
i
cc
o
n
v
c
r
g
e
n
c
candh
a
sr
o
b
u
s
t
n
e
s
sb
y
m
e
t
h
o
d
s
.I
c
a
r
r
y
i
n
gf
o
r
w
a
r
di
n
f
o
r
m
a
t
i
o
nfromI
1
3
Ep
r
e
v
i
o
u
si
t
e
r
a
t
i
o
n
s
.
T
h
ed
i
f
f
e
r
e
n
c
E
b
E
t
w
e
e
n
DFPa
ndBFGSi
st
h
ewayt
h
em
e
l
r
i
ci
su
p
d
a
l
e
d
.Thef
o
r
m
e
rc
o
n
v
e
r
g
e
st
ot
h
e
i
n
v
e
r
s
eo
ft
h
eH
e
s
s
i
a
n
.w
h
i
l
et
h
el
a
t
t
e
rc
o
n
v
e
r
g
e
st
ot
h
eH
e
s
s
i
a
ni
t
s
e
l
f
.I
nas
e
n
s
e
eBFGS i
s morc d
i
r
e
ct
. The BFGS h
a
sr
e
p
l
a
c
e
dt
h
eC
o
n
j
u
g
a
t
eG
r
a
d
i
e
n
l
t
e
c
h
n
i
q
u
e
sa
saw
o
r
k
h
o
r
s
ei
ns
o
l
v
i
n
gn
o
n
l
i
n
e
a
re
q
u
a
t
i
o
n
s
.

Hjil--J41JHAHill-J

6
.
3N
U
M
E
R
I
C
A
LT
E
C
H
N
I
Q
U
E3
R
A
D
I
E
N
T
.
B
A
S
E
DMETHODS 249

248 N
U
M
E
R
I
C
A
LT
E
C
H
N
I
O
U
E
SF
O
RU
N
C
O
N
S
T
R
A
I
N
E
DO
P
T
I
M
I
Z
A
T
I
O
N

250 N
U
M
E
R
I
C
A
LT
E
C
H
N
I
Q
U
E
SFORU
N
C
O
N
S
T
R
A
I
N
E
DOP
T
lM
I
Z
AT
lO
N

6.
4

RU

Forc
o
n
v
e
r
g
e
n
c
et
h
em
e
t
r
i
cmustb
ep
o
s
i
t
i
v
ed
e
f
i
n
i
t
e
.Ani
n
i
t
i
a
lc
h
o
i
c
eo
f
p
o
s
i
t
i
v
ed
e
f
i
nI
l
m
a
t
r
i
xf
o
rt
h
em
e
t
r
i
ci
su
s
u
al
Jys
u
f
f
i
c
i
e
n
l
l
oe
n
s
u
r
et
h
i
sp
r
o
p
e
r
t
y
f
o
rq
u
a
d
r
a
t
i
cp
r
o
b
l
e
m
s
.Thei
d
e
n
t
i
t
ym
a
t
r
i
xi
su
s
u
a
l
l
yad
e
f
a
u
l
tc
h
o
i
c
e
.

N
U
M
E
R
I
C
A
LT
E
C
H
N
I
Q
U
Es-5ECONDORDER 2
5
1

A
l
g
o
r
i
t
h
m
:Bro
F er-Goldfarb-Shanno(BFGS)Method(
A
6
.
S
t
e
p1
. ChoeX"[
A
t
l(
in
i
t
i
a
lm
e
t
r
i
c
)
N
E"E
2
E
3
:(
t
o
l
e
r
a
n
c
ef
o
rs
t
o
p
p
i
n
gc
r
i
t
e
r
i
a
)
S
c
ti=1(
i
n
i
t
i
a
l
i
z
ei
t
e
r
a
t
i
o
nc
o
u
n
t
e
r
)
S
t
e
p2
. Thes
e
a
r
c
hd
i
r
e
c
t
i
o
ni
so
b
t
a
i
n
e
da
sas
o
l
u
t
i
o
nt
o
[
A
;
]
S

=-Vf(X)

l
CS;
dJ(=
S

x=~+

ii
sd
e
t
e
n
n
i
n
e
dbyMin
z
i
n
g
f
(
X
)
S
t
e
p3
.I
fV
f(X
)
TV
f(X
) E
3
converged

cl=2m-3 21
3
.
0
4
2
9 6
.
5
0
9
5

I
fr
I
(Xi
+l
)f
(X)
1~ E
Isp (
f
u
n
c
t
i
o
nn
o
tc
h
u
n
g
i
n
g
)
T

I
f!
1
X dJ(~ E
2
; s
t
o
p (
d
e
s
i
g
nn
o
tc
h
a
n
g
i
n
g
)
I
fi+1=Ns
t
o
p(
i
t
e
r
a
t
i
o
nl
i
m
i
t
)
Y= V
f(Xi
+l)- V
f(X
)

Thet
r
a
n
s
l
a
t
i
o
no
ft
h
eBFGSa
l
g
o
r
i
l
h
mi
n
t
ocodci
sl
e
f
ta
sa
ne
x
e
r
c
i
s
ef
o
rt
h
e
.I
ti
srecommcndedt
h
a
tU
l
ec
a
l
c
u
l
a
t
i
o
no
ft
h
es
e
a
r
c
hd
i
r
e
c
t
i
o
ni
nS
t
c
p2be
s
t
u
d
e
nt
a
c
c
o
m
p
l
i
s
h
e
da
sas
o
l
u
t
i
o
nt
oas
e
to
fl
i
n
e
a
re
q
u
a
t
i
o
n
s(
a
sl
i
s
t
e
d
)r
a
U
l
e
rt
h
a
ni
n
v
e
r
t
i
n
g
t
h
em
a
t
r
i
xA
.Oneo
t
h
e
rt
h
i
n
gt
on
o
t
ef
o
rExample6
.
1i
st
h
a
ti
ft
h
ep
o
l
y
n
o
m
i
a
l
h
ec
a
l
c
u
l
a
t
i
o
no
f
a
p
p
r
o
x
i
m
a
t
i
o
nwasu
s
e
df
o
rt
h
eo
n
e
d
i
m
e
n
s
i
o
n
u
lp
r
o
c
e
s
st
wouldh
a
v
eb
e
c
ne
x
u
c
t(
w
h
y
?
)
.

yyT

[
B
)=

y
'dJ(

Vf(X
)
Vf(X
?
[C]=
V
f
(
X
;
)
'
S
j
[
A
j
+
d=
[
A
;
]+[
B
)+[
c
]
if
i
+1

SECONDORDER
6
.
4 NUMERICALTECHNIQUES-

GoToS
t
e
p2

'
0

A
p
p
l
i
c
t
l
o
n BFGSMethod:Themethodi
sa
p
p
l
i
c
dt
oExample6
.1
.O
nly
c
s
s
e
n
t
i
a
lc
o
m
p
u
t
a
t
i
o
n
sa
r
es
h
o
w
n
.
1
)=5
.
3
1

=
3
.
0
; )
=[g]

Converged

E
l
s
e

c
p
I XI=lsl;

S24221; =1 4
4
; X3=[H

=[~ ~l

; 1

Vf(X1)

WSl=lzl; =0.19 X2=ltzzl;Vf'~]


step3Y=12Jri;m=ltml

Sc
o
n
d
o
r
d
e
rm
e
t
h
o
d
sf
o
ru
n
c
o
n
s
t
r
a
i
n
e
do
p
t
i
m
i
z
a
t
i
o
na
r
en
o
tu
s
c
dbc
a
u
s

v
a
l
u
a
l
i
o
n
o
ft
h
eH
e
s
s
i
a
nm
a
t
r
i
xd
u
r
i
n
ge
a
c
hi
t
e
r
a
t
i
o
nt
h
eH
c
s
s
i
a
nm
a
t
r
i
xi
sc
o
n
s
i
d
e
r
e
dt
ob
e
r
o
p
e
r
t
yo
fq
u
a
d
r
a
t
i
c
c
o
m
p
u
t
a
t
i
o
n
a
l
l
ye
x
p
e
n
s
i
v
e
.A s
e
c
o
n
d
o
r
d
e
rmethodw
i
t
h ep
n
ei
t
e
r
a
t
i
o
n
.
c
o
n
v
e
r
g
e
n
c
ei
sv
e
r
yi
m
p
r
e
s
s
i
v
e
.Ann
v
a
r
i
a
b
l
ep
r
o
b
l
e
mc
a
nc
o
n
v
e
r
g
ci
no
o
rr
e
a
ld
e
s
i
g
np
r
o
b
l
e
m
s
whered
e
c
i
s
i
o
n
sa
r
er
e
q
u
i
r
e
dt
ob
e
Asm
e
n
t
i
o
n
e
db
e
f
o
r
ef
a
r
i
a
b
l
e
s e
x
i
s
t
e
n
c
eo
ft
h
ef
i
r
s
t
madeonU
l
ed
i
s
c
r
e
t
en
a
t
u
r
eo
fsomeo
f ev
o
tt
om
e
n
t
i
o
ns
e
c
o
n
dd
e
r
i
v
a
t
i
v
e
s i
sq
u
e
s
t
i
o
n
a
b
l
e
. Moreovcr U
l
e
d
e
r
i
v
a
t
i
v
e n
q
u
a
s
i
N
e
w
t
o
nm
e
t
h
o
d
so
f l
a
s
ttwos
e
c
t
i
o
n
sa
r
ea
b
l
et
oe
f
f
c
c
t
i
v
e
l
yf
u
n
c
t
i
o
na
s
s
e
c
o
n
d
o
r
d
e
rm
e
t
h
o
d
sa
se
ya
p
p
r
o
a
c
h solution-andt
h
e
yd
on
o
tn
e
e
dt
h
e
ls
e
s
t
i
m
a
t
i
o
no
fs
e
c
o
n
dd
e
r
i
v
a
t
i
v
e
s
.Ones
e
c
o
n
d
o
r
d
e
rmethodi
sp
r

e
n
t
e
dh
e
r
ef
o
rt
h
e
s
a
k
eo
fc
o
m
p
l
e
t
e
n
s
s
.
h
e
r
ei
so
n
l
yo
n
eb
a
s
i
cs
e
c
o
n
d
o
r
d
e
rt
e
c
h
n
i
q
u
ef
o
ru
n
c
o
n
s
t
r
a
i
n
e
d
I
n
d
e
e
dt
o
p
t
i
m
i
z
a
t
i
o
n
.I
ti
sb
a
s
e
dont
h
ee
x
t
e
n
s
i
o
no
ft
h
eNewton-Raphsont
e
c
h
n
i
q
u
et
o
m
u
l
t
i
v
a
r
i
a
b
l
ep
r
o
b
l
c
m
.Manyd
i
f
f
c
r
e
n
tc
x
t
e
n
s
i
o
n
sa
r
ea
v
a
i
l
a
b
l
eb
u
ti
nt
h
i
st
e
x
ta

e
n
e
r
a
la
l
g
o
r
i
t
h
m
i
cs
u
c
t
(
A
6
.
1
)i
sp
r
e
s
e
n
t
e
d
.Once
d
i
r
e
c
te
x
t
e
n
s
i
o
nr
e
c
a
s
ti
neg
l
1
o
w
s
.
a
g
a
i
nt
h
eg
e
n
e
r
a
lp
r
o
b
l
e
ma
n
ds
p
e
c
i
f
i
cexamplea
r
ea
sf
o

(
6
.1
)

M
i
n
i
m
i
z
e f
(
X
)
; [X]n
S
u
b
j
e
c
tt
o
:

x
lXixjt

i=1

2
.
.
.
1
1

(
6
.
2
)

M
i
n
i
r
n
i
z
e f
6
.
3
)
(
X
)=
f
(
X
I
'X
2
)=
3+(
X
I-1
.
5
x
2
)
2+(
X
2-2
)
2 (
S
u
b
j
e
c
tt
o
: O
:
:
;
;
X
I
:
5
;0
:
:
;
;
x
2
:
:
;
;
5

(
6.
4
)

&

-EEE E -E

r
I

1
)f
(
X
i
)
1:
;E
I
' s
t
o
p (
f
u
n
c
t
i
o
nn
o
tc
h
a
n
g
i
n
g
)
I
f (X
TsX:
IfAX
:
;
;
E
2
SlOp (designnotchanging)

i
z
l

S
t
e
p3
.I
fVf(X
)
TVf(Xi
+I
)E
3
' c
o
n
v
e
r
g
e
d

31
2
5
)=5.
Step1431; f(Xt
nU

(
6
.
1
4
)

INewton-Raphsoni
sn
o
tknown f
o
ri
t
sr
o
b
u
s
t
n
e
s
so
rs
t
a
b
i
l
i
t
ya
n
d
Theo
r
i
g
i
na
E
q
u
a
t
i
o
n(
6
.
1
4
)s
h
a
r
e
st
h
esamed
i
s
a
d
v
a
n
t
a
g
e
.Toc
o
n
t
r
o
lt
h
ed
e
s
i
g
nv
a
r
i
a
b
l
ec
h
a
n
g
e
s
andt
ob
r
i
n
gi
tu
n
d
e
reschem
o
fa
1
g
o
r
i
t
h
m(A6.
l
)t
h
el
e
f
t
h
a
n
ds
i
d
ei
sd
e
f
i
n
e
d
t
h
es
e
a
r
c
hd
i
r
e
c
t
i
o
nv
c
t
o
rS f
o
l
l
o
w
e
d by as
t
a
n
d
a
r
do
n
e
d
i
m
e
n
s
i
o
na
Is
t
e
p
s
i
z
e
o
d
i
f
i
e
dNewtonm
e
t
h
o
d
.Thec
o
m
p
l
e
t
ea
l
g
o
r
i
t
h
mi
s
c
o
m
p
u
t
a
t
i
o
n
.T
h
i
si
st
e
r
r
n
e
dt
h
em
a
sf
o
l
l
o
w
s
.

Xi =X

/Si;
AX=.

S
I
i+
)
i
"i
sd
e
t
e
r
r
n
i
n
e
dbyM
i
n
i
m
i
z
i
n
gJ
(
Xi
+l

ApplicationofModifiedNewtonMethod: A
l
g
o
r
i
t
h
m(
A
6
.
9
)i
su
s
e
dt
os
o
l
v
e
Example6
.
1
.Onlye
s
s
e
n
t
i
a
lc
o
m
p
u
t
a
t
i
o
n
sa
r
ei
n
c
l
u
d
e
d
:

0
5

wherefi
st
h
s
i
n
g
l
e
v
a
r
i
a
b
l
eo
b
j
e
c
l
i
v
ef
u
n
c
t
i
o
n
.Them
u
l
t
i
v
a
r
i
a
b
l
ee
x
t
e
n
s
i
o
nl
o
c
o
m
p
u
t
i
n
gas
i
m
i
l
a
rc
h
a
n
g
ei
nev
a
r
i
a
b
l
ev
e
c
l
o
ri
s

Algori
m
:ModlfledNewtonMe od{A6.9)
S
t
e
p1
. ChooseX
JN
E"E
2
'E
3
:(
t
o
l
e
r
a
n
c
ef
o
rs
t
o
p
p
i
n
gc
r
i
t
e
r
i
a
)
S
e
ti=1(
i
n
i
t
i
a
l
i
7
.
ei
t
e
r
a
t
i
o
nc
o
u
n
t
r
)
S
t
e
p2
.Thes
e
a
r
c
hd
i
r
e
c
t
i
o
ni
so
b
t
a
i
n
e
da
sas
o
l
u
t
i
o
nt
o
[
H
(
X
j
)
]
S
i=-Vf(X
i
seH
e
s
s
i
a
n
)
; [
1

i i
+1
GoToSt
p2

o
n
v
e
r
g
e
d
.
0
; c
s

=
1
~.;~~~
:
1
u1

=
:
=
1
~1

7'~~;;;;
;

=1
.0
;

2)=3
1
1
.5000
'
l
t
X?
1;
2
1 f(X
Uo

J.
v

"~2

Asexpectedo
n
l
yas
i
n
g
l
e
i
t
e
r
a
t
i
o
nwasnecessaryt
o
o
b
t
a
i
nt
h
e
s
o
l
u
t
i
o
n
.T
r
a
n
s
l
a
t
i
o
n
o
f
t
h
e
a
l
g
o
r
i
t
h
mintoMATLABcodei
soncea
g
a
i
nl
e
f
ta
sane
x
e
r
c
i
s
e
.

f
"(
x
)

X= [ '~f(X)

E
l
s
e

'
(
X
)

253

i
t
e
r
a
t
i
o
ni
Im
i
t
)
t
o
p (
I
fi+1= N
s

QU

d
x
=L
=

(
6
.
1
3
)

6
.
5A
D
D
I
T
I
O
N
A
LE
X
A
M
P
L
E
S

ModifledNewton'sMethod:TheNewton-Raphsonmethodu
s
e
df
o
ras
i
n
g
l
e
v
a
r
i
a
b
l
es
o
l
v
e
st
h
ep
r
o
b
l
e
m(
X
)=O
.Thise
q
u
a
t
i
o
nc
a
n p
r
e
s
e
n
teFOCf
o
ra
n
u
n
c
o
n
s
t
r
a
i
n
e
dp
r
o
b
l
e
mi
nonev
a
r
i
a
b
l
e(
C
h
a
p
t
e
r5
)
.Thei
t
e
r
a
t
i
v
ec
h
a
n
g
ei
ne
v
a
r
i
a
b
l
ei
scomputedt
h
r
o
u
g
h

i
iftJ
3 33
3341131411ffits?hititi--1
3

U
M
E
R
I
C
A
LT
E
C
H
N
I
Q
U
E
SF
O
U
N
C
O
N
S
T
R
A
I
N
E
OO
P
T
I
M
l
Z
A
T
I
O
N
252 N

6
.
5 ADDITIONALEXAMPLES
ss
e
c
t
i
o
n
.Withmanyo
ft
h
en
u
m
e
r
i
c
a
l
T
h
r
e
ea
d
d
i
t
i
o
n
a
le
x
a
m
p
l
e
sa
r
ep
r
e
s
e
n
t
e
di
n i
b
t
a
i
n
i
n
gt
h
es
o
l
u
t
i
o
ni
sam
a
t
t
e
ro
fj
u
d
i
c
i
o
u
s
t
e
c
h
n
i
q
u
e
sa
l
r
e
a
d
y programmedo
a
p
p
l
i
c
a
t
i
o
no
ft
h
en
u
m
e
r
i
ca
Ip
r
o
c
e
d
u
r
e
s
.T
h
i
ss
e
c
t
i
o
na
l
s
op
r
e
s
e
n
t
ssomec
r
e
a
t
i
v
eu
s
e
o
ft
h
e methods a
n
de
x
p
l
o
r
e
sn
o
n
q
u
a
d
r
a
t
i
cp
r
o
b
l
e
m
s
.l
ef
r
r
s
l examplei
s e
R
o
s
e
n
b
r
o
c
kproblem[
9
]
.lIsexamplewasamongt
h
o
s
ec
r
e
a
t
e
dt
oc
h
a
1
1
e
n
g
e
n
u
m
e
r
i
c
a
lt
e
c
h
n
i
q
u
e
sf
o
ru
n
c
o
n
s
t
r
a
i
n
e
dm
i
n
i
m
i
z
a
t
i
o
n
.I
ti
sa
l
s
os
o
m
e
t
i
m
e
sr
e
f
e
r
r
dt
o
a
st
h
eb
a
n
a
n
a
f
u
n
c
t
i
o
n
.Ifyouhaveacopyo
f
t
h
O
p
t
i
m
i
<
.
a
t
i
o
nT
o
o
l
b
o
xfromMATLAB
youwi
1
ls
e
ei
sa
sp
a
r
to
f t
o
o
l
b
o
xdemo.Youwi
J
1a
l
s
on
o
t
ea
tt
h
et
o
o
l
b
o
x
e
c
h
n
i
q
u
e
sd
e
v
e
l
o
p
e
di
nt
h
i
ss
e
c
t
i
o
n
.Thes
e
c
o
n
dexamplei
sa
c
o
n
t
a
i
n
smanyo
f t
s
o
l
u
t
i
o
nt
oan
o
n
l
i
n
e
a
rt
w
o
p
o
i
n
tb
o
u
n
d
a
r
yva
u
ep
r
o
b
l
e
mt
h
a
ti
sd
u
e10eN
a
v
i
e
r
S
t
o
k
e
s
u
a
t
i
o
n
sd
e
s
c
r
i
b
i
n
gf
l
o
wd
u
et
oas
p
i
n
n
i
n
gd
i
s
k.1
l
a
s
ti
sa
nu
n
u
s
ua
d
a
t
af
i
ng
t
i
r
n
i
z
a
t
i
o
nT
o
o
l
b
o
xp
r
e
s
e
n
t
si
t
sv
e
r
s
i
o
n
e
x
a
m
p
l
eu
s
i
n
gB
e
z
i
e
rc
u
r
v
e
s
.Oncea
g
a
i
nt
h
eOp
o
ft
h
ed
a
t
af
i
t
t
i
n
ge
x
a
m
p
l
ewhenr
u
n
n
i
n
go
p
t
d
e
m
oa
tt
h
eMA
T
L
A
Bprompt
6
.
5
.
1 Example6.2-RosenbrockProblem
TheR
o
s
e
n
b
r
o
c
kproblemi
s

(
x
l
X
2
)=1
0
0
(
x
2-x
T
)
2+(
1-X
I
)
2
Minimize f
2
:
:
;
;
X
2
:
:
;
;
5
2
:
:
;
;
x
l:
:
;
;
5
; S
u
b
j
e
c
tt
o
: -

(
6
.
1
5
)

254 N
U
M
E
R
I
C
A
LT
E
C
H
N
I
Q
U
E
SFORU
N
C
O
N
S
T
R
A
I
N
E
DO
P
T
I
M
I
Z
A
T
I
O
N

6
.
5A
O
D
I
T
I
O
N
A
LE
X
A
M
P
L
E
S 255

Thes
i
d
ec
o
n
s
t
r
a
i
n
t
sa
r
eu
s
e
df
o
rd
r
a
w
i
n
gc
o
n
t
o
u
r
s
.Thes
o
l
u
t
i
o
nt
ot
h
i
sp
r
o
b
l
e
mi
s

;
x=1.0; ;
x=1.0;

I=
0
.
0

u
i
r
i
n
gal
a
r
g
enumbero
fi
l
e
r
a
t
i
o
n
sf
o
rc
o
n
v
e
r
g
c
e
s
.
Thcp
r
o
b
l
e
mi
sn
o
t
o
r
i
o
u
sf
o
r q
Changesi
nd
e
s
i
g
na
r
es
m
a
l
la
st
h
es
o
l
u
t
i
o
ni
sb
e
i
n
ga
p
p
r
o
a
c
h
e
d
.Heret
h
eRandom
Walka
n
dC
o
n
j
u
g
a
t
eG
r
a
d
i
e
n
lmethodsa
r
eu
s
e
dt
oexaminet
h
es
o
l
u
t
i
o
n
.B
o
t
ht
h
e
s
e
methodsa
r
ec
h
a
n
g
e
d1
0drawa
p
p
r
o
p
r
i
a
l
ec
o
n
l
o
u
rl
e
v
e
l
s
.Toa
v
o
i
dc
1u
t
t
e
r
d
i
s
a
b
l
e
p
r
i
n
t
i
n
gl
e
x
li
n
f
o
r
m
a
t
i
o
non l
i
n
e
sr
e
p
r
e
s
e
n
t
i
n
gd
e
s
i
g
nc
h
a
n
g
e
s
.A
l
s
o
s
i
n
c
e
s
e
methodsa
r
er
u
nf
o
r200i
t
e
r
a
t
i
o
n
st
h
ef
m
a
lp
r
i
n
t
i
n
g1
0t
h
eCommandwindowi
s
a
v
o
i
d
e
d
.Thet
r
a
c
o
ft
h
ed
e
s
i
g
nv
a
r
i
a
b
l
e
sont
h
ec
o
n
t
o
u
rp
l
o
ti
sp
r
o
v
i
d
e
da
st
h
eo
u
t
p
u
l
f
r
o
ml
h
eprogram
a
l
o
n
gw
l
t
hI
h
l
i
n
a
lv
a
l
u
e
s
.Ast
h
ef
i
g
u
r
ci
sb
e
i
n
gdrawn
i
ti
sc
1e
a
r
t
h
a
tt
h
ep
r
o
g
r
e
s
st
ot
h
es
o
l
u
t
i
o
ni
ss
l
o
w
.

S
o
l
u
t
l
o
nUsingRandomWalkMethod: F
i
g
u
r
e6
.
7p
r
e
s
e
n
l
st
h
er
e
s
u
l
lo
f
e
x
e
c
u
t
i
n
gRandomWalk
.m f
o
rt
h
l
se
x
a
m
p
l
e
.T
h
p
r
o
g
r
a
mi
n
v
o
c
a
t
i
o
na
n
do
u
t
p
u
tf
r
o
m
eCommandw
indowi
s

>
> RandomWalk('Example6_2'[
32
]
200
l
.Oe-08
0
l
20)
The problem: Example6_2
ans
2.0000e+002 1.0000e+000 1.0000e+000 1.2327e-007

S
i
n
c
et
h
r
ea
r
enos
t
o
p
p
i
n
gc
r
i
t
e
r
i
ab
u
i1
ti
n
t
ot
h
emethod(
w
a
ss
u
g
g
e
s
t
e
da
sa
n
e
x
e
r
c
i
s
e
)
t
h
emethodr
a
nf
o
r200i
l
e
r
a
t
i
o
n
s(
f
i
r
s
tv
a
l
u
e
)
. t
ef
i
n
a
lv
a
l
u
ef
o
rt
h
ed
e
s
i
g
n
V o
randeo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
na
r
ev
e
r
yi
m
p
r
e
s
s
i
v
e
:

;
x=1.0; ;
x=1.0; I=1.23e-07

randomwalk:

S
o
l
u
t
i
o
nUslngCo
ugateGradlentMethod:Thei
n
v
o
c
a
t
i
o
nando
u
t
p
u
tfrom
t
h
eCommandwindowi
s

njugateGradient('Example6_2'
The problem: Example6_2
No. of iterations: 200
ans
9.6B07e-001
9.3700e-001

2
)
200
1
.Oe 08
0
1
2
0
)

1.0222e-003

Thes
o
l
u
t
i
o
ni
s
c
o
n
j
u
g
a
t
eg
r
a
d
i
e
n
t:x
;=
0
.
9
6
8
0
7
;x
;=
0
.
9
3
7
;

r=1.0222e-03

T
h
i
sa
p
p
e
a
r
snomatchf
o
rt
h
eRandomWalkmethodf
o
ri
si
n
v
e
s
t
i
g
a
t
i
o
n
.F
i
g
u
r
e6
.
8
t
r
a
c
k
sed
e
s
i
g
nc
h
a
n
g
e
sont
h
ec
o
n
t
o
u
rp
l
o
t
.

6
.
5
.
2 Example6.3 Three-DlmensionalFlownearaRotatingDisk

T
h
i
sexampler
e
p
r
e
s
e
n
l
sa
n
o
t
h
e
ro
fI
h
ee
x
a
c
ts
o
l
u
t
i
o
n
st
ocN
a
v
i
e
r
S
t
o
k
e
se
q
u
a
t
i
o
n
s
fromf
l
u
i
dm
e
c
h
a
n
i
c
s
.Then
o
n
l
i
n
e
a
rt
w
o
p
o
i
n
tb
o
u
n
d
a
r
yv
a
l
u
ep
r
o
b
l
e
md
e
s
c
r
i
b
e
sa
v
i
s
c
o
u
sf
l
o
wa
r
o
u
n
daf
l
a
td
i
s
ka
tr
o
t
a
t
e
sa
b
o
u
t a
x
i
sp
e
r
p
e
n
d
i
c
u
l
t
oi
t
sp
l
a
n
ewi
ac
o
n
s
t
a
n
ta
n
g
u
l
a
rv
e
l
o
c
i
t
y
.A
f
t
e
rs
u
i
t
a
b
l
e d
e
f
m
i
t
i
o
n[
1
0
]t
h
em
a
t
h
e
m
a
t
i
c
a
l
d
e
s
c
r
i
p
t
i
o
no
ft
h
ep
r
o
b
l
e
mi
st
h
ef
o
l
l
o
w
i
n
gs
e
to
fc
o
u
p
l
e
dn
o
n
l
i
n
e
a
re
q
u
a
t
i
o
n
sw
i
t
h
b
o
u
n
d
a
r
yc
o
n
d
i
t
i
o
n
sa
ttwop
o
i
n
l
s
:

2F+H'=0

(
6.
16
a
)

F2+F'
H-G2-F"=0

(
6
.
1
6
b
)

2FG+HG'-G"=0

16
c
)
(
6.

F
l
g
u
r
e6
.
7 Randomw
a
l
km
e
t
h
o
d
:E
x
a
m
p
l
e6
.
2
.

256 N
U
M
E
A
I
C
A
lT
E
C
H
N
I
Q
U
E
SF
O
RU
N
C
O
N
S
T
R
A
I
N
E
DO
P
T
I
M
I
Z
A
T
I
O
N

6
.
5

Y
2
Y
I
Y
2
Yi+Y

Y
4
Y
3
Y
4
2
Y
I
Y
3+Y

v
'
5
Y
;
2
Y
I

A
D
D
I
T
I
O
N
A
LE
X
A
M
P
L
E
S 257

Y
I
(
O
)

Y
2
(
O
)

~II

Y
3
(
0
)
Y
4
(
O
)
Y
5
(
O
)

(
6
.
1
9
)

x2

Thcs
t
a
t
ce
q
u
a
t
i
o
n
s(
6
.
1
9
) .
ci
n
t
c
g
r
a
l
e
du
s
i
n
gMATLABf
u
n
c
t
i
o
no
d
e
4
5
.

.
.

S
o
l
u
t
l
o
ntotheProblem:Thet
h
r
e
e
v
a
r
i
a
b
l
eproblemi
ss
o
l
v
e
du
s
i
n
gt
h
eDFP
t
ums t
h
eo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
n
.I
tc
a
l
l
st
h
ef
u
n
c
t
i
o
n
m
e
t
h
o
d
. Example6_3.m r
Ex
6_3_state.mwhichr
c
t
u
r
n
st
h
es
y
s
t
e
mq
u
a
t
i
o
n
si
ns
t
a
t
ef
o
n
n
.Thec
a
l
lt
ot
h
e
n
i
t
i
a
ld
e
s
i
g
nv
e
c
t
o
ro
f[
1-18
Jw
i
t
h20i
t
e
r
a
t
i
o
n
so
ft
h
eDFP
DFPmethodw
i
t
hei
andat
o
l
e
r
a
n
c
ef
o
rt
h
eg
o
l
d
e
ns
e
c
t
i
o
no
f0
.
0
0
0
1i
sshownb
e
l
o
w
.A
J
s
oi
nt
h
e
method
c
a
l
li
st
h
ei
n
p
u
tt
ol
h
eu
p
p
e
rboundc
a
l
c
u
l
a
t
i
o
n
.

U
s
a
g
e
: DFP(
'Example6_3'[
1 18]
20
0.0001
0
1
20) (
6
.
2
0
)

Thef
i
n
a
lv
a
l
u
ef
o
rt
h
ed
e
s
i
g
na
f
t
e
r20i
t
e
r
a
t
i
o
n
swas

F
(
O
)=
0
;G
(
O
)=
1
; H(O)=
0

(
6
.l
7a
)

F
(
)=0; G( )=0

(
6
.
1
7
b
)

Eq
u
a
t
i
o
n
s(
6
.
1
6
)a
r
ec
o
n
v
e
r
t
e
d

s
t
a
t
es
p
a
f
onn(
s
C
h
a
p
t
e
r5
Example5.
3
)
.The
G'(O)andt
h
ef
i
n
a
l
d
e
s
i
g
nv
a
r
i
a
b
l
e
sa
r
Et
h
em
i
s
s
i
n
gboundaryc
o
n
d
i
6
o
r
t
sonF
'
(
O
)
v
a
l
u
eoft
h
ei
n
d
e
p
e
n
d
e
n
tv
a
r
i
a
b
l
e(
i
nl
i
e
uof
)
.Theo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
ni
ses
q
u
a
r

c
r
r
o
ri
nt
h
ei
n
t
e
g
r
a
t
e
dv
a
J
u
ea
lt
h
ef
i
n
a
lc
o
n
d
i
t
i
o
n
.
TheO
p
t
l
m
l
z
a
t
i
o
nProblem
Minimize f
(
xl'x2
x3
)
: Y
I
(
X
)
2+Y
3
(
x
)
2
3

.
t
;
(
t
.
t

)=
-0.
6
1
5
7
;x
j
(
=1
0
.
7
6
2
7
;j=
5 08e06(
6
.
2
1
)

1
1
1
et
o
l
e
r
a
n
c
ef
o
rt
h
eg
o
l
d
e
ns
e
c
t
i
o
nwaso
n
l
y0
.
0
0
01
.h i
sp
o
s
s
i
b
l
cl
h
co
b
j
e
c
t
i
v
c
o
l
e
r
a
n
c
ei
ss
p
e
c
i
f
i
e
d
.
c
a
n
n
o
td
e
c
r
c
a
s
cu
n
l
i
las
m
a
l
lrt

Thep
u
b
l
i
s
h
e
dv
a
l
u
e
sf
o
rt
h
ed
e
s
i
g
nv
a
r
i
a
b
l
e
sa
r

F
l
g
u
r
e6
.
8C
o
n
j
u
g
a
t
eg
r
a
d
l
e
n
tm
e
t
h
o
d
:E
x
a
m
p
l
e6
.
2
.

x
;
(
) 0.
51
0
1
;

(
6
.
1
8
)

wheret
h
cs
l
a
l
cv
a
r
i
a
b
1
e
sa
r
eo
b
l
a
i
n
e
df
r
o
mt
h
es
o
l
u
t
i
o
no
ft
h
cf
o
l
l
o
w
i
n
gi
n
i
t
i
a
lv
a
l
u
e
p
r
o
b
l
e
m
:

;
x=0.5101; ()=-0.6159

(
6
.
2
2
)

whichi
sr
e
m
a
r
k
a
b
l
yc
l
o
s
et
oI
h
ev
a
l
u
e
so
b
t
a
i
n
e
dbyt
h
eDFPm
e
t
h
o
d
.Thev
a
J
u
eo
f
sI
L
s
l
L
a
l
l
yn
o
ts
p
e
c
i
f
i
e
db
U
Ii
sa
b
o
u
t7
.S
i
n
c
ei
ti
sas
u
b
s
t
i
t
u
t
ef
o
ri
n
f
i
n
i
t
y
t
h
eJ
a
r
g
e
r
x
ji
numbers
h
o
u
l
dn
o
tbeap
r
o
b
J
e
m
.
Example6.
3i
sn
o
tt
r
i
v
i
a
lp
r
o
b
l
e
m
.Thei
n
t
e
g
r
a
t
i
o
ni
sh
i
g
h
l
ys
e
n
s
i
t
i
v
et
ot
h
e
i
n
i
t
i
a
lv
a
l
u
e
s
.A r
e
a
s
o
n
a
b
l
es
t
a
r
t
i
n
gp
o
i
n
ti
se
s
s
e
n
t
i
a
lt
op
r
e
v
e
n
tt
h
ei
n
t
e
g
r
a
t
i
o
n
s(
n
o
lan
u
m
b
e
r
)
.Thes
e
c
o
n
dd
e
s
i
g
nv
a
r
i
a
b
l
emuslhavea
fromg
e
n
e
r
a
t
i
n
gNaN'
l
l
u
s
t
r
a
t
e
st
h
ea
p
p
l
i
c
a
t
i
o
no
fs
t
a
n
d
a
r
do
p
t
i
m
i
z
a
t
i
o
n
n
c
g
a
t
i
v
es
i
g
n
.Theexampl i
t
e
c
h
n
i
q
u
et
os
o
J
v
ean
o
n
l
i
n
e
a
rd
i
f
f
e
r
e
n
t
i
a
ls
y
s
t
e
m
.A s
i
m
i
l
a
ra
p
p
J
i
c
a
t
i
o
nc
a
nb
e
lI
ti
se
s
s
e
n
t
i
a
lt
o
madet
oproblemsi
ns
y
s
l
e
mdynamicsando
p
t
i
m
a
1c
o
n
t
r
o.
h
ea
l
g
o
r
i
t
h
m
so
u
t
l
i
n
e
dh
e
r
ea
r
en
u
m
e
r
i
c
a
lt
o
o
1
s a
tt
r
a
n
s
c
e
n
dany
u
n
d
e
r
s
t
a
n
d t
p
a
r
t
i
c
u
J
a
rd
i
s
c
i
p
l
i
n
e
.
Th
e
r
ea
r
en
u
m
e
r
i
c
a
Jt
e
c
h
n
i
q
u
e
sa
ta
d
d
r
e
s
st
w
o
p
o
i
n
tn
o
n
1
i
n
e
a
rb
o
u
n
d
a
r
yv
a
l
u
e

c
o
m
p
u
t
i
n
gr
e
s
o
u
r
c
e
s

e
s
p
e
c
i
a
l
l
yonaPC
a
r
en
o
t
p
r
o
b
l
e
m
smoree
f
f
i
c
i
e
n
t
l
y
.However
t
1
1
ep
r
o
c
c
d
u
r
ea
d
o
p
t
e
di
ni
scxamplei
sa
na
c
c
e
p
t
a
b
1
ea
p
p
r
o
a
c
h
ac
o
n
c
e
r
n
.T
h
e
r
e
f
o
r
e

U
M
E
R
I
C
A
LT
E
C
H
N
I
Q
U
E
SF
O
RU
N
C
O
N
S
T
R
A
I
N
E
OO
P
T
I
M
I
Z
A
T
I
O
N
258 N

6
.
5A
O
O
I
T
I
O
N
A
LE
X
A
M
P
L
E
S

259

f
o
rt
h
i
sc
l
a
s
so
fp
r
o
b
l
e
m
s
.Event
h
eRandomWalkmethodi
si
m
p
r
e
s
s
i
v
ei
ni
s
m
s
t
a
n
c
e
.
'
1

6
.
5
.
3 Example6
F
i
t
t
l
n
gBezierParametrlcCurves
.4Theexamplei
n
t
r
o
d
u
c
e
san
o
n
t
r
a
d
i
t
i
a
la
n
da
nu
n
u
s
u
a
lc
u
r
v
ef
i
u
i
l
l
gp
r
o
c
e
d
u
r
e
.
e
z
i
e
rc
u
r
v
e01u
s
e
r
c
h
o
s
e
no
r
d
e
ri
sf
i
l10t
h
ed
a
t
a
.B
e
z
i
e
r
Givenas
e
to
fd
a
t
ap
o
i
n
t
saB
p
e
c
i
a
lc
a
s
eo
fu
n
i
f
o
r
mB
s
p
l
i
n
e
s
.B
e
z
i
e
r
c
u
r
v
e
sa
r
ep
a
r
a
m
e
t
r
i
cc
u
r
v
e
s as
e
m
s
t
e
i
nb
a
s
i
sf
u
n
c
t
i
o
n
s
. Any p
o
i
n
t P on a
p
a
r
a
m
e
t
r
i
z
a
t
i
o
ni
sb
a
s
e
d on t
h
eB
I
w
o
d
i
m
e
n
s
i
o
n
a
lB
e
z
i
e
rc
u
r
v
e(
a
n
yp
a
r
a
m
e
t
r
i
cc
u
r
v
e
)i
sa
C
l
u
a
l
l
yo
b
t
a
i
n
e
d P
(
x
(
v
)
y
(
v
)
)
where0SvS1.l
ea
c
t
u
a
lr
e
l
a
t
i
o
n
sa
r
e

0.
7

./'

.~

0
.
6

--

805

~.

1
.
1
.
1
/

.
'/
1
/

(
1
)
"
.
1

whereB; E ev
e
r
t
i
c
e
so
fep
o
l
y
g
o
nt
h
a
td
e
t
e
r
m
i
n
s
t
h
ec
u
r
v
e
.B
;r
e
p
r
e
s
e
n
tap
a
i
r
I-th
h
i
g
h
e
s
tpower
o
fv
a
l
u
e
si
nt
w
o
d
i
m
e
n
s
i
o
n
a
ls
p
a
c
e
.Theo
r
d
e
roft
h
ec
u
r
v
ei
sl
i
nl
h
eb
a
s
i
sf
u
n
c
t
i
o
n
s
.Thea
c
l
u
a
lc
o
m
p
u
t
a
t
i
o
n
so
ft
h
ep
o
i
n
l
SonI
h
eC
U
l
v
e

'
ee
l
l
s
i
e
r
da
n
ddocumentedi
nR
e
f
e
r
e
n
c
e1
1
.F
i
g
u
r
e
u
s
i
n
gm
a
t
r
i
xalgebma
n
da
r
ew
e
l
le
x
p
l
a
i
n
u
b
i
cBz
i
e
rc
u
r
v
e
.Thef
o
l
l
o
w
i
n
ga
r
esome0'
1l
h
emostu
s
e
f
u
l
6
.
9represenL~ ac
p
r
o
p
i
e
so
ft
h
ec
u
r
v
ea
so
b
s
r
v
e
di
nt
h
ef
i
g
u
r
e
:
Thec
u
r
v
ei
sc
o
m
p
l
e
t
e
l
yd
e
f
i
n
e
dbyt
h
ep
o
l
y
g
o
no
b
t
a
i
n
c
dbyj
o
i
n
i
n
gt
h
ev
e
r
t
i
c
e
s
i
no
r
d
e
r
.

'hedegreeofepolynomidefiningthecurveisonelessanthenumberof
s
o
ft
h
ep
o
l
y
g
o
n
.
v
e
r
t
i
. Th
ef
i
r
s
la
n
dl
a
s
tp
o
i
n
t
so
ft
h
ec
u
r
v
ea
r
ec
o
i
n
c
i
d
c
n
tw
i
t
ht
h
ef
i
r
s
ta
n
dl
a
s
tv
e
r
t
e
x
.
I
eont
h
cc
u
r
v
c
.
Thcr
e
m
a
i
n
i
n
gv
e
r
t
e
xp
o
i
n
t
sdon
o
tl
y
p
i
c
a
l
l
yi

. Thes
l
o
p
c
sa
lt
h
ee
n
d
so
ft
h
ec
u
r
v
ch
a
v
et
h
esamed
i
r
e
c
t
i
o
na
st
h
es
i
d
e
so
ft
h
e
p
o
l
y
g
o
n
. Thec
u
r
v
ei
sc
o
n
t
a
i
n
e
dw
i
t
h
i
nt
h
econvexh
u
l
lo
fe
p
o
l
y
g
o
n
.
p
l
et
h
ec
o
n
v
e
n
i
e
n
tm
a
t
r
i
xr
e
p
r
e
s
e
n
t
a
t
i
o
n
sf
o
rt
h
cc
u
r
v
ea
r
cu
s
e
d
.Th
e
ya
r
e
I
nt
h
i
se
x n
h
o
w
e
v
e
r
.b
u
ta
r
ea
v
a
i
l
a
b
l
ci
nt
h
ec
o
d
c
.F
o
ri
se
x
a
m
p
l
e
.t
h
ed
c
s
i
g
n
n
o
td
i
s
c
u
s
s
e
dh
e
r
e
t
h
c
r
ea
r
c6v
e
c
.Thef
i
r
s
ta
n
d
v
a
r
i
a
b
l
e
sa
r
et
h
ei
n
s
i
d
ep
o
l
y
g
o
nv
e
r
t
i
c
e
s
.F
o
rn 5
i
g
nv
a
r
i
a
b
l
e
s
l
a
s
tv
c
r
t
e
xa
r
eknownfromt
h
ed
a
t
ap
o
i
n
t
s
.T
h
i
sl
e
a
v
e
s4v
e
r
t
i
c
e
so
r8ds
f
o
r
n=
5
.

i
sm-f
i
Ici
sas
t
a
n
d
a
l
o
n
eprogramt
or
u
nt
h
ccxamplci
nt
h
i
s
Example6_
4.m: l
s
c
c
t
i
o
n
.TypingExample6_4a
lt
h
eprompts
h
o
u
l
ds
t
a
r
tt
h
cp
r
o
g
r
a
m
.A
c
t
u
a
l
l
yt
h
e

l'

J
/

0
.
3

J
I
I
.
;
(
v
)

~.

/ "

(
6
.
2
3
)

;
=
0

-~

0
.
4

"
/

a
'

/i
f /

P
(

v)
=
[

x
y)=
BJ

.
1

(
v

0
.
8

C
!
I

0
.
2

E3

' "

0
.
1

0
.
2

0.
4

xd
a
t
a

1
.
6

F
l
g
u
r
e6
.
9 A
c
u
b
l
cBz
lr
c
u
r
v
e
.

examplei
sg
e
n
e
r
i
ca
smostd
e
c
i
s
i
o
n
sa
b
a
s
e
don et
e
x
tf
i
l
et
h
a
tc
o
n
t
a
i
n
st
h
ed
a
t
a
t
obef
i
u
e
d
.1
1c
a
l
l
sonm ym
f
i
l
e
s
.TheB
e
z
i
e
rc
a
I
cu
l
a
t
i
o
n
sa
r
ef
a
i
r
l
ym
o
d
u
l
a
r
.T
h
is
programi
sn
o
t v
i
a
l
.Amorei
m
p
o
r
t
a
n
t a
s
o
nf
o
ri
t
si
n
c
l
u
s
i
o
ni
sa
t
i
fu
n
d
e
r
s
t
o
o
d

i
ts
i
g
n
a
l
sar
e
a
s
o
n
a
b
l
em
a
t
u
r
i
t
yw
i
t
hMATLABp
r
o
g
r
a
m
m
i
n
g
.Morei
m
p
o
r
t
a
n
l
l
yi
t
oe
x
p
l
o
i
lt
h
eimmensepowero
fb
u
i
l
t
i
nMATLABsuppo
.Th
e
e
x
e
m
p
l
i
f
i
e
sewayt
a
u
t
h
o
re
n
c
o
u
r
a
g
e
st
h
eu
s
e
rt
oe
x
e
r
tsomce
f
f
o
r
tt
r
a
c
i
n
gt
h
ep
r
o
g
r
e
s
so
ft
h
ec
a
l
c
u
l
a
t
i
o
n
s
r
oughec
o
d
e
.A 1
0
1o
fprogramc
o
n
t
r
o
li
sb
a
s
e
doneu
s
e
r
'sc
h
o
i
c
econceming
t
h
eo
r
d
e
ro
ft
h
eB
e
z
i
e
rc
u
r
v
e
.A
l
s
ot
h
ei
n
i
t
i
a
lv
a
l
u
e
sf
o
rt
h
ed
e
s
i
g
nv
a
r
i
a
b
l
e
sa
r
e

"basedont
h
ed
a
t
a10bef
i
t
t
e
d
.Thef
u
n
c
t
i
o
nc
a
l
l
i
n
gs
e
q
u
e
n
c
e
g
e
n
e
r
a
t
e
da
u
t
o
m
a
t
i
c
a
l
sexamplec
a
nbeb
e
t
t
c
ru
n
d
e
r
s
t
o
o
dbyt
h
ef
o
l
l
o
w
i
n
gs
t
r
u
c
t
u
r
e
:
u
s
e
di
ni
Example6_5_3.m
*.
t
x
t
)I
ca
tc
o
n
t
a
i
n
stwo-columnd
a
t
aa
tw
i
l
lbe
CaUsu
i
g
e
t
f
i
l
ct
op
i
c
kupe(
f
i
t(
E
x
6
_
5
_
3
.
t
x
t
)

'
h
r
owsupa
ni
n
p
u
t
d
i
a
l
o
gboxt
oo
b
t
a
i
nt
h
eo
r
d
e
roff
i
tfromt
h
eu
s
e
r(0 5
)
C
a
l
l
scoeff.mw
i
t
ht
h
eo
r
d
e
ri
n
f
o
r
m
a
t
i
o
n
1c
o
m
b
i
n
a
t
i
o
nf
o
rB
c
r
n
s
t
e
i
nb
a
s
i
s
C
a
l
l
scomblnatlon.m-c
a
l
c
u
l
a
t
i
o
n0'
I
cu
l
a
t
e
sf
a
c
t
o
r
i
a
lo
fa
ni
n
t
e
g
e
r
C
a
l
l
sFactorial.m-c
a

260

6
.
5 ADDITIONALEXAMPLES

NUMERICALTECHNIQUESFORUNCONSTRAINEDOPTIMIZATION

C
a
l
l
sC
u
r
v
e
_
f
i
t
.m whichhandlest
h
eo
p
t
i
m
i t
i
o
n
l
IsDFP
Ca
l
IsBez
Sq_Err.mc
a
l
c
u
l
a
t
i
o
no
fO~
je
:ctiv

ef
u
n
c
t
i
o
n
Ca
C
a
l
l
sg
r
i

'
s
d
f

un.m-c
a
l
c
u
l
a
t
i
o
no
fd
e
r
i
v
a

t
i
v
s
C
a
l
l
sGold_Section
nV

'
.

m
- 1
Ds
t
e
p
s
i
z
e
C
a
l
l
sUpperBound_nVar.m-b
r
a
c
k
e
tt
h
eminimum

objective funcionvalue:
iteration number:
3
gradient length squared:
objective function value:
4
iteration number:
gradient length squared:
objective function value:
iteration number:
5
gradient length squared:
objective function value:
iteration number:
6
gradient length squared:
objective function value:
7
iteration number:
gradient length squared:
objective function value:

iteration number:
gradient length squared:
objective function value:
iteration number:
9
gradient length squared:
objective funcionvalue:

a
11o
ft
h
ei
n
f
o
r
m
a
t
i
o
ne
x
c
e
p
tf
o
ri
t
e
r
a
t
i
o
n
S
i
n
c
et
h
e
r
ea
r
ee
i
g
h
td
e
s
i
g
nv
a
r
i
a
b
l
e
s
n
u
m
b
e
r
.c
o
n
v
e
r
g
e
n
c
ec
r
i
t
e
r
i
aa
n
d o
b
jc
t
i
v
ef
u
n
c
t
i
o
ni
ss
u
p
p
r
e
s
s
e
df
r
o
mt
h
e
Commandwindowt
op
r
e
v
e
n
ts
c
r
o
l
l
i
n
ga
n
dl
o
s
so
fc
o
m
p
r
e
h
e
n
s
i
o
n
.
l
es
i
g
n
i
f
i
c
a
n
te
l
e
m
e
n
t
si
ni
sc
o
d
e
a
r
a
sf
o
l
l
o
w
s
:
P
i
c
k
suped
a
t
af
i
l
et
h
r
o
u
g
baf
i
l
es
e
l
e
c
t
i
o
nb
o
x
.
U
s
e
sa
ni
n
p
u
td
i
a
1
0gboxt
oq
u
e
r
yt
h
eu
s
e
ront
h
eo
r
d
e
ro
ft
h
ec
v
e
.
r
d
e
t
e
r
m
i
n
e
st
h
es
i
z
eo
ft
h
ec
o
e
f
f
i
c
i
e
n
tm
a
t
r
i
xAa
n
di
ti
sa
u
t
o
m
a
t
i
ca
11
y
Theord
g
e
n
e
r
a
t
.
Thenumbero
fd
e
s
i
g
nv
a
r
i
a
b
l
e
si
sa
u
t
o
m
a
t
i
c
a
l
l
yd
e
t
e
r
m
i
n
e
d
.Oncea
g
a
i
ni
ti
s
r
d
e
ro
ft
h
ec
u
r
v
e
.
basedon o
Thei
n
i
t
i
a
lv
a
l
u
e
sf
o
rt
h
ed
e
s
i
g
na
r
ea
u
t
o
m
a
t
i
c
a
l
l
yg
e
n
e
r
a
t
e
db
a
s
e
dont
h
eda
t
ob
ef
i
.
t
TheDFPm
e
t
h
o
di
su
s
e
dt
of
i
n
dt
h
eo
p
t
i
m
u
m
.
Theo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
ni
sc
o
m
p
u
t
e
di
nt
h
eg
e
n
e
r
i
cf
i
IeBeziecSq_Err.m.The
a
m
p
l
ea
r
ep
a
s
s
e
dt
h
r
o
u
g
heMATLABg
l
o
b
a
ls
t
a
t
e
m
e
n
.
t
d
a
t
as
p
e
c
i
f
i
ct
oex
Theo
r
i
g
i
n
a
ld
a
t
as
e
ta
n
d B
e
z
i
e
rc
u
r
v
ef
i
t
t
i
n
gt
h
ed
a
t
aa
r

d
i
s
p
l
a
y
e
da
tt
h
e
e
n
d
.
Theo
b
j t
iv
ef
u
n
c
t
i
o
na
n
dt
h
c
o
n
v
e
r
g
e
n
c
ed
a
t
af
o
re
a
c
hi
t
e
r
a
t
i
o
na
r
ep
r
i
n
t
e
d
t
ot
h
eCommandw
i
n
d
o
w
.

O
b
J
e
c
t
l
v
eF
u
n
c
t
l
o
n
:Th o
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
ni
st
h
e1
s
ts
q
u
a
r
e
de
r
r
o
ro
v
e
ra
l
lo
f
t
h
ed
a
t
ap
o
i
n
t
sb
e
t
w
e
e
nea
c
t
u
a
ld
a
t
aa
n
dt
h
ef
i
t
t
e
dB
e
z
i
e
rc
u
r
v
e
.I
ti
se
x
p
r
'
s
s
e
da
s

I
yY
B
T

IIDa

Mi
n
i
m
i
z
e
: X):

1
<
=1
'
"

(
6
.
2
4
)

'
1
1
'

.
J

t
m
i
t
t

The f
o
l
l
o
w
i
n
gi
n
f
o
r
m
a
t
i
o
ni
sc
o
p
i
e
df
r
o
mt
h
eCommand window (
t
h
ef
i
l
ei
s
r
d
e
ri
s5
)
:
E
x
6
_
5
_
3
.
t
x
ta
n
deo
Bxample6_4
iteration number:
1
gradient length squared:
objective function value:
iteration number:
2
gradient length squared:

"
'
I d - - R W 43

U
I

5.6890

Eli--

5.0001
0.5302

2
6
1

0.2497
0.0546
0.1830
0.1322
0.1605
0.0573
0.1467
0.0374
0.1437
0.0071
0.1433
0.2793
0.1415
0.4337
0.1401

Fromt
h
ed
a
t
aa
b
o
v
ei
ta
p
p
e
a
r
st
h
a
tt
h
ee
o
ri
sa
r
o
u
n
d0
.
1
4a
n
di
sd
i
f
f
i
c
u
l
tt
or

i
u
c
e
ta
l
s
oa
p
p
e
a
r
st
h
er
e
d
u
c
t
i
o
ni
nt
h
eo
b
j
e
c
t
i
v
ei
sa
c
c
o
m
p
a
n
i
e
db
yt
h
ei
n
c
r
e
a
s
e
f
u
r
t
h
e
r
.I
i
nt
h
eg
r
a
d
i
e
n
t
.a
f
t
e
rac
e
r
t
a
i
nnumbero
fi
t
e
r
a
t
i
o
n
s
.
l
ti
smovingawayf
r
o
ms
a
t
i
s
f
y
i
n
g
t
h
eFOC.Theb
e
s
tv
a
l
u
ef
o
rt
h
eFOCi
sa
b
o
u
t0
.
0
0
7a
n
ds
t
a
r
t
st
ow
o
r
s
e
nw
i
t
hf
u
r
t
h
e
r
t
I
o
n
s
.I
ti
sq
u
i
t
ep
o
s
s
i
b
l
e n
u
m
e
r
i
c
a
li
n
a
c
c
u
r
a
c
i
e
smayb
eap
r
o
b
l
e
ma
st
h
e
r
e
i
t
e
r
a
t
e
n
s
i
v
ec
o
m
p
u
t
a
t
i
o
n
si
n
v
o
l
v
e
di
ne
a
c
hi
t
e
r
a
t
i
o
n
.I
ti
sa
l
s
ol
i
k
e
l
y a
te
a
r
ex
1g
r
a
d
i
e
n
tc
o
m
p
u
t
a
t
i
o
nmayn
e
e
dt
ober
e
f
i
n
e
d
.C
u
r
r
e
n
t
l
yegr
i
e
n
t
s e
n
u
m
e
r
i
ca
.T
h
e
s
ea
r
ep
a
r
a
m
e
t
e
r
s
c
o
m
p
u
t
e
du
s
i
n
gf
o
r
w
a
r
df
i
n
i
t
ed
i
f
f
e
r
e
n
c
ew
i
t
has
t
e
po
f
O
.
O
O1
dt
ob x
p
l
o
r
e
di
no
p
t
i
m
i
z
a
t
i
o
na
st
h
e
r
ei
snou
n
i
v
e
r
s
a
ls
e
to
fv
a
l
u
e
s
t
l
t
a
tn
f
o
ra
l
lc
l
a
s
s
e
so
fp
r
o
b
l
e
m
s
.
r
i
g
i
na
1d
a
t
a dt
h
eB
e
z
i
e
rc
u
r
v
ea
e
rel
a
s
ti
t
e
r
a
t
i
o
n
.
F
i
g
u
r
e6
.
1
0i
l
l
u
s
t
r
a
t
e
seo
r
a
t
et
h
ec
v
e
.

h
ef
i
ti
sd
e
f
i
n
i
t
e
l
y
Thep
o
l
y
g
o
n
sa
r
en
o
tshowna
so
u
re
f
f
o
r
ti
st
og
e
n
e
z
i
e
rc
u
r
v
ec
a
ndoa
b
o
u
tac
o
u
p
l
eo
fp
o
i
n
t
s
a
c
c
e
p
t
a
b
l
ea
n
dt
h
e
r
ei
sp
r
o
b
a
b
l
yl
i
t
t
l
eeB
ti
sp
r
o
b
a
b
l
yt
h
e
s
ep
o
i
n
t
st
h
a
tdon
o
ta
110wt
h
e
t
h
a
ta
r
en
o
tv
e
r
ys
m
o
o
t
h
l
yl
o
c
a
t
e
d
.I
t
Io
nt
od
e
c
r
e
a
s
ef
u
r
t
h
e
rw
i
t
h
o
u
ti
n
c
r
e
a
s
i
n
gt
h
g
r
a
d
i
e
n.
tThi
sc
o
n
c
l
u
d
e
s
o
b
j
e
c
t
i
v
ef
u
n
c
ea
u
t
h
o
rh
a
su
s
e
dt
h
e
s
e
a
n
o
t
h
e
ri
n
n
o
v
a
t
i
v
eu
s
eo
fu
n
c
o
n
s
t
r
a
i
n
e
do
p
t
i
m
i
z
a
t
i
o
n
.l
.
t
c
u
r
v
e
st
oo
p
t
i
m
a
l
l
yd
e
s
i
g
na
i
r
f
o
i
l
sf
o
rl
o
w
s
p
e
e
df
l
i
g
h

NoteonFOCf
o
rUnconstrainedM
i
n
i
m
i
z
a
t
i
o
n
: exampleh
a
sa
l
e
r
t
e
du
s
.I
fa
n
a
l
y
t
i
c
a
lc
o
m
p
u
t
a
t
i
o
no
ft
h
ed
e
r
i
v
a
t
i
v
ei
sp
o
s
s
i
b
l
e
t
ot
h
eq
u
e
s
t
i
o
no
fFOCVf=O
.l
nd
e
s
i
g
np
r
o
b
l
e
m
sl
i
k
e
f
o
rap
r
o
b
l
e
m
.t
h
e
nFOCh
a
st
h
el
i
k
e
l
i
l
t
o
o
do
fb
e
i
n
gmet

4
.w
h
e
r
et
h
ee
l
e
m
e
n
t
so
ft
h
eg
r
a
d
i
e
n
tv
e
c
t
o
rh
a
v
et
h
ef
o
r
m
Example6

262

PROBLEMS

NUMERICALTECHNIQUESFORUNCONSTRAINEDOPTIMIZATION

au-

Ed-

d
u

f
o
-

aEF-n"au-

OB-

hv--

O
4

5
1

a
a

{SEEE )S

aq

1
1
-

oo

0
.
5

5
.F
l
e
t
c
h
e
r
R
.
a
n
dR
e
e
v
c
sR
.M.F
u
n
c
t
i
o
nM
i
n
i
m
i
z
a
t
i
o
nb
yC
o
n
j
u
g
a
l
eG
r
a
d
i
e
n
t
s
.The
rJ
o
u
m
a
l
.Vo
.
l7
p
p
.1
4
9
1
8
0
1
9
6
4
.
Compute
.V
a
r
i
a
b
l
eM
e
t
r
i
cM
e
t
h
o
d
sf
o
rM
i
n
i
m
i
z
a
l
i
o
nU
.
S
.A
l
o
m
i
cE
n
e
r
g
y
6
.D
a
v
i
d
o
nW.C
C
o
m
m
i
s
s
i
o
nR
e
s
e
a
r
c
ha
n
dD
e
v
e
l
o
p
m
e
n
tR
e
p
o
r
lN
o
.ANL-5
9
9
0
Ar
g
o
n
n
eN
a
t
i
o
n
a
l
La
b
o
r
a
t
o
r
y1
9
5
9
.
7
.Huang
H
.Y
.U
n
i
f
i
e
dApproachQua
caJl
yC
o
n
v
e
r
g
e
n
tA
1
g
o
r
i
t
h
m
sf
o
rFu
n
c
t
i
o
n
Min.
t
i
o

nJoumalo
fOp
uw
t
I
on'
a
o
r
yandA
p
p
l
i
c
a
t
i
o
/
l
s

Vo
l
.5

p
p
.4
0
5-4
2
3
1
9
7
0
.
8
.V
a
n
d
e
r
p
1
a
a
t
sG
.N
.Nl/ln
e
r
i
c
a
lO
p
t
i
m
i
z
a
t
i
o
/
l T
e
c
h
n
i
q
u
e
sf
o
rE
n
g
i
n
e
e
r
i
n
gD
e
s
i
g
n
McGraw-Hi
l
1
NewY
o
r
k1
9
8
4
.
9
.Hoc

k W.
a
n
dS
c
h
i
t
t
k
o
w
s
k
iK
.
T
e
s
tExa
m
p
l
e
sf
o
rNonL
i
n
e
a
rProgrammi
l
!
gCodes

Le
c
t
u
r
eN
o
t
e
sI
nE
c
o
n
o
m
i
ca
n
dM
a
t
h
e
m
a
t
i
c
a
lS
y
s
t
e
m
s1
8
7S
p
r
i
n
g
e
r
V
e
r
l
a
gB
e
r
l
i
n
1
9
8
0
.
1
0
.S
c
h
l
i
c
h
t
i
n
g

H
.
Bounda
ILa
y
e
rT
h
e
o
r
y

McGraw-H
i
l
I
NewY
ork
1
9
7
9
.
1
.R
o
g
e
r
s

G
.F
.
a
n
dAdams
J
.A
.
M
a
t
h
e
m
a
t
i
c
a
lE
l
e
m
e
n
t
s
f
o
rComputerG
r
a
p
h
i
c
s

2
n
de
d
.
1
McGraw-H
i
l
I
NewY
ork
1
9
9
0
.

PROBLEMS
4

e
.
.
.
_

)o.

263

-e.
.

h
eRandomWalkmethodt
oExample5
.
2
.
6
.
1 Applyt

1
.
5

x
d
a
t
a(
r
e
a
dfromf
l
l
e
)

6
.
2 S
o
l
v
eu
s
i
n
gt
h
eRandomWalkm
e
t
h
o
d
:
2
.
5

Minimize

=
x
1-22+ x1+~-2x. +4

!
(
X
.
x
2
)

5
.
2
.
h
eP
a
t
t
e
mS
e
a
r
c
hmethodt
oExampl
6
.
3 Applyt

r
l
g
l
n
a
ld
a
t
aa
n
dB
e
z
l
e
rc
u
r
v
e
.
F
l
g
u
r
e
6
.
1
0O

d
:
o
l
v
eu
s
i
n
gt
h
ePattemS
e
a
r
c
hm o
6
.
4 S
Minimize

)
i
andwhicha
r
eb
e
s
tcomputedn
u
m
e
r
i
c
a
l
l
y
somes
p
e
c
i
f
i
c
a
t
i
o
n
sa
r
eneededr
e
g
a
r
d
i
n
g
d
e
r
i
v
a
t
ivec
o
m
p
u
t
a
t
i
o
n
si
ft
h
e
ya
r
et
obeu
s
e
df
o
re
n
f
o
r
c
i
n
gt
h
en
e
c
e
s
s
a
r
yc
o
n
d
i
t
i
o
n
s
.

REFERENCES
1
.F
l
e
t
c
h
e
r
.R P
r
a
c
t
i
c
a
lM
e
t
l
lO
d
s
f
o
rO
p
t
i
m
i
w
t
i
o
n
Vo1
.
1
W
i
l
e
y
NewYork
1
9
8
0
.

+
.
q2
x
..4

!
(
.
x
.

x
2
)=x
1
-2
x
1
X2+x
i

t
h
eP
a
t

emS
e
a
r
c
hs
ot
h
a
tprogramc
o
n
t
r
o
lp n
e
t
e
r
sf
o
rc
o
n
t
o
u
r
6
.
5 Modi
g
o
l
d
e
ns
e
c
t
i
o
n
andupperboundc
a
l
c
u
l
a
t
i
o
ncanbes
e
tby e
r
.
p
l
o
t
t
i
n
g
l
l
o
wf
o
rd
e
f
a
u
l
tv
a
l
u
e
si
nc
a
s
et
h
eu
s
e
rd
e
c
i
d
e
sn
o
tt
o
I
n
c
l
u
d
e promptanda
.
t
t
a
k
ea
d
v
a
n
t
a
g
eofi
r
a
n
s
l
a
t
et
h
ePowe
1methodi
l
n
t
oworkingMATLABc
o
d
e
.V
e
r
i
f
yt
h
es
o
l
u
t
i
o
nt
o
6
.
6 T
Examples6
.
1and5
.
2
.S
t
a
r
tfToms
e
v
e
r
a
lp
o
i
n
t
sandv
e
r
i
f
yt
h
a
tt
h
enumberof
c
y
c
l
e
st
oc
o
n
v
e
r
g
ei
sesame.
e
a
r
c
hd
i
r
e
c
t
i
o
n
sa
r
ec
o
n
j
u
g
a
t
ew
i
t
h s
p
e
c
tt
ot
h
eH
e
s
s
i
a
nm
a
t
r
i
x
e
r
i
f
yes
6
.
7 V
5
.
2
.
f
o
rE
x
a
r
n
p
l
e
s6
.
1 d
tv
e
r
i
f
yanytwos
e
a
r
c
hd
i
r
e
c
t
i
o
n
sa
r
e
o
l
v
eu
s
i
n
gP
o
w
e
l
l
'
smod da
6
.
8 S
ugatewi r
e
s
p
e
c
tt
ot
h
eH
e
s
s
i
a
nm
a
t
r
i
xa
tt
h
es
o
l
u
t
i
o
n
:
co
Minimize

+
.
q2
x
.+4

X2+x
!
(
.
x
.

x
2
)=x
1
-2
x
1
i

R
.P
.A
l
g
o
r
i
t
h
m
sf
o
rM
i
l
l
i
m
i
z
a
t
i
o
nw
i
t
h
o
u
tD
e
r
i
v
a
t
i
v
e
s
t
i
c
e
H
a
l
l
2
.B
r
e
n
t
E
n
g
1
e
w
o
o
dC
l
i
f
f
s
NJ
1
9
7
3
.

3
.Powel
J
.M.J
.D
.
N
O
l
l
l
i
n
e
a
rO
p
t
i
m
i
W
l
i
o
n
A
c
a
d
e
m
i
cP
r
e
s
s
NewY
o
r
k
.1
9
81
.
4
.FoxR
.L
.
.O
p
t
i
m
i
z
a
t
i
o
l
lMethodsforEl
!
g
i
n
e
e
r
i
n
gD
e
s
i
g
n
.A
d
d
i
s
o
n
W
e
s
l
e
y
.R
e
a
d
i
n
g
MA.1971
.

o
l
v
eExample5
.
2u
s
i
n
gt
h
eS
t
e
e
p
e
s
tD
e
s
c
e
n
t
C
o
n
j
u
g
a
t
eG
r
a
d
i
e
n
t
DFP
a
nd
6
.
9 S
BFGSme
o
d
s
.
u
s
i
n
gt
h
eS
t
e
e
p
e
s
t Descen

t Conjugate GradientDFPand BFGS


6
.
1
0S
o
l
v
methods problem

z
e

X2+x
f
(
x
.O
X
2
)=x
1
-2
x
1
i+
.
q2
x
.+4

264 N
U
M
E
R
I
C
A
LT
E
C
H
N
I
O
U
E
SF
O
RU
N
C
O
N
S
T
R
A
I
N
E
DO
P
T
I
M
I
Z
A
T
I
O
N
t
i
o
n6
.
3
.
3u
s
i
n
gac
a
l
c
u
l
a
t
o
r
.V
e
r
i
f
yi
ft
h
em
a
t
r
i
xAa
t
6
.
1
1V
e
r
i
f
yt
h
ev
a
l
u
e
si
nSc
es
o
l
u
t
i
o
ni
sei
n
v
e
r
s
eo
ft
h
eH
e
s
s
i
a
n
.

6
.
1
2V
e
r
i
f
yi
ft
h
em
a
t
r
i
xAa
tt
h
es
o
l
u
t
i
o
ni
st
h
ei
n
v
e
r
s
eo
ft
h
eH
e
s
s
i
a
nf
o
rDFP
me
o
df
o
r
Minimize f
(
X
I
x
2
)=
x
1
-2xtx2+xJ+x~ -2x1+4
6
.
1
3 DevelopeBFGSmethodi
n
t
oMATLABc
o
d
eandv
e
r
i
f
yt
h
ec
a
Jc
u
l
a
t
i
o
n
si
n
S
e
c
t
i
o
n6.
3
.
4
.
6
.
1
4 Developt
h
eM
o
d
i
f
i
e
dNewtonmcthodi
n
t
oMATLABc
o
d
eanda
p
p
l
yi
tt
o
Example5
.
2
.
6
.
1
5 Howw
i
l
lyoui
n
c
o
r
p
o
r
a
t
ees
i
d
ec
o
n
s
t
r
a
i
n
t
si
n
t
ot
h
ecod
f
o
ra
Jl
o
ft
h
ev
a
r
i
o
u
s
methodsi
nt
h
i
ss
e
c
t
i
o
n
?I
m
p
l
e
m
c
n
tthemi
nt
h
cn
u
m
e
r
i
c
a
lp
r
o
c
e
d
u
r
e
s
.

NUMERICALTECHNIQUES
FORCONSTRAINED
OPTIMIZATION

6
.
1
6S
o
l
v
eExamples6
.
2
6
.
3
.a
n
d6.
4byoneo
t
h
e
rmt
h
o
do
ft
h
es
e
c
t
i
o
n
.
6
.
1
7I
d
e
n
t
i
f
yands
o
l
v
eas
y
s
t
e
mdynamicsproblemu
s
i
n
ga
n
ymethodo
f
t
h
i
ss
e
c
t
i
o
n
.
6
.
1
8I
d
e
n
t
i
f
ya
n
ds
o
l
v
cy
o
u
rownc
u
r
v
ef
i
tp
r
o
b
l
c
m
.

l
i
sc
h
a
p
t
e
re
x
p
l
o
r
e
sa
Jg
o
r
i
t
h
m
s
/
m
e
t
h
o
d
st
h
a
th
a
n
d
l
e eg
c
n
e
r
a
lo
p
t
i
m
i
z
a
t
i
o
n
t
yandi
n
e
q
ua
Ji
t
yc
o
n
s
t
r
a
i
n
t
sa
r
ei
n
c
l
u
d
e
d
.F
o
ra
ne
n
g
i
n
e
e
r
i
n
g
p
r
o
b
l
e
m
.Bothe
q
ui

p
r
o
b
l
m t
h
i
sw
i
l
li
n
v
o
l
v
eb
r
i
n
g
i
n
gs
e
v
e
r
a
ln
o
n
l
i
n
e
a
rr
e
l
a
t
i
o
n
si
n
t
ot
h
ed
e
s
i
g
ns
p
a
c
e
.
T
h
i
si
sc
e
r
t
a
i
nt
oi
a
s
et
h
ed
e
g
r
e
eo
fd
i
f
f
i
c
u
l
t
yi
no
b
t
a
i
n
i
n
gcs
o
l
u
t
i
o
n
.F
o
rt
h
e
r
ei
sa
na
d
d
i
t
i
o
n
a
lb
u
r
d
e
no
fb
e
i
n
gmorea
t
t
e
n
t
i
v
et
ot
h
ed
e
s
i
g
nc
h
a
n
g
e
s
d
e
s
i
g
n
e
re
andt
h
ec
o
r
r
e
s
p
o
n
d
i
n
gnumberst
oc
o
a
xt
h
es
o
l
u
t
i
o
ni
ft
h
em
a
t
h
e
m
a
t
i
c
a
ld
e
f
i
n
i
t
i
o
no
f
t
h
ep
r
o
b
l
e
mi
sp
a
r
l
i
c
u
l
a
r
l
ys
e
v
e
r
e
.I
na
Jl
o
fl
h
e
s
ep
r
o
b
l
e
m
sl
h
e
r
ea
r
etwoo
u
t
c
o
m
e
st
h
a
t
a
l
g
o
r
i
t
h
m
ss
e
e
kt
oaccomp
i

s
h
. l
ef
i
r
s
ti
st
oe
n
s
u
r
ea
tt
h
ed
e
s
i
g
ni
sf
e
a
s
i
b
l
e
(
s
a
t
i
saJlc
o
n
s
t
r
a
i
n
t
s
)andt
h
es
e
c
o
n
d ti
ti
so
p
t
i
m
a
l(
s
a
t
i
s
f
i
e
st
h
eKuhn
l
c
k
e
r
c
o
n
d
i
t
i
o
n
s
)
.Wh
i
l
et
h
ef
o
c
u
si
sond
e
t
e
r
m
i
n
i
n
ges
o
l
u
t
i
o
n
i
nt
i
m
e
so
fd
i
f
f
i
c
u
l
t
y
i
ti
s
e
s
s
e
n
t
i
a
l1
0r
ememberI
h
a
t
f
e
a
s
i
b
i
l
i
i
smorei
m
p
o
r
t
a
n
ta
to
p
t
i
m
a
l
i
t
y
.Alsou
n
s
t
a
t
e
d
e
s
to
fec
h
a
p
t
e
ri
sa
teo
p
t
i
m
a
ls
o
l
u
t
i
o
nmustbef
e
a
s
i
b
l
e
.Twod
i
s
t
i
n
c
t
i
n r
a
p
p
r
o
a
c
h
e
swi
1
1beusedt
oh
a
n
d
l
et
h
ec
o
n
s
t
r
a
i
n
e
do
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
e
m
.Th
ef
i
r
s
t
a
p
p
r
o
a
c
hi
st
e
r
m
e
dei
n
d
i
r
e
c
ta
p
p
r
o
a
c
ha
n
ds
o
l
v
e
s problembyt
r
a
n
s
f
o
r
m
i
n
gi
t
a
n
d
l
eec
o
n
s i
n
t
s
i
n
t
oa
nu
n
c
o
n
s
t
r
a
i
n
e
dp
r
o
b
l
e
m
.Thes
e
c
o
n
da
p
p
r
o
a
c
hi
s10h
w
i
t
h
o
u
l n
s
f
o
r
m
a
t
i
o
n
t
h
ed
i
r
e
c
ta
p
p
r
o
a
c
h
.
Thei
n
d
i
r
e
c
ta
p
p
r
o
a
c
hi
sa
ne
x
p
r
e
s
s
i
o
no
fi
n
c
r
e
m
e
n
t
a
ld
e
v
e
l
o
p
m
e
n
to
ft
h
es
u
b
j
e
c
t
u
r
r
e
n
tm
e
t
h
o
d
s
.F
o
re
x n
p
l
e
i
tl
e
v
e
r
a
g
e
st
h
eDFPmethod
t
ot
a
k
ea
d
v
a
n
t
a
g
eo
f ec
t
oh
a
n
d
l
ec
o
n
s
t
r
a
i
n
e
do
p
t
i
m
a
lp
r
o
b
l
e
m
s
.Twoi
n
d
i
r
e
c
tmethodsa
r
ep
r
e
s
e
n
t
e
d
.t
h

E
x
t
e
r
i
o
rP
e
n
a
l
t
yF
u
n
c
t
i
o
n(
E
P
F
)method
a
n
d AugmentedLa
g
r
a
n
g
eM
u
l
t
i
p
l
i
e
r
(ALM)m
e
t
h
o
d
.Th
ed
i
r
e
c
ta
p
p
r
o
a
c
hh
a
n
d
l
e
st
h
ec
o
n
s
t
r
a
i
n
t
sandt
h
eo
b
j
e
c
t
iv
et
o
g
e
t
h
e
r
wioutany n
s
f
o
r
m
a
t
i
o
n
.Fourmethods p s
e
n
t
e
d
. l
emethodsi
ni
s
b
o
o
k
a
r
e
S
e
q
u
e
n
t
ia
JL
i
n
e
a
rProgramming(
S
L
P
)
S
e
q
u
e
n
t
i
a
lQ
u
a
d
r
a
t
i
cP
r
ogramming(SQP)
G
e
n
e
r
a
l
i
z
e
dReducedG
r
a
d
i
e
n
tMethod(GRG)
andS
e
q
u
e
n
t
ia
JG
r
a
d
i
e
n
tR
e
s
t
o
r
a
t
i
o
n
A
l
g
o
r
i
t
h
m(SGRA).
265

2
6
6 N
U
M
E
R
I
C
A
LT
E
C
H
N
I
Q
U
E
SF
O
RC
O
N
S
T
R
A
I
N
E
DO
P
T
I
M
I
Z
A
T
I
O
N

7
.
1P
R
O
B
L
E
MD
E
F
I
N
I
T
I
O
N 267

7
.
1 PROBLEMDEFINION
F

Thes
t
a
n
d
a
r
df
o
r
m
a
to
ft
h
en
o
n
l
i
n
e
a
rp
r
o
g
r
a
m
m
i
n
gp
r
o
b
l
e
m(
N
L
P
)i
sr
e
p
r
o
d
u
c
e
dh
e
f
o
rc
o
n
v
e
n
i
e
n
c
e
:

M
i
n
i
m
i
z
e

(
7.
1
)

f
(
x
"X
2 Xn)

S
u
b
j
t
t
o
: h
k
(
x
"X
2
'... X
. k=1

2 .
1
n)=O

(
7
.
2
)

g
j
(
X
"X
2
.

..
X
j 1
.2 .
m
n)~ 0

(
7
.
3
)

XISXj~xll

i=I.2.....1
t

(
7.
4
)

9<0

I
nv
e
c
t
o
rn
o
t
a
t
i
o
n

M
i
n
i
m
i
z
e f
(
X
)
. [
X
l
"

(
7
.
5
)

S
u
b
j
e
c
tt
o
: [
b
(
X
)
]
f=0

(
7
.
6
)

[
g
(
X
)
]" S0

(
7
.
7
)

X
1
n
w
S
X
:
;
;
X
"1'

(
7
.
8
)
3

F
o
rt
h
i
sc
h
a
p
t
e
rt
h
ef
o
l
l
o
w
i
n
gi
n
d
i
c
e
sa
r
er
c
s
e
r
v
e
d
:I
t
.n
umbero
fv
a
r
i
a
b
l
;
1
.n
u
m
b
e
r
o
fe
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
l
s
;m
.n
u
m
b
e
ro
fi
n
e
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
t
s
.Manyo
fI
h
ea
l
g
o
r
i
t
h
m
s
a
n
dm
e
t
h
o
d
sp
r
e
s
e
n
tar
e
a
s
o
n
a
b
l
el
e
v
e
lo
fd
i
f
f
i
c
u
l
t
y
.A t
w
o
v
a
r
i
a
b
l
ep
r
o
b
1
e
m(
f
o
r
g
r
a
p
h
i
c
a
li
1
lu
s
t
r
a
t
i
o
n
)
.w
i
t
has
i
n
g
l
ee
q
u
a
l
i
t
ya
n
di
n
e
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
ti
su
s
e
dt
o
i
l
l
u
s
t
r
a
t
et
h
ev
a
r
i
o
u
sa
l
g
o
r
i
m
s
.Al
1o
ft
h
ef
u
n
c
t
i
o
n
sa
r
en
o
n
l
i
n
e
a
ra
1
t
h
o
u
g
ht
h
e
c
o
n
s
t
r
a
i
n
tf
u
n
c
t
i
o
n
sh
a
v
es
i
m
p
l
eg
e
o
m
e
t
r
i
cs
h
a
p
e
s
.
7
.
1
.
1 ProblemFormuJatlon-Example7
.
1
r
e
s
e
n
ta
n
y
E
x
a
m
p
l
e7
.
1i
s as
i
m
p
l
em
a
t
h
e
m
a
t
i
c
a
lf
o
r
m
u
l
a
t
i
o
n
.I
td
o
e
sn
o
t p
e
n
g
i
n
e
e
r
i
n
gp
r
o
b
1
e
m
.I
ti
sc
o
n
s
c
t
e
dt
oh
a
v
eas
o
l
u
t
i
o
no
fx
j=1 dx
i=1f
o
r
nstrainedprob1emandx
j=0
.
8
5
2
0a
n
dx
i=0.8520f
o
rt
h
eu
n
c
o
n
s
i
n
e
dp
r
o
b
1
e
m
.
Thet
w
oc
o
n
s
t
r
a
i
n
tf
u
n
c
t
i
o
n
sh
a
v
es
i
m
p
l
en
o
n
1
i
n
e
a
r
i
t
y
.The u
a
l
i
t
yc
o
n
s
t
r
a
i
n
ti
sa
c
i
l
ew
h
i
l
et
h
ei
n
e
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
ti
sa
ne

l
ip
s
e
.
M
i
n
i
m
i
z
e f(x"X
2
)
: x1-2x ~X2 +x
r+X
Ix~ -2x1
+4
S
u
b
j
e
c
l
l
o
: h
(
X
"
X
2
)
: xr+xi-2=0

(
7
.
9
)

O
a
)
(
7
.

g
(
x
"X
2
)
:0
.
2
5x
T+0
.
7
5x~ -1S0

(
7
.l
O
b
)

O:
5x
l:
55
; 0:
5x
2:
55

(
7
.
O
c
)

F
1g
u
r
e7
.
1G
r
a
p
h
i
c
s
o
l
u
1
i
o
nf
o
rE
x
a
m
p
l
e7
.
1
.

r
o
b
l
e
m
.I
ti
sc
l
e
a
rf
r
o
mt
h
ef
i
g
u
r
e
F
i
g
u
r
e7
.
1d
i
s
p
l
a
y
st
h
eg
r
a
p
h
i
c
a
ls
o
l
u
t
i
o
nt
o p
t
h
a
tt
h
es
o
l
u
t
i
o
ni
sa
tx
T=1
.x
i=1(why?).Thevalueoft
h
ef
u
n
c
t
i
o
ni
s3
.A s
i
m
p
l
e
a
r
o
a
c
ht
ot
h
es
o
l
u
t
i
o
no
ft
h
ep
r
o
b
l
e
mi
s1
0u
u
a
t
i
o
n(
7.
l
O
a
)1
0e
l
i
m
i
n
a
t
eo
n
eo
fe
v
a
r
i
a
b
l
e
smep
r
o
b
l
e
me
r
e
b
yr
e
d
u
c
i
n
gi
tt
oas
i
n
g
l
e

v
a
r
i
a
b
l
ec
o
n
s
t
r
a
i
n
e
dp
r
o
b
l
e
m
.
l
i
si
sn
o
td
o
n
eh
e
b
u
ti
ti
sag
o
o
de
x
e
r
c
i
s
e
.F
I
g
7
_l.m1w
i
l
lc
r
e
a
t
et
h
ef
i
g
u
r
e
.
7
.
1
.
2 NecessaryCondltions
l
e

n
e
c
e
s
s
a
r
yc
o
n
d
i
t
i
o
n
o
rt
h
e Kuhn-T
u
c
k
e
r(
K
T
)c
o
n
d
i
t
i
o
n
sf
o
r e
problem-wasd
e
v
c
l
o
p
e
di
nC
h
a
p
t
e
r4
.I
ti
sb
a
s
e
dont
h
em
i
n
i
m
i
z
a
t
i
o
no
fE
L
ag
r
a
n
g
i
a
nf
u
n
c
t
i
o
n(

.Thec
o
n
d
i
l
i
o
n
sa
r
e

Min
i
m
i
z
e
F
(
x
l.'
X
I.
.
.A
f
's
I.
.
.s
f
(
xp xn)+Ihl++
+
m)=
n'A

P
1
+ +sg
"

"

1
F
t
!
1
0bed
o
w
n
l
o
a
d
e
df
r
o
m ew
e
b
s
i
a
r
ei
n
d
i
c
a
l
e
db
yb
o
l
d
f
a
c
es
a
n
ss
e
r
i
fI
y
p
e
.

(
7
.
1
1
)

A4

d81.

. d
h

xi+xi-2=0

(
7
.l
7c
)

E
qu
a
t
i
o
ns
t(
7.
17
)i
ss
o
l
v
e
ds
y
m
b
o
l
i
c
a
l
l
yi
nMATLAB.T
h
c
r
ea
r
etwos
o
l
u
t
i
o
n
st
h
a
!

.11 d
8m

(
7
.
1
2
)

x
;=1.4142; =0
;
x=0.9275; x
;=1.0676;=-0.1217

(
7.
1
3
)

n
e
q
u
a
l
i
l
yc
o
n
s
t
r
a
i
n
t
.I
nF
i
g
u
r
e7
.
1
.t
h
ef
i
r
s
ts
o
l
u
t
i
o
n
8
0
t
ho
ft
h
e
s
ev
i
o
l
a
t
eei
b
u
tl
h
es
e
c
o
n
di
sal
i
l
t
l
ed
i
f
f
i
c
u
l
t1
0s
e
e
.Zo
omings
h
o
u
l
dh
e
l
p
.Casea
d
e
f
i
n
i
t
e
l
ydo
d
o
c
sn
o
tp
r
o
v
i
d
eas
o
l
u
t
i
o
n
.

j=1
.
2
.
.
.
.
.n

X
;=0;

1e
q
u
a
t
i
o
n
sa
r
eo
b
t
a
i
n
e
dd
i
r
e
c
t
l
yt
1
l
f
Oughee
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
l
s

.
2
.
.
.
.
.
1
x
hi
X
2
.
.
.
..xn)=O; k=1
m
mc
q
u
a
t
i
o
n
sa
r
ea
p
p
l
i
c
dt
h
r
o
u
g
ht
h
e2
c
a
s
e
s

s
j
8
j=Oi
f=0

O
N 269
PROBLEMD
E
F
I
N'

i
n
t
s(
7.
l0
c
)
s
a
t
i
s
f
yt
h
es
i
d
cc
o
n
s

=~J + + + L - 4 1+ +P
JS7=0;

a
x
i axaIaxttaxi ta
x
i
d
x
t
0

L/

t
t
oecons
a
i
n
t
s
.l
e
a
r
en+1+munknownsi
nt
h
eL
a
g
r
a
n
g
i
a
n
.Th
esame
s
u
b
j
numbero
fe
q
u
a
t
i
o
n
si
sr
e
q
u
I
I
t
os
o
l
v
e p
r
o
b
l
e
m
.Th
e
s
ea
r
ep
r
o
v
i
d
e
dbyt
h
eFOC
o
rKuhn-Tuckerc
o
n
d
i
t
i
o
n
s
.H
e
r
e
'
le
q
u
a
t
i
o
n
sa
r
eo
b
t
a
i
n
e
da
s

dF d
r d
h
.

7
.
1

U
M
E
R
I
C
A
LT
E
C
H
N
I
O
U
E
SF
O
C
O
N
S
T
R
A
'
N
E
DOP
T
lM
I
Z
A
T
I
O
N
268 N

Caseb
: g=O
t
h
e
n 8j<0

h
e
n
i
f8j=0 l

>0

l
i
se
s
t
a
b
l
i
s
h
e
sf
o
u
re
q
u
a
t
i
o
n
sa
s

4xt-4x +2x -2+x~+2 x +0.


5sX1=0
(
7.
14
)

(
7
.
1
8
a
)

-2xf+2xx2+2X2+1
.5~=O

(
7
.
1
8
b
)

xi+x~ -2=0

(
7
.
1
8
c
)

0
.
2
5x~+O x~ -1=0

(
7
.
1
8
d
)

TheKTc
o
n
d
i
t
i
o
n
sa
p
p
l
yo
n
l
yi
ft
h
ep
o
i
n
t
sa
r
er
e
g
u
l
a
r(
s
e
eExample4
.
3
)
.

A
p
p
l
l
c
a
t
i
o
n0't
h
eKTC
o
n
d
i
t
i
o
n
s
: eLagrangian
F(x"x2 ):

x~ -2xTx2+xT+xx~ 2
x
+4+

x
i+0.75~ -1)

(
7
.
15
)

Th
r
e
ee
q
u
a
t
i
o
n
sa
r
es
e
tupt
h
r
o
u
g
h
(
7
.
1
6
a
)

dF

E;=44-4xtX2+2XI+xi-2+2hi+05PXl=0

dF

= 2xT+2xx2+2i
.x

2+1
.5sX2=0
C
1
X2

h=xi+x~ -2

(
7
.
1
6
b
)

(
7
.
16
c
)

s=O

Witht
h
i
si
n
f
o
r
m
a
t
i
o
nE
q
u
a
t
i
o
n
s(
7.
16
)become

4x1-4Xx2+2x 1 x~-2+2 x =O

(
7.
17
a
)

(
7
.
1
7
b
)

-2xT+2x x2+2
x
2=0

E
q
u
a
t
i
o
n
s(
7.
18
)
n
t
ss
c
o
n
s
t
r

s
o
l
v s
y
m
b
o
l
i
c
a
l
l
yandl
h
eo
n
l
ys
o
l
u
t
i
o
ns
a
t
i
s
f
y
i
n
gI
h
es
i
d
e

x
i=1
.
0
; =1
.0
;

i
.
.
"=-0.
5
0
0
0
; s"=O

whichwase
x
p
e
c
t
e
d(
f
o
rt
h
ed
e
s
i
g
nv
a
r
i
a
b
l
c
sn
o
tl
h
em
u
l
t
i
p
l
i
c
r
s
)
.Thev
a
l
u
co
fs
"=0
i
sn
o
te
x
p
e
c
l
e
da
sl
h
i
si
sa
na
c
t
i
v
ec
o
n
s
t
r
a
i
nt
.TheKTc
o
n
d
i
t
i
o
n
sc
a
nb
ec
o
n
s
i
d
e
r
e
d1
0
b
ew
e
a
k
l
ys
a
l
i
s
f
i
e
da
ssi
sn
o
ts
t
r
i
ct
1yp
o
s
i
t
i
v
e
.Asi
nC
h
a
p
t
e
r6
.w
h
e
r
e
v
e
rp
o
s
s
i
b
l
e
.
eKTc
o
n
d
i
t
i
o
n
sa
r
eu
s
e
dt
oc
h
e
c
kf
o
rc
o
n
v
e
r
g
e
n
c
e
.

7
.
1
.
3 ElementsofaNumericalTechnlque
T
h
e
r
ei
sad
i
s
t
i
n
c
Id
ie
r
e
n
c
ei
nt
h
en
u
m
e
r
i
c
a
lt
e
c
h
n
i
q
u
e
sw
i
t
hi
n
c
o
r
p
o
r
a
t
i
o
no
ft
h
e
.
lThemajor
c
o
n
s
t
r
a
i
n
l
S
.e
v
e
ni
fl
h
ei
d
e
ao
fu
s
i
n
gas
e
a
r
c
hd
i
r
e
c
t
i
o
nv
e
c
l
o
ri
ss
t
i
l
lu
s
e
f
u.
d
i
f
f
e
r
e
n
c
e
si
nt
h
ea
1gorithmsoft
h
ep
r
e
v
i
o
u
sc
h
a
p
t
e
rwas wayt
h
es
e
a
r
c
hd
i
r
e
c
t
i
o
n
s
w
e
r
ee
s
t
a
b
l
i
s
h
e
d
.Tou
n
d
e
r
s
t
a
n
dt
h
en
e
c
e
s
s
a
r
yc
h
a
n
g
e
s
.t
h
eg
e
n
e
r
a
la
p
p
r
o
a
c
hf
o
r
u
n
c
o
n
s
t
r
a
i
n
e
do
p
t
i
m
i
z
a
t
i
o
na
ss
e
e
ni
nt
h
ep
r
e
v
i
o
u
sc
h
a
p
t
e
rw
a
s
:

Thef
o
u
r
l
he
q
u
a
t
i
o
ni
sdev
e
10pedt
h
r
o
u
g
h
Casea
:

; t i -a-

( +~-2)+ (0.25

GeneralAlgorlthm6
.
1
)
S
t
e
p1
. ChooseX
o

270

T
lO
N 2
7
1
7
.
21
N
D
I
R
E
C
TMETHODSFORC
O
N
S
T
R
A
I
N
E
DO
P
T
I
M
I
Z
A

N
U
M
E
R
I
C
A
LT
E
C
H
N
I
O
U
E
SFORC
O
N
S
T
R
A
I
N
E
DO
P
T
I
M
I
Z
A
T
I
O
N

S
t
e
p2
.F
o
re
a
c
hi
t
e
r
a
t
i
o
ni
D
e
t
e
n
n
i
n
es
e
a
r
c
hd
i
r
e
c
t
i
o
nv
e
c
t
o
rS
j
(
I
twouldb
en
i
c
ei
ft
h
eo
b
j
e
c
t
i
v
ed
e
c
r
e
a
s
e
da
l
o
n
gt
h
i
sd
i
r
e
c
t
i
o
n
)

S
t
e
p3
.C
a
l
c
u
l
a
t
eAXj
= S
j
N
o
t
e
:AXj
i
snowaf
u
n
c
t
i
o
no
ft
h
es
c
a
l
a
r(Xj
a
sS
ji
sknownfromS
t
e
p2
j
(X
j
i
sc
a
l
l
e
des
t
e
ps
i
z
ea
si
te
s
t
a
b
l
i
s
h
e
sel
e
n
g
t
hofAX

a
;i
sd
e
t
e
r
r
n
i
n
e
dbyM
i
n
i
m
i
z
i
n
g
f
(
X
;
+
I
)
'where
X
j
+
1=X +AXj

r
t44j532

e
v
i
e
w
e
d
i
nS
t
e
p2adilemman
d
st
ober
e
s
o
l
v
e
d
:S
h
o
u
l
d
Ast
h
ea
b
o
v
es
t
e
p
s er
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
n0's
h
o
u
l
di
ta
t
t
e
m
p
tt
os
a
t
i
s
f
y
t
h
es
e
a
r
c
hd
i
r
e
c
t
i
o
nS
jd
e
c
r
e
a
s
eeo
t
h
ec
o
n
s
t
r
a
i
n
t
s
?(
I
ti
sassumeda
ti
tc
a
n
n
o
tdob
o
t
ht
h
o
u
g
hv
e
r
yo
f
t
e
ni
tw
i
ll
.
)S
i
n
c
e
f
e
a
s
i
b
i
l
i
t
yi
smorei
m
p
o
r
t
a
n
ta
no
p
t
i
m
a
l
i
t
yi
ti
s s
i
e
rt
oi
d
e
n
t
i
f
yt
h
el
a
t
t
e
ra
sa
s
i
g
n
i
f
i
c
a
n
ti
n
f
l
u
e
n
c
ei
nd
e
t
e
r
r
n
i
n
i
n
gt
h
ed
i
r
e
c
t
i
o
n
.T
h
i
si
smored
i
f
f
i
c
uI
tt
h
a
ni
tsounds
a
se
a
c
hc
o
n
s
t
r
a
i
n
tc
a
nh
a
v
ei
t
sownf
a
v
o
r
i
t
ed
i
r
e
c
t
i
o
na
tt
h
ec
u
r
r
e
n
tp
o
i
n
.
tSuch
o
c
c
a
s
i
o
n
sc
a
l
lf
o
rat
r
a
d
e
o
f
fandt
os
e
ei
fi
ti
sp
o
s
s
i
b
l
et
on
e
g
o
t
i
a
t
ead
e
c
r
e
a
s
ei
nt
h
e
ti
sc
l
e
a
rt
h
a
tS
t
e
p2w
i
l
li
n
v
o
l
v
ei
n
v
e
s
t
m
e
n
to
fe
f
f
o
r
tand
o
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
na
sw
e
l
l
.I
somes
o
p
h
i
s
t
i
c
a
t
i
o
n
.
sn
o
tf
a
rb
e
h
i
n
di
nt
h
e q
u
i mentf
o
r
s
p
i
a
lh
a
n
d
l
i
n
g
.On
c
et
h
ed
i
r
e
c
t
i
o
n
S
t
e
p3i
i
se
s
t
a
b
l
i
s
h
e
d(
S
t
e
p2
)whatk
i
n
do
fs
t
e
p
s
i
z
ew
i
l
lbea
c
c
e
p
t
a
b
l
e
1S
e
v
e
r
a
lr
e
s
u
l
t
sa
r
e
p
o
s
s
i
b
l
c
.F
i
r
s
t
t
h
co
b
j
c
c
t
i
v
ef
u
n
c
t
i
o
nc
a
nb
d
e
c
r
e
a
s
e
da
l
o
n
gt
h
ed
i
r
e
c
t
i
o
n
.Second
c
u
r
r
e
n
ta
c
t
i
v
ec
o
n
s
t
r
a
i
n
t
s(
i
n
c
l
u
d
i
n
g e
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
t
s
)c
a
nbecomei
n
a
c
t
i
v
e
v
i
o
l
a
t
e
d
o
rc
a
ns
t
i
l
lp
s
e
r
v
e ea
c
t
i
v
cs
t
a
t
e
.T
h
i
r
d
c
u
r
r
e
n
ti
n
a
c
t
i
v
ec
o
n
s
t
r
a
i
n
t
sc
a
n
u
n
d
e
r
g
oas
i
m
i
l
a
rc
h
a
n
g
e0"
1s
t
a
t
e
.F
o
u
r
t
h
p
r
o
b
a
b
l
ymosti
m
p
o
r
t
a
n
t

c
u
r
r
e
n
tv
i
o
l
a
t
e
d
c
o
n
sa
i
n
t
sc
a
nbecomea
c
t
i
v
eo
ri
n
a
c
t
i
v
e
.I
ti
sd
i
f
f
i
c
u
l
tt
oe
n
c
a
p
s
u
l
a
t
et
h
i
sd
i
s
c
u
s
s
i
o
n
i
nag
e
n
e
r
i
ca
l
g
o
r
i
t
h
m
.Somca
l
g
o
r
i
t
h
m
scombineS
t
e
p
s2and3i
n
t
oas
i
n
g
l
es
t
e
p
.
Somed
i
v
i
d
et
h
ei
t
e
r
a
t
i
o
ni
n
t
oaf
e
a
s
i
b
i
l
i
t
ycomponenlanda
no
p
t
i
m
a
l
i
t
ycomponen
.
t
s
s
e
n
c
et
h
eg
e
n
e
r
i
ca
1g
o
r
i
t
h
mf
o
r i
sc
h
a
p
t
e
r

whichi
sn
o
tp
a
r
t
i
c
u
l
a
r
l
y e
f
u
l
i
s
:
I
ne

i
t
e
r
a
t
i
o
n
.l
eu
n
c
o
n
s
a
i
n
e
da
l
g
o
r
i
t
h
mi
t
s
e
l
fe
x
e
c
u
t
e
smanyi
t
e
r
a
t
i
o
n
s
.T
h
i
swould
u
i ar
o
b
u
s
tu
n
c
o
n
s
t
r
a
i
n
e
dm
i
n
i
m
i
z
e
rt
oh
a
n
d
l
eal
a
r
g
ec
l
a
s
so
fp
r
o
b
e
m
s
.Th
e
r
BFGSmethodi
sr
o
b
u
s
tandi
m
p
r
e
s
s
i
v
eo
v
e
ral
a
r
g
ec
l
a
s
so
fp
r
o
b
l
e
m
s
.
f
o
r
m
i
n
gt
h
ec
o
n
s
t
r
a
i
n
e
d problem i
n
t
oa
n
Ap
r
e
p
r
o
c
e
s
s
i
n
gl
a
s
ki
n
v
o
l
v
e
st
r
u
n
c
o
n
s
t
r
a
i
n
e
dp
r
o
b
l
e
m
.T
h
i
si
sl
a
r
g
e
l
ya
c
c
o
m
p
l
i
s
h
e
dbya
u
g
m
e
n
t
i
n
gt
h
eo
b
j
e
c
t
i
v
e
f
u
n
c
t
i
o
nw
i
t
ha
d
d
i
t
i
o
n
a
lf
u
n
c
t
i
o
n
s r
e
f
l
e
c
t ev
i
o
l
a
t
i
o
no
ft
h
en
s
t
r
a
i
n
t
sv
e
r
y
s
i
g
n
i
f
i
c
a
n
t
l
y
.These n
c
t
i
o
n
sa
r
er
e
f
e
r
r
e
dt
oa
sep
e
n
a
l

f
c
t
i
o
n
s
.l
e
W
s
i
g
n
i
f
i
c
a
n
ta
c
t
i
v
i
t
yi
ni
s a
tonet
i
m
ewhichl
e
dt
oe
n
t
i
r
ef
a
r
n
i
l
i
e
so
fd
i
f
f
e
r
e
n
t
t
y
p
e
so
fp
e
n
a
l
l
yf
u
n
c
t
i
o
nmethods[
2
)
.Thef
i
r
s
lo
ft
h
e
s
ewast
h
eE
x
1
e
r
i
o
rP
e
n
a
l
t
y
F
u
n
c
t
i
o
n(
E
P
F
)m
e
t
h
o
d
. Th
ef
i
r
s
tv
e
r
s
i
o
no
f ANSYS [
3
] i
n
c
o
r
p
o
r
a
1
e
d
o
p
t
i
m
i
z
a
t
i
o
ni
ni
t
sf
i
n
i
t
ee
l
e
m
e
n
lprogramr
e
l
i
e
dont
h
eEP
F
.TheEPF(
p
r
e
s
e
n
t
e
di
n
en
e
x
ts
e
c
t
i
o
n
)hads
e
v
e
r
a
ls
h
o
r
t
c
o
m
i
n
g
s
.Toa
d
d
r
e
s
st
h
o
s
e eI
n
t
e
r
i
o
rP
e
n
a
l
t
y
F
u
n
c
t
i
o
n(
I
P
F
)methodswered
e
v
e
l
o
p
e
dl
e
a
d
i
n
g1
0t
h
eV
a
r
i
a
b
l
eP
e
n
a
l
t
yF
u
n
c
t
i
o
n
(VPF)m
e
t
h
o
d
s
.I
n i
st
e
x

.
to
n
l
yt
h
eEPFi
sa
d
d
r
e
s
s
e
dl
a
r
g
e
l
yduet
oa
c
a
d
e
m
i
c
.
tI
nviewo
fee
x
c
e
l
l
e
n
tp
e
r
f
o
r
m
a
n
c
eo
f d
i
r
e
c
tmethodst
h
e
s
emethods
i
n
t
e
r
e
s
w
i
l
lp
r
o
b
a
b
l
yn
o
tbeu
s
c
dt
o
d
a
yf
o
rc
o
n
t
i
n
u
o
u
sp
r
o
b
l
e
m
s
.Theya
r
concea
g
a
i
n
i
m
p
o
r
t
a
n
ti
ng
l
o
b
a
lo
p
t
i
m
i
z
a
t
i
o
r
lt
e
c
h
n
i
q
u
e
sf
o
rc
o
n
s
t
r
a
i
n
e
dp
r
o
b
l
e
m
s
.Thes
e
c
o
n
d
me
odp
s
e
n
t
e
di
nt
h
i
ss
e
c
t
i
o
n
l
h
eAugmentedLa
g
r
a
n
g
eMethod(ALM)
i
sl
h
eb
e
s
l
o
ft
h
eSUMT.l
t
se
x
c
e
e
d
i
n
g
l
ys
i
m
p
l
ei
m
p
l
e
m
e
n
t
a
t
i
o
n
i
t
sq
u
a
l
i
t
yo
fs
o
l
u
t
i
o
n
andi
t
s
1
i
e
r
sa
l
l
o
wi
tt
os
e
r
i
o
u
s
l
y
a
b
i
l
i
t
yt
og
e
n
e
r
a
t
ei
n
f
o
n
n
a
t
i
o
nont
h
eL
a
g
r
a
n
g
em
u
l
t
i
p
c
h
a
l
l
e
n
g
e d
i
r
e
c
tt
e
c
h
n
i
q
u
e
s
.

Method
7
.
2
.
1 E
x
t
e
r
l
o
rPenaltyFunctlon(EP
Thet
r
a
n
s
f
o
r
m
a
t
i
o
no
ft
h
co
p
t
i
m
i
z
a
t
i
o
nproblem(
7
.1
)
(
7.
4
)1
0a
nu
n
c
o
n
s
t
r
a
i
n
e
d
problemi
smadep
o
s
s
i
b
l
et
h
r
o
u
g
hap
e
n
a
l
t
yf
u
n
c
t
i
o
nf
o
n
n
u
l
a
t
i
o
n
.Thet
r
a
n
s
f
o
n
n
e
d
u
n
c
o
n
s
t
r
a
i
n
e
dproblemi
s
:
Minimize

GenerlcAlgorlthm(
A
7
.
1
)

S
t
e
p1
. Ch
seXo
S
t
e
p2
.F
o
re
a
c
hi
t
e
r
a
t
i
o
ni
D
e
t
e
n
n
i
n
es
e
a
r
c
hd
i
r
e
c
t
i
o
nv
e
c
t
o
rS
j

S
t
e
p3
.C
a
l
c
u
l
a
t
eAX;=jSj
Xj
+1
=Xj
+AXj

F(X'
1
1
''
g
)=f
(
X
)+p
(
X'
1
1
'

X::
S
;X
/:
S
;

The
s
emco
d
swered
e
v
e
l
o
p
e
dt
ot
a
k
ea
d
v
a
n
t
a
g
eo
fc
o
d
e
s a
ts
o
l
v
eu
n
c
o
n
s
t
r
a
i
n
e
d
o
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
e
m
s
. They a
r
ea
l
s
or
e
f
e
r
r
e
dt
oa
sS
e
q
u
e
n
t
i
a
lU
n
c
o
n
s
t
r
a
i
n
e
d
M
i
n
i
m
i
z
a
t
i
o
nT
e
c
h
n
i
q
u
e
s(SUMT)(
1
)
.Thei
d
e
ab
e
h
i
n
dea
p
p
r
o
a
c
hi
st
or
e
p
e
a
t
e
d
l
y
i
1
/e
do
p
t
i
m
i
z
a
t
i
o
1
/a
l
g
o
r
i
t
h
mu
s
i
n
gt
h
es
o
l
u
t
i
o
no
ft
h
ep
r
e
v
i
o
u
s
c
a
l
lt
h
eu
n
c
o
n
s
ta

(
7.
4
)

whereP(X'
h
'
g
)i
st
h
ep
e
n
a
l
t
yf
u
n
c
t
i
o
n
.'
1
1a
n
d'
ga
r
ep
e
n
a
l
t
yc
o
n
s
t
a
n
t
s(
a
l
s
oc
a
l
l
e
d
J
l
i
es
.
)
m
u
l
t
Thep
e
n
a
l
t
yf
u
n
c
t
i
o
ni
se
x
p
r
e
s
s
e
d

L
.

7
.
2 INDIRECTMETHODSFORCONSTRAINEDOPTIMIZAT10N

i=1

2

1
/

(
7
.
1
9
)

p
(X
'
h
''

)='h1 "
k
(
X
)
2
1+'
8
1 (m {
O

(X)}
)
2
1
1
"
'
1
l
t
1

(
7
.
2
0
)

I
nEq
u
a
t
i
o
n(
7
.
2
0
)
i
fee
q
u
a
l
i
t
yc
o
n
s
i
n
t
sa
r
en
o
tz
e
r
o
t
h
e
i
rv
a
l
u
eg
e
t
ss
q
u
a
r
e
dand
m
u
l
t
i
p
l
i
e
dbyt
h
ep
e
n
a
l
t
ym
u
l
t
i
p
l
i
e
randt
h
e
ng
e
t
saddedt
o eo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
n
.I
f
t
h
i
n
e
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
ti
si
nv
i
o
l
a
t
i
o
n
i
tt
o
og
e
t
ss
q
u
a
r
e
danda
d
d
e
d1
0 o
b
j
e
c
t
i
v
e
e
n
a
l
t
ym
u
l
t
i
p
l
i
e
r
s
.I
nas
e
n
s
e
i
ft
h
ec
o
n
s
t
r
a
i
n
t
s
f
u
n
c
t
i
o
na
f
t
e
rb
e
i
n
ga
m
p
l
i
f
i
c
dby p

2
7
2 N
U
M
E
R
I
C
A
LT
E
C
H
N
I
Q
U
E
SF
O
RC
O
N
S
T
R
A
I
N
E
DO
P
T
I
M
I
Z
A
T
I
O
N

7
.
2I
N
D
I
R
E
C
TM
E
T
H
O
D
SF
O
RC
O
N
S
T
R
A
I
N
E
DO
P
T
I
M
I
Z
A
T
I
O
N 2
7
3

a
r
en
o
1s
a
t
i
s
f
i
e
d
t
h
e
ne
y
a
r
e
p
e
ni
z
e
d

h
e
n
c
et
h
ef
u
n
c
t
i
o
n
'
sn
n
e
.I
tc
a
nbes
h
o
w
n
t
h
c l
s
f
o
n
n
c
du
n
ns n problemsolveseoriginalm nedproblema
st
h
e
"
r
ga
p
p
r
o
a
c
h.I
no
r
d
e
rf
o
rPt
or
e
m
a
i
nf
i
n
i
l
ea
tev
a
l
u
e
so
fe
m
u
l
t
i
p
l
i
e
r
sr
m
u
l
t
i
p
l
i
e
r
s(
1
d
c
df
o
rav
a
l
i
ds
o
l
u
t
i
o
n
)ec
o
n
s
t
r
a
i
n
t
sm
u
s
tbes
a
t
i
s
t
i
e
d
.I
nc
o
m
p
u
t
e
r
i
m
p
l
e
m
e
n
t
a
t
i
o
ni
sl
i
m
i
ti
s p
l
a
c
e
db
yal
a
r
g
ev
a
l
u
ei
n
s
a
do
f.Ano
t
h
e
rf
a
c
e
to
f
mputerimplementationoft
h
i
sm
e
t
h
o
di
s al
a
r
g
ev
a
l
u
eo
ft
h
em
u
l
t
i
p
l
i
e
r
sa
tt
h
et
i
l
e
r
a
t
i
o
ni
sb
o
u
n
dt
oc
r
e
a
t
en
u
m
e
r
i
c
a
ld
i
f
f
i
l
t
i
e
s
.l
e
s
e
m
t
i
p
l
i
e
r
a
r
es
t
a
r
t
e
dw
i
t
hsma
l
1
v
a
l
u
e
s du
a
gme
l
l
y c
hi
t
e
r
a
t
i
o
n
. eu
n
c
o
n
s
l
r
a
I
n
c
h
n
i
q
u
e

f
o
r
e
x
DFPwi
1
s
o
l
v
eE
q
u
a
t
i
o
n(
7
.
1
9
)f
o
raknownv
a
l
u
eo
f m
u
l
t
i
p
l
i
c
r
s
.The
c
df
r
o
mt
h
eDFPc
a
nbec
o
n
s
i
d
e
r
a
saf
u
n
c
t
i
o
no
ft
h
cm
u
l
t
i
p
l
i
e
ra
n
dc
a
n
s
o
l
u
t
i
o
n n
bet
h
o
u
g
h
to
fa
s

X.=X
.
(
r
"

r)
8

(
7
.
21
)

TheSUMTi
t
e
r
a
t
i
o
ni
n
v
o
l
v
e
su
p
d
a
t
i
n
gt
h
em
u
l
t
i
p
l
i
e
r
sa
n
dei
n
i
t
i
a
ld
e
s
i
g
nv
e
c
t
o
r

a
n
dc
a
l
l
i
n
geu
n
c
o
n
s
t
r
a
i
n
e
dm
i
n
i
m
i
z
e
ra
g
a
i
n
.I
n a
l
g
o
r
i
t
h
mi
ti
sa
s
s
u
m
e
dt
h
a
te
DFPm
e
t
h
o
di
su
s
e
d(
a
l
t
h
o
u
g
ha
n
ym
e
t
h
o
df
r
o
mC
h
a
p
t
e
r6c
a
nb
eu
s
e
dBFOSi
s
c
o
m
m
e
n
d
e
d
)
.
r
Algorlthm:E
x
t
e
r
l
o
rPenaltyFunction(EP

M
ethod2
)
S
t
e
p1
.ChooseXI

N.(
n
o
.o
fSUMTi
t
e
r
a
t
i
o
n
s
)
Nu(
n
o
.ofDFPi
t
e
r
a
t
i
o
n
s
)

E
;
'
S(
f
o
rc
o
n
v
e
r
g
e
n
c
ea
n
ds
t
o
p
p
i
n
g
)
r~ r
!(
i
n
i
t
i
a
lp
e
n
a
l
t
ym
u
l
t
i
p
l
i
e
r
s
)
C
h
.c
/
I(
s
c
a
l
i
n
gv
a
l
u
ef
o
rm
u
l
t
i
p
l
i
e
r
s
)
q=1(SUMTi
t
e
r
a
t
i
o
nc
o
u
n
t
e
r
)
q
S
t
e
p2
.C
a
l
lDFPt
om
i
n
i
m
i
z
eF(X
r
%

r~)
O
u
t
p
Ul
:x
q
.
S
t
e
p3
.C
o
n
v
e
r
g
e
n
c
ef
o
rEPF
I
fhk=
0f
o
rk=
1

2
..
.
.

1
;
I
fg
j:
5O
.f
o
rj=1
.2
..
.
.
m;
I
fa
l
ls
i
d
ec
o
n
s
t
r
a
i
n
t
sa
r
es
a
t
i
s
f
i
e
d
T
h
e
nC
o
n
v
e
r
g
e
d

S
t
o
p
S
l
o
p
p
i
n
gC
r
i
t
e
r
i
a
:
d
.
F=
Fq F
Xq*
-X(q-I)*
q
-I dX=
I
f(
d
.
F
l:
5E
I
:S
t
o
p(
f
u
n
c
t
i
o
nn
o
tc
h
a
n
g
i
n
g
)
E
l
s
eI
fd
.
XTd
.
X:
5E
I
:S
t
o
p(
d
e
s
i
g
nn
o
tc
h
a
n
g
i
n
g
)
s
el
fq=
N
.
:S
l
O
p(maximumi
t
e
r
a
t
i
o
n
sr
e
a
c
h
e
d
)
C
o
n
t
i
n
u
e
Q -

qq+l
rg r~'Ch;

x
qx
q
.

0
"

r
:r
1

Cg

00t
oS
t
e
p2

s
c
a
l
i
n
g
TheEPFi
sv
e
r
ys
e
n
s
i
t
i
v
et
ot
h
es
t
a
r
t
i
n
gv
a
l
u
eo
ft
h
em
u
l
t
i
p
l
i
e
r
sa
n
dt
ot
h
f
a
c
t
o
r
sa
sw
e
l
l
.D
i
f
f
e
r
e
n
tp
r
o
b
l
e
m
sr
e
s
p
o
n
df
a
v
o
r
a
b
l
yt
od
i
f
f
e
r
e
n
tv
a
l
u
e
so
ft
h
e
r
s
.I
ti
srecommendedt
h
a
tt
h
ei
n
i
t
i
a
lv
a
l
u
e
so
ft
h
em
u
l
t
i
p
l
i
e
r
sb
ec
h
o
s
e
na
s
m
u
l
t
i
p
l
i
l
Io
nt
ot
h
ec
o
r
r
e
s
p
o
n
d
i
n
gt
e
n
ni
n p
e
n
a
l
t
yf
u
n
c
t
i
o
na
1
t
h
er
a
t
i
oo
ft
h
eo
b
j
e
c
t
i
v
ef
u
n
c
tb
o
t
ht
h
eo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
na
n
dt
h
ec
o
n
s
t
r
a
i
n
t
sa
r
e
t
h
ei
n
i
t
i
a
ld
e
s
i
g
n
.T
h
i
se
n
s
u
r
e
sa
n
i
n
i
n
gt
h
ec
h
a
n
g
e
si
nt
h
ed
e
s
i
g
nf
o
r s
u
c
c
e
e
d
i
n
g
e
q
u
a
l
l
yi
m
p
o
r
t
a
n
ti
nd
e
t
e
l
t
e
r
a
t
i
o
n
.
e
r
i
o
rP
e
n
a
l
t
yi
st
h
a
ta
tt
h
ee
n
do
fe
a
c
hSUMTi
t
e
r
a
t
i
o
n
Oner
e
a
s
o
nf
o
rt
h
et
e
n
nExt
d
e
s
i
g
nw
i
l
lb
ei
l
a
s
i
b
l
e(
u
n
t
i
lt
h
es
o
l
u
t
i
o
ni
so
b
t
a
i
n
e
d
)
.T
h
i
si
m
p
l
i
e
st
h
a
tc
r
ea
p
p
r
o
a
c
h
i
n
g ef
e
a
s
i
b
l
er
e
g
i
o
nf
r
o
mt
h
e
m
e
t
h
o
dd
e
t
e
n
n
i
n
e
sd
e
s
i
g
nv
a
l
u
e
s a
o
u
t
s
i
d
e
.Thisi
sas
e
r
i
o
u
sd
r
a
w
b
a
c
ki
ft
h
em
e
t
h
o
df
a
i
l
sp
m
a
t
u
r
e
l
y
a
si
tw
i
l
lo
f
t
e
nd
o
.
Thci
n
f
o
r
m
a
t
i
o
ng
e
n
e
r
a
t
e
ds
of
a
ri
sv
a
l
u
e
l
e
s
sa
st
h
ed
e
s
i
g
n
sw
e
r
cn
e
v
e
rf
e
a
s
i
b
l
e
.As
o
n
l
i
n
e
a
r
i
t
yo
ft
h
ep
r
o
b
l
e
m
s
e
e
ni
nt
h
ee
x
a
m
p
l
eb
e
l
o
w
.t
h
eEPFs
e
v
e
r
e
l
yi
n
c
r
e
a
s
e
sen
o
dt
of
a
i.
lI
li
se
x
p
e
c
t
e
d t
h
ei
n
c
r
e
a
s
ei
n
c
r
e
a
t
i
n
gc
o
n
d
i
t
i
o
n
sf
o
rt
h
eme
o
s
e
rs
t
a
r
t
i
n
gv
a
l
u
ef
o
rt
h
ed
e
s
i
g
na
se
a
c
hSUMT
n
o
n
l
i
n
e
a
r
i
t
yi
sb
a
l
a
n
c
e
db
yac
o
s
e
rt
ot
h
es
o
l
u
t
i
o
nt
h
a
ncp
r
e
v
i
o
u
so
n
e
.
i
t
e
r
a
t
i
o
ns
t
a
r
t
sc
l
nt
h
ef
o
l
l
o
w
i
n
gt
h
eEPFi
sa
p
p
l
i
e
dt
oE
x
a
r
n
p
l
e7
.
1t
h
r
o
u
g
has
e
r
i
e
so
fc
a
l
c
u
l
a
t
i
o
n
s
nt
h
r
o
u
g
ht
h
et
r
a
n
s
l
a
t
i
o
no
fea
l
g
o
r
i
t
h
mi
n
t
oMA
TlABc
o
d
e
.The
r
ea
r
ea
r
a
t
h
e
r a
c
o
u
p
l
eo
fc
h
a
n
g
e
sw
i
t
hr
e
s
p
e
c
tt
oA
l
g
o
r
i
t
h
m(
A
7
.
1
)
.Tor
e
s
o
l
v
et
h
ep
e
n
a
l
t
yf
u
n
c
t
i
o
n
n
e
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n

tt
h
ec
o
n
s
t
r
a
i
n
ti
sa
s
s
u
m
e
dt
oa
l
w
a
y
sb
ei
n
w
i
t
hr
e
s
p
e
c
tt
oei
v
i
o
l
a
t
i
o
n80a
ter
e
t
u
r
nf
r
o
mt
h
em
l
l
xf
u
n
c
t
i
o
ni
s ec
o
n
s
t
r
a
i
n
tf
u
n
c
t
i
o
ni
t
s
e
lf
.T
h
i
s
1
ld
r
i
v
et
h
ei
n
e
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
tt
obea
c
t
i
v
ew
h
i
c
hweknowt
ob
et
r
u
ef
o
rt
h
i
s
wi
e
x
a
m
p
l
e
.N
u
m
e
r
i
c
a
li
m
p
l
e
m
e
n
t
a
t
i
o
na
so
u
t
l
i
n
e
di
nt
h
ea
l
g
o
r
i
t
h
ms
h
o
u
l
da
l
l
o
w
d
e
t
e
n
n
i
n
a
t
i
o
no
fi
n
a
c
t
i
v
ec
o
n
s
t
r
a
i
n
t
s
.I
n
s
t
e
a
do
fn
u
m
e
r
i
c
a
li
m
p
l
e
m
e
n
t
a
t
i
o
no
ft
h
e
na
n
a
l
y
t
i
c
a
ls
o
l
u
t
i
o
ni
sd
e
t
e
r
m
i
n
e
du
s
i
n
gMATLABs
y
m
b
o
l
i
c
u
n
c
o
n
s
t
r
a
i
n
e
dp
r
o
b
l
e
ma
c
o
m
p
u
t
a
t
i
o
n
.Example7
.
1i
sr
e
p
r
o
d
u
c
e
df
o
rc
o
n
v
e
n
i
e
n
c
e
.

M
i
n
i
m
i
z
e f
(
X
j
o
X
2
)
: x1-2xTx2+xT+XIX~-2xl +4 (
7
.
9
S
u
b
j
e
c
t

: h(x( X
2
)
: xT+x~-2=0
g
(
x

( x

O. 25x r+ O x~-1 :
50

o:
5X
l:
55
;

0:
5X
2:
55

(
7
.
l
O
a
)

(
7
.
L
O
b
)
(
7.
l0
c
)

F
i
g
u
r
e7
.
2i
st
h
cc
o
n
t
o
u
rp
l
o
to
ft
h
et
r
a
n
s
f
o
r
m
e
du
n
c
o
n
s
t
r
a
i
n
c
df
u
n
c
t
i
o
nf
o
rv
a
l
u
e
so
f
r
h=
1a
n
d =1
.Thei
n
c
r
e
a
s
ei
nn
o
n
l
i
n
e
a
r
i
t
yi
sr
c
a
d
1
i
ya
p
p
a
r
n
.
tF
i
g
u
r
e7.
3i
st
h
ep
l
o
t
h=
5a
n
dr
g=
5
.H
a
n
d
l
i
n
ge
s
ef
u
n
c
t
i
o
n
sn
u
m
e
r
i
c
al
1
ye
v
e
nw
i
t
ht
h
eBFOSw
il
1
f
o
rr
beac
o
n
s
i
d
e
r
a
b
l
ec
ha
11
e
n
g
e
.B
o
t
ht
h
ef
i
g
u
r
e
ss
u
g
g
e
s
ts
v
e
r
a
lp
o
i
n
t
s a
ts
a
t
i
s
f
yf
i
r
s
t
td
i
f
f
i
c
u
l
tf
o
ra
n
yn
u
m
e
r
i
c
a
lt
e
c
h
n
i
q
u
et
of
i
n
d
o
r
d
e
rc
o
n
d
i
t
i
o
n
s
.T
h
e
i
rc
l
o
s
e
n
e
s
sm i
h
eEPFs
e
v
e
r
e
l
yi
n
c
r
e
a
s
e
st
h
en
o
n
l
i
n
e
a
r
i
t
yo
f
t
h
ep
r
o
b
l
e
m
.
t
h
eo
p
t
i
m
u
m
.
l
ti
sc
l
e
a
r t
plolmcontainst
h
ec
o
d
e tg
e
n
e
r
a
t
e
st
h
ep
l
o
.
tS
i
n
c
et
h
ec
o
d
eu
s
e
s
Sec7_2_1s
y
m
b
o
l
i
cm
a
n
i
p
u
l
a
t
i
o
n
.a
c
t
u
a
l
l
yd
r
a
w
i
n
gt
l
l
ep
l
o
tt
a
k
e
st
i
m
e
.E
v
a
l
u
a
t
i
n
gt
h
ed
a
t
a
D
u
m
e
r
i
c
a
l
l
yw
i
l
lmakeab
i
gd
i
f
f
e
r
e
n
c
e
.

274

7.
2

N
U
M
E
R
I
C
A
LT
E
C
H
N
I
Q
U
E
SFORC
O
N
S
T
R
A
I
N
E
DO
P
T
I
M
I
Z
A
n
O
N

I
N
D
I
R
E
C
TMETHODSFORC
O
N
S
T
R
A
I
N
E
DO
P
T
I
M
I
Z
A
T
I
O
N 275

5
$e
v
e
r
cn
o
n
l
l
n
e
a
r
l
t
yw
l
t
hm
a
n
yl
o
c
o
lm
l
n
l
m
u
m
s

.
.
;
.
.
;
.
.
:

4
1

.
.

..

.'"

~-c

.
.

xt

F
l
g
u
r
e7
.
2 E
o
rP
e
n
a
l
t
yF
u
n
c
l
l
o
nM
e
l
h
o
d
:E
x
a
m
p
l
e7
.
1(
n
.=
1
rg=
1
)
.

F
l
g
u
r
e7
.
3E
x
t
e
r
i
o
rP
e
n
a
l
t
yF
u
n
c
t
i
o
nM
e
t
h
o
d
:E
x
a
m
p
l
e7
.
1(
n
.=5
.fg=5
)
.

'
0

A
p
p
l
l
c
a
t
i
o
n EPFAorlthm(
A
7
.
2
)
:Thec
o
r
r
e
s
p
o
n
d
i
n
gu
n
c
o
n
s
t
r
a
i
n
e
df
u
n
c
t
i
o
nf
o
ri
sp
r
o
b
l
e
mi
s

F(X I .X2 . 'g)=(x~-2x1X2+xf+X1X~-2xl +4)+

(x~ +xi-2
)
2+'
8
(
0
xf+O
1)2

(
7
.
2
2
)

i
Iet
h
a
tw
i
l
lc
a
l
c
u
l
a
t
eesoh
ona
n
dt
h
ev
a
l
u
e
so
fe
Sec7
_2_1_calc.mi
sa
nm-f
f
u
n
c
t
i
o
n
.f
o
rExample7
.
1
.f
o
rap
r
e
d
e
t
e
r
m
i
n
e
ds
e
tofv
a
l
u
e
so
ft
h
ep
e
n
a
l
t
ym
u
l
t
i
p
l
i
e
r
s
.
I
tr
e
q
u
i
r
e
stwoi
n
p
u
t
s
mt
h
eu
s
e
ra
td
i
f
f
e
r
e
n
ts
t
a
g
e
s
.Thef
i
r
s
ti
n
p
u
ti
st
h
ev
a
l
u
e
sf
o
r
em
u
l
t
i
p
l
i
e
r
sf
o
rwhiches
o
l
u
t
i
o
ns
h
o
u
l
db
eo
b
t
a
i
n
e
d
. l
ei
Is
to
fs
o
l
u
t
i
o
n(
t
h
e
r
e
a
r
en
i
n
ef
o
rt
h
i
sp
r
o
b
l
e
m
)i
sd
i
s
p
l
a
y
e
di
n Commandwindow.Theu
s
e
rf
i
n
d
se
s
o
l
u
t
i
o
nt
h
a
ts
a
t
i
s
f
i
e
st
h
es
i
d
ec
o
n
s
t
r
a
i
n
t
s(
u
s
u
a
l
l
yo
n
e
)whichmustb
ee
n
t
e
r
e
du
te
prompt(
n
o
t
e
:i
tm
u
s
tb
ee
n
t
e
r
e
da
sav
e
c
t
o
r
)
.Theprogramt
h
e
np
r
i
n
t
so
u
tev
a
l
u
e
s
a
r
i
o
u
sf
u
n
c
t
i
o
n
si
n
v
o
l
v
e
di
nee
x
a
m
p
l
e
.Thef
o
l
l
o
w
i
n
gi
sp
o
s
t
e
dfromt
h
e
o
fev
Commandwindowf
o
rb
o
t
ho
ft
h
ep
e
n
a
l
t
ym
u
l
t
i
p
l
i
e
r
s
e
tt
o2
5
.
Sec7 2 1 calc
enter value for rh [default

1
]

25

xt

enter value for rg [default


1
]
25
ans
-1.4125 -0.1015
-0.8692
1.0980
-0.6263 -1.2107
-0.0278 -1.3266
-0.0124
1.3267
-0.0090 -0.0000
0.7893 -1.1267
0.9775 1.0165
1.4184
-0.1318
Input the design vector chosen for evaluation
[0.9775 1.01651
The design vector [0.9775 1.0165 1
objective function: 2.9810
equality constraint: 0.0112
inequality constraint: 0.0138

276

7
.
2 INDIRECTMETHODSFORCONSTRAINEDOPTIMIZATION

NUMERICALTECHNIOUESFORCONSTRAINEDOPTIMIZATION

I
nt
h
ea
b
o
v
er
u
nt
h
ee
q
u
a
l
i
t
yandt
h
ei
n
e
q
u
a
l
i
t
yn
s
t
r
n
ta
r
n
o
ts
a
t
i
s
f
i
e
d(

e
x
p
e
c
t
e
d
)
.T
a
b
l
e7
.
1d
o
c
u
m
e
n
t
st
h
er
e
s
u
l
t
so
ft
h
ea
p
p
l
i
c
a
t
i
o
no
f eEPFmethodt
o
Example7
.
1t
h
r
o
u
g
hSec7
_2_1_calc.m.l
es
c
a
l
i
n
gf
a
c
t
o
ri
s5f
o
rb
o
t
hm
u
l
t
i
p
l
i
e
r
s
.
Ag
l
a
n
c
ea
lT
a
b
l
e7
.
1c
l
e
a
r
l
yi
1
Iu
s
t
r
a
t
e
st
h
ec
h
a
r
a
c
t
e
r
i
s
t
i
c
so
ft
h
eEPFmethod.
Ast
h
e
a
s
e
:
v
a
l
u
e
so
ft
h
em
u
l
t
i
p
l
i
e
r
si
n
c

277

Here)
.i
st
h
em
u
l
t
i
p
l
i
e
rv
e
c
t
o
rt
i
e
dt
ot
h
e u
a
l
i
t
yc
o
n
sa
i
n
t
s
pi
st
h
em
u
l
t
i
p
l
i
e
r
v
e
c
t
o
ra
s
s
o
c
i
a
t
e
dw
i
l
ht
h
ei
n
e
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
t
sandr
handr
sa
r
et
h
ep
e
n
a
l
t
y
m
u
l
t
i
p
l
i
e
r
su
s
e
ds
i
m
i
l
a
rt
ol
h
eEPFm
e
t
h
o
d
.F i
ss
o
l
v
e
da
sa
nu
n
c
o
n
s
t
r
a
i
n
e
d
f
u
n
c
t
i
o
nf
o
rp
r
e
d
e
t
e
r
m
i
n
e
dv
a
l
u
e
so
fp
r
"andr
g
.Thereforet
h
es
o
l
u
t
i
o
nf
o
r
e
a
c
hSUMTi
t
e
r
a
t
i
o
ni
s

X.=X

(p
r
r
g
)
h

. Thed
e
s
i
g
na
p
p
r
o
a
c
h
e
seo
p
t
i
m
a
lv
a
l
u
e
.

Thec
o
n
s
t
r
a
i
n
tv
i
o
l
a
t
i
o
n
sd
e
c
r
e
a
s
e
.
Thes
o
l
u
t
i
o
ni
sb
e
i
n
ga
p
p
r
o
a
c
h
e
df
r
o
mo
u
l
s
i
d
et
h
ef
e
a
s
i
b
l
er
e
g
i
o
n
.
Ani
m
p
o
r
t
a
n
tn
o
t
e d
st
ob
emad
.l
ea
n
a
l
y
t
i
c
a
lc
o
m
p
u
t
a
t
i
o
nu
s
i
n
gs
y
m
b
o
l
i
c
s
y
.Thesamec
a
n
n
o
tbes
a
i
df
o
rt
h
en
u
m
e
r
i
c
a
lc
o
m
p
u
t
a
t
i
o
n
.
c
o
m
p
u
t
a
t
i
o
na
p
p
e
a
r
1l
yl
o
o
k
i
n
ga
tF
i
g
u
r
e7.
3
.
e
s
p
e
c
ia

Atee
n
do
ft
h
eSUMTi
l
e
r
a
t
i
o
nl
h
ev
a
l
u
e
so
ft
h
em
u
l
t
i
p
l
i
e
r
sa
n
dpn
a
l
t
yc
o
n
s
t
a
n
t
s
a
r
eu
p
d
a
t
e
d
.Thel
a
t
t
e
ra
r
eu
s
u
a
l
l
yg
e
o
m
e
t
r
i
c
a
l
l
ys
c
a
l
e
db
u
tu
n
l
i
k
eEPFdon
o
th
a
v
et
o
v
e
nt
o f
o
rc
o
n
v
e
r
g
e
n
c
e
.
b
e i

Al
gorlthm:AugmentedLs
grangeMu
i
p
l
l
e
rLM)Method 7
.
3
)
S
t
e
p1
.ChooseX1
N
.(
n
o
.o
fSUMTi
t
e
r
a
t
i
o
n
s
)

Nu(
n
o
.o
fDFPi~rations)

7
.
2
.
2 AugmentedLagrangeM
u
l
t
i
p
l
i
e
r(ALM)Method

Ej
'

S(
f
i
o
r
c
o
n

v
r

g
e
n
c
ea
n
ds
t
o
p
p
i
n
g
)

T
h
i
si
st
h
em
o
s
tr
o
b
u
s
to
fep
e
na
1t
yf
u
n
c
t
i
o
nm
e
t
h
o
d
s
.Morei
m
p
o
r
t
a
n
t
l
yi
ta
1s
o
g
r
a
n
g
em
u
l
t
i
p
l
i
e
r
sa
tes
o
l
u
t
i
o
n
.T
h
i
si
sa
c
h
i
e
v
e
dby
p
r
o
v
i
d
e
si
n
f
o
r
m
a
t
i
o
non eLa
n
o
ts
o
l
v
i
n
gf
o
rt
h
em
u
l
t
i
p
l
i
e
r
sb
u
tm
e
r
e
l
yu
p
d
a
t
i
n
gthemd
u
r
i
n
gs
u
c
c
e
s
s
i
v
eSUMT
4
)
.I
tovercom
sm
anyo
ft
h
ed
i
f
f
i
c
u
l
t
i
e
sa
s
s
o
c
i
a
t
e
dw
i t
h
epena
1t
y
i
t
e
r
a
t
i
o
n
s[
2
f
u
n
c
t
i
o
nf
o
r
m
u
l
a
t
i
o
nw
i
t
h
o
u
ta
n
ys
i
g
n
i
f
i
c
a
n
to
v
e
r
h
e
n
d
.

Trans
r
matlont
otheUnconstralnedProblem:The g
e
n
e
r
a
1o
p
t
i
m
i
z
a
t
i
o
n
p
r
o
b
l
e
m(
7
.
1
)
(
7.
4
)i
s l
s
f
o
r
m
in o
do
fL
a
g
r
a
n
g
em
u
l
t
i
p
l
i
e
r
s
Minimize

)
+

)
dX)zd

r
8)
: f
(X)

(
7
.
2
3
)

(
X
)

hk

x
l
s Sx!

i=1
.2
.
.
.
.
.n

(
7
.
4
)

x
l
e
r
i
o
rP
e
n
a
l
l
yF
u
n
c
t
i
o
nM
e
l
h
o
d
:Exa
m
p
l
e7
.
1
T
a
b
l
e7
.
1E

I
t
e
r
a
t
i
o
n
'aq
JA--J

r
h

r
8

1
.0

1
.
0

X1

X2

0
.
9
2
2
8
0
.
9
4
6
4
0
.
9
7
7
5
0
.
9
9
4
2
0
.
9
9
8
8

1
.0
3
9
1

;
.
r
4
;(
O
i

n
i
t

i
a
1pe

n
a
I
t
y
mI

l
I

u
l
l
t

i
P

l
i

Ch
c
sc
a
1
i
ngv
a
l
u
ef
o
rm
u
l
t
i
p
l
i
e
r
s
)
g(

v
a
l
u
ef
o
rm
u
l
t
i
p
l
i
e
r

.
J
I(
i
n
i
t
i
a
lm
u
l
t
i
p
l
i
e
rv
e
c
t
o
r
s
)
q=1(SUMTi
t
e
r
a
t
i
o
nc
o
u
n
t
e
r
)
S
t
e
p2
. Ca
11DFPt
om
i
n
i
m
i
z
eF(X
q
)
.
.
q

p
"
Output
:xq.
S
t
e
p3
.C
o
n
v
e
r
g
e
n
c
ef
o
rALM
I
fh

f
o
rk=1

2
...
.

1
;
k=0
l
f
g

f
o
r
j
=
1
.
2

m;
:
j
r~O" r~ (maximum

r
Zr
:
)

( >Ofor=0)

(IfVf+k'
k
V
h
k+ ~V =0)
I
fa
l
ls
i
d
c
o
n
s
t
r
a
i
n
t
sa
r
es
a
t
i
s
f
i

ThenConverged
S
t
o
p
S
t
o
p
p
i
n
gC
r
i
t
e
r
i
a
:
M'=Fq-F
q-1
aJ(=x
q
*-X<q-I
)
l
f(M')2SE
J
:S
t
o
p(
f
u
n
c
t
i
o
nn
o
tc
h
a
n
g
i
n
g
)
s
eI
faJ(Td
.
XSEI:S
t
o
p(
d
e
s
i
g
nn
o
tc
h
a
n
g
i
n
g
)
E
l
s
eI
fq=N
.
:S
t
o
p(maximumi
t
e
r
a
t
i
o
n
sr
e
a
c
h
e
d
)
C
o
n
t
i
n
u
e

5
.
0
5
.
0
2
5
.
0 25.0
1
2
5
1
2
5
6
2
5
6
2
5

1
.
0
3
6
4
1
.0
1
6
5
1
.0044

1
.
0
0
0
9

0
.
0
6
8
7
2
.
9
5
7
8
0
.
0
2
2
7
0
.
0
3
0
2
2
.
9
6
5
1
0
.
0
2
9
5
0
.
0
1
1
2
2
.
9
8
1
0
0
.
0
1
3
8

2
.
9
9
0
.
0
0
3
7
0
. 2
7
5
.
9
7
7
5
e
0
0
4 7
2
.
9
9
8
8
0
0
4
.
5
0
9
7
e

qq+l
+2r
hh
(
X
"
O
)
p
qp
q+2rl(max(xq*).-pqI2r8])
r
1
.
rrCh;r:rfCg

278 N
U
M
E
R
I
C
A
LT
E
C
H
N
I
O
U
E
SF
O
RC
O
N
S
T
R
A
I
N
E
DO
P
T
I
M
I
Z
A
T
I
O
N

7.
2

xqxq*

o
f
v
a
l
u
e
sf
o
ra
1
1o
f
t
h
eSUMTi
t
e
r
a
t
i
o
n
sa
tt
h
et
e
r
m
i
n
a
t
i
o
no
f
t
h
ep
r
o
g
r
a
m
.Thei
n
i
t
i
a
l
n
d~ =1
.
d
e
s
i
g
nv
e
c
t
o
ri
s[32]
.I
n
i
t
i
a
l=1a

00t
oS
t
e
p2
TheALMa
l
g
o
r
im(
A
7
.
3
)i
sa
v
a
i
l
a
b
l
ea
sas
t
a
n
d
a
l
o
n
ep
r
o
g
r
a
mi
nALM.m.l
e
p
r
o
g
r
a
mi
sl
o
n
gw
h
i
c
hi
sn
o
tu
n
e
x
p
e
c
t
e
d
.Th
e
s
ea
1
g
o
r
i
t
h
r
n
sa
r
en
o
t v
i
a
l
.Tot
r
a
n
s
l
a
t
e
a
l
g
o
r
i
t
h
m(A7.3)i
n
t
ow
o
r
k
i
n
gc
o
d
ei
samendousa
c
c
o
m
p
l
i
s
h
m
e
n

tp
a
r
t
i
c
u
l
a
r
l
yi
f
y
o
u
r
e
x
p
o
t
oMA
T
I
.
A
Bo
c
c
u
r
r
e
dt
h
r
o
u
g
hi
sb
o
o
k
.Ther
e
a
d
e
rs
h
o
u
l
dw
a
l
kt
h
r
o
u
g
h
t
h
ec
o
d
el
i
n
eb
yl
i
n
e
. eo
r
g
a
n
i
z
a
t
i
o
na
n
ds
t
r
u
c
t
u
r
ea
r
ek
e
p
ts
i
m
p
l
ea
n
d
s
t
r
a
i
g
h
t
f
o
r
w
a
l
'
d
.C
o
n
u
n
e
n
l
sh
a
v
b
e
e
nl
i
b
e
r
a
l
l
yu
s
e
da
sa
nc
n
c
o
u
r
a
g
e
m
e
n
tt
o
u
n
d
e
r
st
.
1
n
dt
h
ec
o
d
e
.

ALM.m:Thec
o
d
eg
a
e
r
si
n
p
u
tf
r
o
mt
h
eu
s
e
rt
h
r
o
u
g
hs
e
v
e
r
a
lp
r
o
m
p
t
si
nt
h
e
Conunandw
i
n
d
o
w
.Th
ep
r
o
m
p
t
sa
n
d f
i
n
a
li
t
e
r
a
t
i
o
na
r
c
o
p
i
e
db
e
l
o
w
.The
s
i
g
n
i
f
i
c
a
n
tf
e
a
t
u
r
e
so
ft
h
ec
o
d
ea
r
e
:

A
p
p
l
i
c
a
t
i
o
nt
oExample7
.
1
:Thc f
o
l
l
o
w
i
n
gr
e
p
r
e
s
e
n
t
st
h
ei
n
i
t
i
a
lp
r
o
m
p
t
s
s
e
c
ni
nt
h
eCommandwindowa
n
dt
h
ev
a
l
u
e
sf
o
rt
h
cs
t
a
r
t
i
n
gi
t
e
r
a
t
i
o
n
s
.O
t
h
e
r
i
t
e
r
a
t
i
o
n
sa
r
es
i
m1
ia
r
l
yd
i
s
p
l
a
y
e
d
.A
l
s
oi
nc
1u
d
e
dbelowi
sac
o
n
s
o
l
i
d
a
t
e
dp
r
i
n
t
i
n
g

ALM

1nput the starting design vector


This is mandatory as there is no default vector setup
The length of your vector indicates the number of
unknowns (
n
)
Please enter it now and hit return for example [1 2
3 ...]
[3 2]

The initial design vector:


t

1nput the minimum values for the design vector. .


These are input as a vector. The default values are
3 *start value below the start value unless it is zero.
1n that case it is -5:
[
00
]

The minimumvalues for the design variables are

1
!
c
i
-

Thes
t
a
r
t
i
n
gd
e
s
i
g
nv
e
c
t
o
ri
st
h
ef
i
r
s
ti
n
p
u
tt
o p
r
o
g
r
n
.Thenumbero
fd
e
s
i
g
n
v
a
r
i
a
b
l
e
si
sa
u
t
o
m
a
t
i
c
a
l
l
ye
x
t
r
a
c
t
e
df
r
o
mt
h
i
si
n
f
o
r
m
a
t
i
o
n
.
l
el
o
w
e
ra
n
du
p
p
e
rb
o
u
n
d
son d
e
s
i
g
nv
a
b
l
e
s(
s
i
d
ec
o
n
s
n
t
s
)
o
b
t
a
i
n
e
dr
o
u
g
hu
s
e
rp
r
o
m
p
t
s(
t
h
o
u
g
he
ya
r
en
o
tu
s
e
di
nep
r
o
g
r
a
r
n
)
.
Numbero
fe
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
t
sa
n
dc
o
r
r
e
s
p
o
n
d
i
n
gi
n
i
t
i
a
lm
u
l
t
i
p
l
i
e
r
s(

)
p
r
o
m
p
t
c
d
.
Numbero
fi
n
e
q
ua
Ji
t
y n
s
t
r
a
i
n
t
sa
n
dc
o
r
r
e
s
p
o
n
d
i
n
gi
n
i
t
i
a
lm
u
l
t
i
p
l
i
e
r
s()
p
r
o
m
p
t
c
d
.
r
i
n
t
c
di
nt
h
eCommandw
i
n
d
o
w
.Wamings
V
a
l
u
e
sa
tc
a
c
hSUMTi
t
c
r
a
t
i
o
n p
g
e
n
e
r
a
t
e
db
yMATLAB s
w
i
t
c
h
e
do
f
f
.
The DFP.m i
su
s
e
df
o
ru
n
c
o
n
s i
n
e
do
p
t
i
m
i
z
a
t
i
o
n
.T
h
i
sp
r
o
g
r
a
mu
s
e
s
GoldSecti
o
n
nV
ar

oep
r
o
g
r
a
me
x

P
c
t

st
h
ef
o
l
l
o
w
i
n
ge
s
t

ob
ea
v
a
i
l
a
b
l
e
i
ne
p
a

:
O
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
n
Of
un.m
E
qu
a
l
i
t
yc
o
n
s
t
r
a
i
n
t
s
Hfun.m
I
n
e
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
t
s
Gfun.m
U
n
c
o
n
s
t
r
a
i
n
e
df
u
n
ct
Io
n FALM.m
Thep
r
o
g
r
a
mu
s
e
sg
l
o
b
a
ls
t
a
t
e
m
e
n
t
st
oc
o
m
m
u
n
i
c
a
t
em
u
l
t
i
p
l
i
e
rv
a
l
u
e
s
.
Thei
n
i
t
i
a
lp
e
n
a
l
t
ym
u
l
t
i
p
l
i
e
r
sa
r
ec
o
m
p
u
t
c
da
u
t
o
m
a
t
i
c
a
l
l
y
.
r
o
g
r
a
m
.e
s
p
e
c
i
a
I
lyt
h
o
s
en
e
e
d
e
df
o
rt
h
e
S
e
v
e
r
a
lp
a
r
a
m
e
t
e
r
sa
r
ec
o
d
e
di
n
t
oep
g
o
l
d
e
ns
e
c
t
i
o
na
n
dt
h
eu
p
p
e
rb
o
u
n
dc
a
l
c
u
l
a
t
i
o
n
.T
h
e
s
ec
a
nb
ec
h
a
n
g
e
db
y
u
s
r
.i
fn
e
c
e
s
s
a
r
y
.

I
N
D
I
R
E
C
TM
E
T
H
O
D
SF
O
RC
O
N
S
T
R
A
I
N
E
DO
P
T
I
M
I
Z
A
T
I
O
N 279

00

1nput the maximum values for the design vector.


These are input as a vector. The default values are
3 *start value above the start value.
1f start value is 0 then it is +5:
[
55
1
The maximum values for the design variables are
Number of equality constraints [
0
]
1nitial values for lambda 1
Number of inequality constraints [
0
]
1nitial values for beta 1

55

ALM iteration number: 0


Design Vector (
X
)
32
Objective function 64
Square Error in constraints(hg
)
1.2100e+002
1.8063e+001
Lagrange Multipliers (lambda beta): 1 1
Penalty Multipliers (rh rg): 5.2893e-001 3.5433e+000

280

N
U
M
E
A
I
C
A
LT
E
C
H
N
I
Q
U
E
SFOAC
O
N
S
T
R
A
I
N
E
DO
P
T
I
M
I
Z
A
T
I
O
N

(other iteration omitted)


X not changing

281

. Thes
t
a
r
t
i
n
gd
e
s
i
g
nc
a
nbef
e
a
s
i
b
l
eo
ri
n
f
e
a
s
i
b
l
e
.
m
u
l
t
i
p
l
i
e
r
s nber
e
l
a
x
e
d
.
Thei
n
i
t
i
a
lc
h
o
i
c
e
sf
o
rt
h

Thereisonlyamodestincreaseinthecalculations.

O.OOOE+OOO reached in 5 iterations

The values for x and f and 9 and h are


3.0000e+000 2.OOOOe+OOO 6
.4000e+001 1.8063e+001
1.2100e+002
7.9603e-001 7.1223e-001 2.9443e+000 0 7.379ge-001
9.9937e-001 1.0092e+000 2.9995e+000 1.8223e-004
2.9434e-004
9.9607e-001 1.0013e+000 2.9962e+000 0 2.731ge-005
9.9940e-001 1.0002e+000 2.9994e+000 2.5872e-011
6.2772e-007
9.9940e-001 1.0002e+000 2.9994e+000 2.5872e-011
6.2772e-007

Atc
o
n
v
e
r
g
e
n
c
et
h
eLagrangem
u
l
t
i
p
l
i
e
r
swi
1
lbed
r
i
v
e
n1
0t
l
l
e
i
roptimumv
a
l
u
e
s
.
J
lowsv
e
r
i
f
i
c
a
t
i
o
no
fKTc
o
n
d
i
t
i
o
n
st
h
r
o
u
g
ht
h
es
i
g
noft
h
em
u
l
t
i
p
l
i
e
r
s
T
h
i
sa
ea
c
t
i
v
ei
n
e
q
u
a
l
i
t
yc
o
n
sa
i
n
t
s
.
a
s
s
o
c
i
a
t
e
dwi

7
.
3 DIRECTMETHODSFORCONSTRAINEDOPT
lMIZAT
lON

:
1

U
n
e
s
r
l
z
a
t
l
o
n
:Th
eT
a
y
l
o
rs
e
r
i
e
sf
o
ralWo
.
v
a
r
i
a
b
l
eexpandedf
u
n
c
t
i
o
nj
{
x
y)
'
Y
p
)i
q
u
a
d
r
a
t
i
c
a
l
l
ya
b
o
u
ti
l
i
ec
u
r
r
e
n
tp
o
i
n
t(
x
se
x
p
r
e
s
s
e
da
s
:
p

[
J
1
3

{
J
+
I
! +)

p
rogramt
o
o
kf
i
v
ei
t
e
r
a
t
i
o
n
s(
l
i
f
t
hwasn
o
tn
e
c
e
s
s
a
r
y
)t
of
i
n
dt
h
es
o
l
u
t
i
o
na
s
:
Th

x
;
=
O
.
9
4
;;
x=1.0002; r
=2.9994; h.=0.0000; g.=0.0007 (7.24)
J

(
7
.
2
5
)

i
nS
e
c
t
i
o
n7
.1
.2 t
h
ev
a
l
u
e
so
ft
h
e
T
h
i
ss
a
t
i
s
f
i
e
st
h
e KTc
o
n
d
i
t
i
o
n
s
. However
o
b
t
a
i
n
e
du
s
i
n
gs
y
m
b
o
l
i
cc
o
m
p
u
t
a
t
i
o
n
a
r
eq
u
i
t
ed
i
f
f
e
r
e
nt
.I
ti
sd
i
f
f
i
c
u
l
tt
o
m
u
l
t
i
p
l
i
e
r
s
h
es
y
m
b
o
l
i
cc
a
l
c
u
l
a
t
i
o
nf
a
i
l
e
dt
od
i
s
c
o
v
e
r
e
x
p
l
a
i
nt
h
ed
i
s
c
r
e
p
a
n
c
ye
x
c
e
p
tt
os
a
y t
ev
a
l
u
e
si
n(
7
.
2
5
)a
r
mo

t
h
i
sp
a
r
t
i
c
u
l
a
rs
o
l
u
t
i
o
nwhichi
si
n
d
e
e
dp
o
s
s
i
b
l
e
. l
a
c
c
e
p
t
a
b
l
eant
h
o
s
ei
nS
e
c
t
i
o
n7
.
1
.
2c
o
n
s
i
d
e
r
i
n
geKTc
o
n
d
i
t
i
o
n
s
.
od

a
sd
e
m
o
n
s
t
r
a
l
e
dabove

i
s e
f
f
e
c
l
i
v
eandu
s
e
f
u
lmethodi
ns
D
i
t
e
TheALMme
o
fb
e
l
o
n
g
i
n
gt
ot
h
ec
l
a
s
so
fi
n
d
i
r
e
c
tme
o
d
s
.I
th
a
ss
e
v
e
r
a
ls
i
g
n
i
f
i
c
a
n
ta
d
v
a
n
t
a
g
e
so
v
e
r
t
h
eo
t
h
e
rp
e
n
a
l
l
yf
u
n
c
t
i
o
nf
o
r
m
u
l
a
t
i
o
n
s[
2
]
.Som
o
fthema
r
el
i
s
l
e
db
e
l
o
w
:
Themethodi
sn
o
ts
e
n
s
i
t
i
v
et
ot
h
ei
n
i
t
i
a
lp
e
n
a
l
t
ym
u
l
t
i
p
l
i
e
r
so
rt
h
e
i
rs
c
a
l
i
n
g
s
t
r
a
t
e
g
y
.

Themethoddoesn
o
t q
u
i
r
eep
e
n
a
l
t
ym
u
l
t
i
p
l
i
e
r
st
obed
r
i
v
e
nt
oe
x
t
r
e
m
e
v
a
l
u
e
s1
0
t
a
b
l
i
s
hc
o
n
v
e
r
g
e
n
c
e
.

Thee
q
u
a
l
i
t
yanda
c
t
i
v
ec
o
n
s
t
r
a
i
n
tc
a
nbes
a
t
i
s
f
i
e
dp
r
l
i
s
e
l
y
.

)
=f(x
/

)1
"y
f

+2

L_
.
.I
(
x

.
J
I

y
c
J
u
u
d

r i

whichi
sv
e
r
yc
10set
ot
h
ea
c
t
u
a
ls
o
l
u
t
i
o
n
.Th
ev
a
l
u
e
sf
o
rem
u
l
t
i
p
l
i
e
r
s

s
"=0.9922

Thed
i
r
e
c
tmethodsi
nc
1
udeb
o
t
ht
h
eo
b
j
e
c
t
i
v
e d.
ec
o
n
s
t
r
n
t
st
os
e
a
r
c
hf
o
rt
h
e
o
p
t
i
m
a
ls
o
l
u
t
i
o
n
.Whileemethodsdon
o
ti
n
v
o
l
v
ec
o
n
v
e
r
s
i
o
nt
oad
i
f
f
e
r
e
n
tc
1a
s
sof
p
r
o
b
l
e
m
s(
l
i
k
eS
e
c
t
i
o
n7
.
2
)
mosto
fthema
r
eb edonl
i
n
e
a
r
i
z
a
t
i
o
noft
h
ef
u
n
c
l
i
o
n
s
p
o
i
nl
.L
in
e
a
r
i
z
a
t
i
o
ni
sb
a
s
e
done
x
p
a
n
s
i
o
noft
h
ef
u
n
c
t
i
o
n
a
b
o
u
tt
h
ec
u
r
r
e
n
ld
e
s
i
a
b
o
u
tt
h
ec
u
r
r
e
n
tv
a
r
i
a
b
l
ev
a
l
u
e
su
s
i
n
gt
h
eT
a
y
l
o
rs
e
r
i

(
C
h
a
p
t
e
r4
)
.

The values for lamda and beta are


1.0000e+000 1.0000e+000
-9.0876e-001 0
-1.6493e001 9.5663e-001
9.1192e-002 9.5616e-001
7.6430e-002 9.9221e-001

A
.
'=
0
.
0
7
6
3
;

7
.
3D
I
A
E
C
TMETHODSFOAC
O
N
S
T
R
A
I
N
E
DO
P
T
I
M
I
Z
A
T
I
O
N

"I
(
x
p
J
'

"I
(

Ifed
i
s
p
l
a
c
e
m
e
n
t
sa
r
eo
r
g
a
n
i
z
e
d acolumnv
[
ax 8
y
]
T
. e
x
p
a
n
s
i
o
ni
n
Equ
a
t
i
o
n(
7
.
2
6
)canbee
x
p
r
e
s
s
e
di
nacondensedmanner

f
()=
f

L
[

(7
.
2
7
)

y]
T[

H(

;
X
n
.

I
n
J

]
]
I

.
.
I
sYI

ForI
tv
a
r
i
a
b
l
e
s
.w
i
t
hXpt
h
ec
u
r
r
e
n
tp
o
i
n
tand8xt
h
ed
i
s
p
l
a
c
e
m
e
n
tv
e
c
t
o
r

f
)=
f(X
)+V
f
l
p
)
T8X+tdXTH

Eq
u
a
t
i
o
n(
7
.
2
8
)c
b
ew
r
i
l
t
e
ni
nt
e
r
m
so
fed
i
f
f
e
r
e
n
c
ei
nf
u
n
c
t
i
o
nv
a
l
u
e
s

T
=f(X
)=V
f
(
X
p
)
TdX+ dX
Xp) )
p
p+ f(X

df=Of+of

282

7
.
3 DIRECTMETHODSFORCONSTRAINEDOPTIMIZATION

283

NUMERICALTECHNIQUESFORCONSTRAINEDOPTIMIZATION

whereof
=Vj
T
aXi
st
e
n
n
e
dJ
i
r
s
tv
a
r
i
a
t
i
o
n
.o2
j
i
st
h
es
e
c
o
n
dv
a
r
i
a
l
i
o
nandi
sg
i
v
e
n
o
ndt
e
n
ni
nt
h
eabove u
a
t
i
o
n
.I
nl
i
n
e
a
r
i
z
a
t
i
o
noft
h
ef
u
n
c
t
i
o
n
f
a
b
o
u
tt
h
e
by es
n
l
yt
h
ef
i v
a
r
i
a
t
i
o
ni
su
s
e
d
.Th
n
e
i
g
h
b
o
r
i
n
g u
eo
f
c
u
r
r
e
n
tv
a
l
u
eofd
e
s
i
g
nXp'o
t
h
ef
u
n
c
t
i
o
nc
a
nb
ee
x
p
r
e
s
s
e
da
s

l(X
f(X
TaX
p) f(X
p)+V
p)

(
7.
30
)
3

A
l
loft
h
ef
u
n
c
t
i
o
n
s eproblemc
b
el
i
n
e
a
r
i
z
e
ds
i
m
i
l
l
y
.I
ti
se
s
s
e
n
t
ia
It
o
i
f
f
e
r
e
n
c
ebetweenf(
X
)andl(X
p
)
'T
h
i
si
si
l
Iu
s
t
r
a
t
e
di
nF
i
g
u
r
e7.
4
u
n
d
e
r
s
t
a
n
d d
u
s
i
n
gt
h
eo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
nofExnmple7
.
1expandeda
b
o
u
tt
h
ec
u
r
r
e
n
td
e
s
i
g
n
x=3
x2=2.Th curvedi
I
nesf(X)a
r
et
h
ec
o
n
t
o
u
r
so
feo
r
i
g
i
n
a
If
u
n
c
t
i
o
n
.Thes
B
i
g
h
t
l
i
n
e
sJ
(
Xp)a
r
et
h
ec
o
n
t
o
u
r
so
ft
h
el
i
n
e
a
r
i
z
e
df
u
n
c
t
i
o
n(
l
i
n
e
sofc
o
n
s
t
a
n
tv
a
Iueoft
h
e

f
u
n
c
t
i
o
n
)o
b
t
a
i
n
e
dt
h
r
o
u
g
ht
h
ef
o
l
l
o
w
i
n
g
:

x
i

f
(
x
)
.X2): x1-2xrX2+ +X~

2x1+4

L
in
e
a
r
i
z
e
df
u
n
c
t
i
o
n
:]
(
!
J
.X
I
'!
J
.X2)=+92!
J
.XI-6!
J
.X2
wheret
h
ec
o
e
f
f
i
c
i
e
n
t
si
n seconde
x
p
r
e
s
s
i
o
ni
n
c
l
u
d
e
st
h
ee
v
a
l
u
a
t
i
o
noft
h
e
X2=2a
nds
u
b
s
t
i
t
u
t
e
di
nE a
t
i
o
n(
7.
30
)
.I
nF
i
g
u
r
e
f
u
n
c
t
i
o
nandg
r
a
d
i
e
n
tatx =3
r
e
s
ta
r
ei
nt
h
eneighborhood
7
.
4s
e
v
e
r
a
lc
o
n
t
o
u
r
sa
r
eshowna
l
t
h
o
u
g
ht
h
o
s
eofi
n
t
u
r
r
e
n
td
e
s
i
g
n
.Anyn
o
n
l
i
n
e
a
rf
u
n
c
t
i
o
nc
a
nbeexpandedi
nt
h
esamem
a
n
n
e
r
.
ofec
I
fa
n
o
t
h
e
rp
o
i
n
twasc
h
o
s
e
n
.t
h
e
nes
l
o
p
e
s0"
1L
hel
i
n
e
swouldbed
i
f
f
e
r
e
n
.
t
Theq
u
a
d
r
a
t
i
ce
x
p
a
n
s
i
o
noft
h
ef
u
n
c
t
i
o
nc
a
nb
eo
b
t
a
i
n
e
dbyu
s
i
n
gE
q
u
o
n(
7
.
2
8
)
.
l
ee
x
p
a
n
d
e
dc
u
r
v
e
sw
i
l
lben
o
n
l
i
n
e
a
r(
q
u
a
d
r
c
)
. l
e
y
a
p
p
e
a
r e
l
l
i
p
s

i
nF
igw
7
.
5
r
a
Ic
o
n
t
o
u
r
so
fe 0e
c
t
i
v
ef
u
n
c
t
i
o
no
fExam
p
l
e 7.1 e
x
p
a
n
d
e
d
where s
e
v
e

l
I
ya
u
tXI=
3
X2=
2
.I
ti
si
m
p
o
r
t
a
n
tt
o o
g
n
i
z
et
h
a
te
c
o
n
t
o
u
r
w
ouldbe
q
u
a
d
r
a
t
i
c
h
o
s
e
nf
o
re
x
p
a
n
s
i
o
n
.
d
i
f
f
e
r
e
n
ti
fa
n
o
t
h
e
rp
o
i
n
twe
d
i
s
c
u
s
s
e
dThef
i
r
s
ti
sS
e
q
u
e
n
t
ia
IL
in
e
a
rP
r
o
g
r
a
r
n
m
i
n
g
Fourd
i
r
e
c
tmethodsa
r
h
es
o
l
u
t
i
o
ni
so
b
t
a
i
n
e
dbys
u
c
c
e
s
s
i
v
e
l
ys
o
l
v
i
n
gt
h
ec
o
r
r
e
s
p
o
n
d
i
n
g
(
S
L
P
)wheret
o
ndS
e
q
u
e
n
t
ia
IQ
u
a
d
r
a
t
i
cP
r
ogramming
l
i
n
e
a
r
i
z
e
do
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
e
m
.Thes
s
e
seq
u
a
d
r
a
t
i
ce
x
p
a
n
s
i
o
nf
o
reo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
n
.L
ikeeSUMTe
(SQP)u
t
a
r
t
i
n
gv
a
Iu
e
sf
o
rt
h
en
e
x
ti
t
e
r
a
t
i
o
n
.Thet
h
i
r
d
.e
c
u
r
r
e
n
ts
o
l
u
t
i
o
np
r
o
v
i
d
e
ses
a
Ii
z
e
d ReducedG
r
a
d
i
e
n
tMethod (GRG).d
e
v
e
l
o
p
sas
o
p
h
i
s
t
i
c
a
t
e
ds
h
Gener
d
i
r
e
c
t
i
o
na
n
df
o
l
l
o
w
si
tw
i
t
hane
l
a
b
o
r
a
t
eo
n
e
d
i
m
e
n
s
i
o
n
a
lp
r
o
c
e
s
s
.Thef
o
u
r
t
h
.L
he
1GradientR
e
s
t
o
r
a
t
i
o
nA
l
g
o
r
i
t
h
m(SGRA)u
s
e
sat
w
o
c
y
c
l
ea
p
p
r
o
a
c
h
S
e
q
u
e
n
t
ia
1tematelyt
of
i
n
dt
h
eoptimum.l
e
r
ea
r
e
workingonf
e
a
s
i
b
i
l
i
t
yando
p
t
i
m
a
l
i
t
ya
a
1o
t
h
e
rmethodsduet
omanyr
e
s
e
a
r
c
h
e
r
sb
u
te
yd
i
f
f
e
rfromo
s
el
i
s
t
e
di
n
s
e
v
er
pt
h
el
e
n
g
t
hoft
h
i
sc
h
a
p
t
e
rr
e
a
s
o
n
a
b
l
e
.eyhaven
o
tb
e
e
n
s
m
a
l
ld
e
t
a
i
l
s
.Tok
Ju
d
e
d
.Withecode do
t
h
e
rp
r
o
c
e
d
u
r
e
sa
v
a
i
l
a
b
l
efromt
h
i
ss
e
c
t
i
o
n
.i
ts
h
o
u
l
dn
o
t
i
nc
c
u
l
tt
oprogramt
h
a
d
d
i
t
i
o
n
a
lt
e
c
h
n
i
q
u
e
sa
ta
r
ew
o
r
t
he
x
p
l
o
r
i
n
gf
o
rp
a
r
t
i
c
u
l
a
r
b
e
d
i
f
c
Ja
s
s
e
sofp
r
o
b
l
e
m
s
.
Ex
c
e
p
tf
o
rs
p
e
c
i
a
lc
l
a
s
s
e
sofp
r
o
b
l
e
m
si
nt
h
ef
o
l
l
o
w
i
n
gtwoc
h
a
p
l
e
r
s
. i
sc
h
a
p
t
e
r
i
st
h
ep
r
i
n
c
i
p
a
lr
e
a
s
o
nf
o
rt
h
eb
o
o
k
.I
tp
r
o
v
i
d
e
s ar
e
s
o
l
u
t
i
o
no
ft
h
ec
o
m
p
l
e
t
e
o
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
e
ma
sp
o
s

l
a
t
e
d
.F
o
ral
o
n
gt
i
m
enowt
h
e
r
ehaven
o
tb
e
e
nanynew
o
rg
r
e
a
t
l
yi
m
p
r
o
v
e
dn
u
m
e
r
i
c
a
lt
e
c
h
n
i
q
u

.l
e
r
ei
sal
o
to
fworkong
l
o
b
a
lo
p
t
i
m
i
z
a
t
i
o
n
.

4Af
t222dt

o
n
:
O
r
i
g
i
naIi

F
d

-1

xt

F
1g
u
r
e7
.
4l
Iu
s
t
r
a
t
l
o
no
l
l
l
n
e
a
r
i
z
a
t
l
o
n
.

C
u
r
r
e
n
ti
n
t
e
r
e
s
ti
nt
r
a
d
i
t
i
o
n
a
ld
e
s
i
o
p
t
i t
i
o
n(
i
schap

)
i
sd
i
r
e
c
t
e
dt
o
w
a
r
d
a
c
t
u
a
ld
e
s
i
g
nappucadonsandm
i
g
r
a
t
i
o
nfmmthemainfmmeandUNIXE
n
v
i
r
o
n
m
e
n
t
l
i
k
ei
sbook)Forc
o
m
p
l
e
t
e
n
e
s
s
.t
h
es
t
a
n
d
a
r
df
o
n
n
a
to
ft
h
eNLPi
s
t
oePC(
r
e
p
r
o
d
u
c
e
dh
e
r
ef
o
rc
o
n
v
e
n
i
e
n
c
e
:

(
x
1
oX2. x

)
Mi
n
i
m
i
z
e !

(
7
.1
)

t
<
X"X2.


XI
. k=12.... 1
I) =o
S
u
b
j
e
c
t

: h

(
7
.
2
)

~XI X2....
) O
. j=12
.
.
.
.
m
XlS
:XiS
:Xr.

i=1.2....
n

3
)
(
7.

(
7.
4
)

I
nv
e
c
t
o
rn
o
t
a
t
i
o
n

(X
)
. ]"
Minimize f

h
(
X
)
]=O
t
o
: [
S
u
b
j t

(
7
.
5
)
(
7
.
6
)

2
8
4 N
U
M
E
R
I
C
A
LT
E
C
H
N
I
O
U
E
SF
O
RC
O
N
S
T
R
A
I
N
E
DO
P
T
I
M
I
Z
A
T
I
O
N

7
.
3

M
i
n
i
m
i
z
e
:

l
(
aX)=f
(
X
;
)+V
f
(
X
;
)
TaX

(
7
.
3
1
)

Su
t
l
o
:

aX): "
k
(
X
;
)+v
h
I(X;)aX=
0
; k=
1

2...
1

(
7
.
3
2
)

(
A
X
)
: (
X
;
)+V8
J
{
X
;
)
TAXS0
; j=1

2
.
..
.
m

(
7
.
3
3
)

A
x
!
o
wa
x
;Sax)
I
P
; i=1

2.
..

(
7.
34)

D
I
R
E
C
TM
E
T
H
O
D
SF
O
RC
O
N
S
T
R
A
I
N
E
DO
P
T
I
M
I
Z
A
T
lO
N 2
8
5

Equ
a
t
i
o
n
s(
7.
3
1
)
(
7
.
3
4
)r
e
p
r
e
s
e
n
laL
in
e
a
rP
r
o
g
r
a
m
m
i
n
g(
L
P
)p
r
o
b
l
e
m(
C
h
a
p
t
e
r3
)
.
u
a
t
i
o
n
s(
7
.
3
1
)
(
7.
3
4
)
e
x
c
e
p
tf
o
raX.h
a
v
en
u
m
e
r
i
c
a
lv
a
l
u
e
s
A
l
lo
ft
h
ef
u
n
c
t
i
o
n
si
nEq
a
f
t
e
rs
u
b
s
t
i
t
u
t
i
n
gan
u
m
e
r
i
c
a
lv
e
c
t
o
rf
o
rX
;
.A
s
s
u
m
i
n
ga
nLPp
r
o
g
r
a
mc
o
d
ei
s
a
v
a
i
l
a
b
l
e
i
t nb
ec
a
l
l
e
dr
e
p
e
a
t
e
d
l
ya
f
t
e
rt
h
ed
e
s
i
g
ni
su
p
d
a
t
e
da
s

2
"
'
"

X"=X+AX
l-~.i

I
no
r
d
e
rt
oi
n
c
l
u
d
eas
e
a
r
c
hd
i
r
e
c
t
i
o
na
n
ds
t
c
p
s
i
z
ec
a
l
c
u
l
a
t
i
o
nXi
nEq
u
a
t
i
o
n
s
(
7
.
3
1
)(
7
.
3
3
)i
sc
o
n
s
i
d
e
r
e
da
sas
o
l
u
t
i
o
nf
o
rS(
as
e
a
r
c
hd
i
r
e
c
t
i
o
na
tt
h
ec
u
r
r
e
n
l
g
yi
sn
ol
o
n
g
e
rs
i
m
p
l
e
.I
f

r
ea
r
ev
i
o
l
a
t
e
d
d
e
s
i
g
n
)
.S
t
e
p
s
i
z
ec
o
m
p
u
t
a
t
i
o
ns
l
r
a
c
o
n
s
t
r
a
i
n
t
s
a
1
p
h
aa
t
t
e
m
p
t
st
or
e
d
u
c
ei
sv
i
o
l
a
t
i
o
n
.I
fec
u
r
r
e
n
ts
o
l
u
t
i
o
ni
sf
e
a
s
i
b
l
e
es
t
e
p
s
i
z
ew
i
l
la
t
t
e
m
p
tt
or
e
d
u
c
et
h
e n
c
t
i
o
nw
i
t
h
o
u
tc
a
u
s
i
n
gt
h
ec
o
n
st
r
a
I
n
t
st
ob
e
e
x
c
e
s
s
i
v
e
l
yv
i
o
l
a
t
e
d
.S
u
c
hs
t
r
a
t
e
g
i
e
sa
r
el
a
r
g
e
l
yi
m
p
l
e
m
e
n
t
e
da
tt
h
ed
i
s
c
r
e
t
i
o
no
f
a
l
g
o
r
i
t
h
mo
rc
o
d
ed
e
v
e
l
o
p
e
r
.

F
J
g
u
r
e7
.
5l
Iu
s
t
r
a
t
i
o
n0
1q
u
a
d
r
a
t
l
c9
x
p
a
n
s
l
o
n
.
[
g
(
X
)
]
mS0

(
7
.
7
)

X10wSXSXup

(
7
.
8
)

E
x
a
m
p
l
e7.
1w
i
l
lb
eu
s
e
dt
oi
l
Iu
s
t
r
a
t
et
h
ev
a
r
i
o
u
sa
l
g
o
r
i
t
h
m
s
.

M
i
n
i
m
i
z
e f
(
x xi
)
:

x
1-2xix2+xi+xlrz-2xl+4

S
u
b
j
e
c
t
t
o
: h
(
X
I
x
:
z
)
: x
i+r
z-2=0

(
7
.
9
)

(
7
.
1
0

g
(
X
.
.
X
2
)
: 0.2sxi+0.75~-1 SO

(
7.
10
b
)

OSXI:
:
;
5
; OSX2SS

(
7
.l
O
c
)

7
.
3
.
1 S
e
q
u
e
n
t
l
a
lL
i
nearProgrammlng(SLP)
I
nt
h
eSLP[
5
]a
l
lo
ft
h
ef
u
n
c
t
i
o
n
sa
r
ee
x
p
a
n
d
e
dl
i
n
e
a
r
l
y
.I
fX
;i
sc
o
n
s
i
d
e
r
e
de
c
u
r
r
e
n
t
d
e
s
i
g
nv
t
o
r

e
nel
i
n
e
a
r
i
z
e
do
p
t
i
m
a
lp
r
o
b
l
e
mc
a
nb
es
e
tu
pa
s

Ao
r
l
t
h
m
:S
e
q
u
e
n
t
l
a
lL
l
nearProgrammlng(SLP)(
A4
)
S
t
e
p1
.C
h
o
o
s
eX
'
Ns(
n
o
.o
fi
t
e
r
a
t
i
o
n
s
)
E
;
'
S(
f
o
rc
o
n
v
e
r
g
e
n
c
ea
n
ds
t
o
p
p
i
n
g
)
q=1(
i
t
e
r
a
t
i
o
nc
o
u
n
l
e
r
)
31
)
(
7.
3
4
)
S
t
e
p2
.C
a
l
lLPt
oo
p
t
i
m
i
z
e(
7.
O
u
t
p
u
t
:S
U
s
eac
o
n
s
t
r
n
e
d

c
a
l
c
u
l
a
t
i
o
n(
X
;=x
Q
)
AX=..S
Xq+I=Xq+AX
S
t
e
p3
.C
o
n
v
e
r
g
e
n
c
ef
o
rSLP
I
fhk=0
f
o
rk=1

2...

1
;
I
fg
JS0

f
o
rj=1
.2.
.
.

m;
[
fa
l
ls
i
d
ec
o
n
s
t
r
a
i
n
t
sa
r
es
a
t
i
s
f
i
e
d
I
fKTc
o
n
d
i
t
i
o
n
ss
a
t
i
s
f
i
e
d
ThenC
o
n
v
e
r
g
e
d
S
t
o
p
S
t
o
p
p
i
n
gC
r
i
t
e
r
i
a
:
q
AX=xq+l-x
T
I
fAX
AXSE
I
:S
t
o
p(
d
e
s
i
g
nn
o
tc
h
a
n
g
i
n
g
)
I
fq=N
.
:S
I
O
p(maximumi
t
e
r
a
t
i
o
n
s a
c
h
e
d
)

7
.
3O
I
R
E
C
TM
E
T
H
O
D
SF
O
RC
O
N
S
T
R
A
I
N
E
DO
P
T
I
M
I
Z
A
T
I
O
N 287

2
8
6 N
U
M
E
R
I
C
A
LT
E
C
H
N
I
Q
U
E
SF
O
RC
O
N
S
T
R
A
I
N
E
OO
P
T
I
M
I
Z
A
T
I
O
N
5

C
o
n
t
i
n
u
e
qq+l

00t
oS
t
e
p2
TheSLPi
sa
p
p
l
i
e
d1
0E
x
a
m
p
l
e7
.
1b
e
l
o
w
.I
nC
h
a
p
t
e
r3n
oe
x
p
l
i
c
i
tp
r
o
g
r
a
mf
o
rs
o
l
v
i
n
g
eLPp
r
o
g
r
a
mw
a
sd
e
v
e
l
o
p
e
da
l
t
h
o
u
g
hacommand-by-commandw
a
l
kr
o
u
g
hw
a
s
i
n
t
r
o
d
u
c
e
d
.Ap
r
o
g
r
a
mf
o
ra
p
p
l
y
i
n
geSimplexme
o
d

pslmplex.mi
si
n
c
l
u
d
e
do
n
ew
e
b
s
i
t
e(
w
i
t
h
o
u
ta
n
ycommento
rd
i
s
c
u
s
s
i
o
n
)
.U
s
i
n
gi
n
s
t
r
u
c
t
i
o
n
sf
r
o
mC
h
a
p
t
e
r3

a
d
d
i
n
gsomcp
r
o
g
r
a
mc
o
n
t
r
o
la
n
do
r
g
a
n
i
z
i
n
gsomeb
o
o
k
k
e
e
p
i
n
gc
a
np
r
o
v
i
d
ea
MATLABp
r
o
g
r
a
mf
o
rLPp
r
o
b
l
e
m
s
.A
l
t
e
m
a
t
e
l
y
M.
.AB'
Sl
i
n
p
r
o
gc
a
nb
eu
s
e
di
na
l
o
o
pt
oh
e
l
pl
r
a
n
s
l
a
t
eeSLPa
l
g
o
r
i
t
h
mt
oc
o
d
e
.I
ni
ss
e
conel
i
n
e
a
r
i
z
e
d
l
Iyp
r
e
s
e
n
t
e
d
.Thea
r
e
ao
fi
n
t
e
r
e
s
tc
a
nb
c
s
u
b
p
r
o
b
l
c
mi
ne
a
c
hi
t
e
r
a
t
i
o
ni
sg
r
a
p
h
i
c
a
zoomeda
n
dt
h
es
o
l
u
t
i
o
nr
e
a
df
r
o
mt
h
ep
l
o.
tT
h
i
si
st
r
e
a
t
e
da
sdX(
i
n
s
t
e
a
do
fu
s
i
n
gi
t
a
sS
)
.Th
ed
e
s
i
g
ni
su
p
d
a
t
e
da
n
da
n
o
t
h
e
rp
l
o
tc
a
nb
eo
b
t
a
i
n
e
db
yr
u
n
n
i
n
gt
h
emf
i
l
e
f
o
ranewv
a
l
u
eo
ft
h
ed
e
s
i
g
nv
e
c
t
o
r
.Sec7
_3_1p
1
o
t
.m a
s
s
i
s
t
si
nd
i
s
p
l
a
y
i
n
gt
h
e
i
I
n
e
a
r
i
z
e
df
u
n
c
t
i
o
n
sf
r
o
mw
h
i
c
hs
o
l
u
t
i
o
nc
a
nb
eo
b
t
a
i
n
e
d
.Th
ee
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
ti
s
eg
r
nl
i
n
e
.l
er
e
di
si
n
e
q
u
a
l
i
l
yc
o
n
s
t
r
i
n
tl
ed
e
s
i
g
nv
e
c
t
o
ri
s p
u
t
t
oe
p
r
o
g
r
a
m
.Thd
eu
s
e
ss
y
m
b
o
l
i
cc
a
l
c
u
l
a
t
i
o
n1
0s
o
l
v
ea
n
dd
i
s
p
l
a
yt
h
es
u
b
p
r
o
b
l
e
m
.

A
p
p
l
l
c
a
t
l
o
nofSLP
sP1
. Xl=
(32)
S
t
e
p2
.L
in
e
a
r
i
z
e
ds
u
b
p
r
o
b
l
e
m
r.
M
i
n
i
m
i
z
e1
(
dX)=64+[
9
26
] ~XI
sX
I
2
1

Iso
I
!
lX
2
1

FromFigure7.6(zoomed)thesolutionis~ =ax
=1.3
3
5
&2=ax
7
5
.
F
o
r
l
2=-O.

.6
6
51
.2
5
]
.R
e
p
e
a
tS
t
e
p2
.
t
h
e
n
e
x
ti
t
e
r
a
t
i
o
nX2=[1
in
a
r
i
z
e
ds
u
b
p
r
o
b
l
e
m
S
t
e
p2
.L

S :():23347+[333251lh=0
sX
I
2
1
l
a
x
.
l
g
(
dX9+ 2
51
.8
7
5
ISO

1
/

in
e
a
r
i
z
e
ds
u
b
p
r
o
b
l
e
m
S
t
e
p2
.L

Minim

M
i
n
i
m
i
)=6
.
7
9
8
5+[
1
3 0
51
.
3
8
2
]1~XII
I
LU21

FromF
i
g
u
r
e7
.
7(
z
o
o
m
e
d
)t
h
es
o
l
u
t
i
o
ni
saxl=0
.
5
4

ax
0
.
2
3
.F
o
rt
h
en
e
x
t
2=
1
.
12
51
.0
2
]
.R
e
p
e
a
tS
t
e
p2
.
i
t
e
r
a
t
i
o
nX3=[
-i
i
3
i
f
z
j
-

g
(
d
X
)
:4
.
2
5+[
1
.53
1
1

F
l
g
u
r
e7
.
6E
x
a
m
p
l
e7
.
1S
L
Pi
l
e
r
a
t
i
o
n1
.

c
t I
I
(
d
X
)
:1
1+[
64
]I 1
=
0
!
l

OX
1

II

21
IX

i
I
l

1
(
.
)=3
.
2
0
6+[
2
.
3
9
5
70
.
2
3
6
3
]I

l "
'
2
1

1=0

b
j
e
c
t1
0
:():0.
3060+[
2
.
2
52
.
0
4
]1

l"'

2
l

)
:

0
6

FromF
i
g
u
r
e7
.
8(
z
o
ome
d
)t
h
es
o
l
u
t
i
o
n
i
sd
x
(=0
.
1

2
5

d
x
2=
0
.
0
2
.F
o
rt
h
en
e
x
t
.
01
.0]
.S
i
n
c
et
h
i
si
st
h
eknowns
o
l
u
t
i
o
nf
u
r
t
h
e
ri
t
e
r
a
t
i
o
n
s
i
t
e
r
a
t
i
o
nX4=[1
s
t
o
p
p
e
d
.
l
eSLPt
o
o
kt
h
r
e
ei
t
e
r
a
t
i
o
n
st
o
.f
i
n
dt
h
es
o
l
u
t
i
o
n
.T
h
i
si
sv
e
r
yi
m
p
r
e
s
s
i
v
e
.I
ti
st
r
u
e
o
.
s
ec
a
s
e
sw
h
e
r
et
h
enumbero
.
fe
q
u
a
l
i
t
yp
l
u
sa
c
t
i
v
ei
n
c
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
t
si
s
o
n
l
yf
o
rt
h

288

7
.
3 DIAECTMETHODSFOACONSTRAINEDOPTIMlZATION

NUMEAICALTECHNIQUESFORCONSTRAINEDOPTIMIZATION

>
<0

289

<0

-3

-3

-4

L
5
5

-2~ Lttl 2;

-2~A1 S2

l
ti
sc
l
e
a
rf
r
o
mF
i
g
u
r
e7
.
9I
h
a
tw
i
l
h
o
u
lI
h
emovel
i
m
i
t
st
h
cs
o
l
u
t
i
o
ni
sunbounded
whichi
sn
o
tv
e
r
yh
e
l
p
f
u
.
lUsingl
h
es
o
l
u
t
i
o
no
b
t
a
i
n
e
di
nl
h
ef
i
g
u
r
e
l
h
en
e
x
ti
t
e
r
a
t
i
o
n

-3

SX1

F
l
g
u
r
e7
.
8E
x
a
m
p
l
e7
.
1SLPI
l
e
r
a
t
i
o
n3
.

F
l
g
u
r
e7
.
7E
x
a
m
p
l
e7
.
1SLPI
l
e
r
a
l
l
o
n2
.
l
h
e same a
s e number o
fd
e
s
i
g
nv
a
r
i
a
b
l
e
s
af
u
l
lc
o
m
p
[
e
m
e
n
t
. The main
d
i
s
a
d
v
a
n
t
a
g
eo
fl
h
emelhoda
p
p
e
s
i
fe
r
ei
snof
u
l
lcomplement
.I
nl
h
a
tc
a
s
e s
i
d
e
c
o
n
s
i
n
t
s(
7.
3
4
)a
r
ec
r
i
t
i
c
a
lt
oed
e
t
e
n
n
i
n
a
t
i
o
no
fl
h
es
o
l
u
t
i
o
na
se
yd
e
v
e
l
o
pi
n
t
o
a
c
t
i
v
ec
o
n
s
t
r
a
i
n
t
s
.I
nl
h
e
s
es
i
t
u
a
t
i
o
n
sl
h
el
i
m
i
t
sonl
h
es
i
d
ec
o
n
s
t
r
a
i
n
t
s ec
a
l
l
e
dl
h
e
movel
i
m
i
t
s
.l
n s
eo
fa
c
t
i
v
es
i
d
ec
o
n
s
t
r
n
i
n
l
sI
h
emovel
i
m
i
l
se
s
t
a
b
l
i
s
ht
h
ev
a
l
u
e
so
f
l
h
ed
e
s
i
g
nc
h
a
n
g
e
s
.I
fl
h
ed
e
s
i
g
nc
h
a
n
g
e
sa
r
ec
o
n
s
i
d
e
r
a
sl
h
es
e
a
r
c
hd
i c
t
i
o
n(
s

a
l
g
o
r
i
l
h
m
)e
nbo
v
a
J
ueandd
i
r
e
c
t
i
o
na
r
ea

c
t
e
d
.Forl
h
e
s
ec
h
a
n
g
e
st
ob
es
m
a
l
l
a
sl
h
es
o
l
u
t
i
o
ni
sa
p
p
r
o
a
c
h
e
d

emovel
i
m
i
t
shavet
obea
d
j
u
s
t
e
d(
l
o
w
e
r
e
d
)w
i
t
he
v
e
r
y
i
t
e
r
a
t
i
o
n
. l
es
t
r
n
t
e
g
yf
o
ri
sa
d
j
u
s
t
r
n
e
n
tmayi
n
f
l
u
e
n
c
e es
o
l
u
t
i
o
n
.
Example7
.
1w
i
l
h
o
u
tt
h
ei
n
e
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
t
b
u
tw
i
l
hee
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
t
p
r
o
v
i
d
e
sa
ni
l
l
u
sl
i
o
no
ft
h
emovel
i
m
i
t
s
.F
i
g
u
7
.
9r
e
p
r
e
s
e
n
t
sl
h
el
i
n
r
i
z
a
t
i
o
n
a
b
o
u
tl
h
ep
o
i
n
t[3

2.
JThemovelimitsi
nl
h
ef
i
g
u (
b
o
x
)r
e
p
r
e
s
e
n
tl
h
es
i
d
ec
o
n
s
t
r
a
i
n
t
l
i
m
i
t
s

-4

bee
x
e
c
u
t
e
d
.Oncea
g
a
i
nl
h
emovel
i
r
n
i
t
sw
i
l
li
n
f
l
u
e
n
c
ees
o
l
u
t
i
o
n
.I
ft
h
emove

t
h
es
o
l
u
t
i
o
nw
i
l
la
l
w
a
y
sbeonl
h
es
q
u
ew
i
t
hno
l
i
m
i
t
sa
r
el
e
f
ta
tl
h
esamev
a
l
u
e
p
o
s
s
i
b
i
l
i
t
yo
fc
o
n
v
e
r
g
e
n
c
e
.Thesel
i
r
n
i
t
smuslbel
o
w
e
r
e
dw
i
t
h c
hi
t
e
r
a
t
i
o
n
.U
s
u
a
l
l
y
t
h
e
ya
r
edoneg
e
o
m
e
t
r
i
c
a
l
l
yl
h
r
o
u
g
has
c
a
l
i
n
gf
a
c
l
o
r
.l
i
si
m
p
l
i
e
sl
h
ec
o
n
v
e
r
g
e
n
c
e
w
i
l
ldependonl
h
es
t
r
a
l
e
g
yf
o
rc
h
a
n
g
i
n
gl
h
emovel
i
m
i
t
s
n
o
ta
na
p
p
e
a
l
i
n
gs
i
t
u
a
t
i
o
n
.
Th
eSLPi
sn
o
toneofl
h
ep
o
p
u
l
a
rm
e
l
h
o
d
s
.1
1
1
emovel
i
r
n
i
t
sa
r
ea
v
o
i
d
e
dw
i l
h
en
e
x
t
o
d
.
me

7
.
3
.
2 SequentlalQuadratlcProgrammlng(SQP)
Thel
a
c
ko
fr
o
b
u
s
t
n
e
s
si
neSLPduetoen
e
e
df
o
rmovel
i
r
n
i
t
sc
a
nbec
o
u
n
t
e
l
b
y
i
n
c
l
u
d
i
n
ga
ne
l
e
m
e
n
to
fn
o
n
l
i
n
r
i
t
yi
nep
r
o
b
l
e
m
.The
a
r
cs
e
v
e
r
a
lwaysl
h
i
sc
a
n
bea
c
h
i
e
v
e
d
.Onewayi
s1
0r
e
q
u
i
r
el
h
a
l s
e
a
r
c
hd
i
r
e
c
t
i
o
n
s(rememberi
na
c
a
1
i
m
p
l
e
m
e
n
t
a
t
i
o
nsXi
sS
)i
nE
q
u
a
t
i
o
n
s(
7.
31

)(
7.
34
)b
el
i
m
i
t
e
dbyt
h
eu
n
i
tc
i
r
c
l
ewhich
c
a
nbed
e
s
c
r
i
b
e
d c
o
n
s
t
r
a
i
n
l
:
STSS
:1

35
)
(
7.

290 N
U
M
E
A
I
C
A
LT
E
C
H
N
I
C
U
E
SF
O
AC
O
N
S
T
A
A
I
N
E
DOP
T
lM
I
Z
A
T
I
O
N

7
.
3

optimum MATLAB i
m
p
l
e
m
e
n
t
s(
a
so
ft
h
i
sw
r
i
t
i
n
g
)
. Having a
c
c
e
s
st
ot
h
e
O
p
t
i
m
i
z
a
t
i
o
nToolboxa
l
l
o
w
sNLPt
obes
o
l
v
e
du
s
i
n
gSQP.A d
e
t
a
i
l
e
dd
i
s
c
u
s
s
i
o
no
f
t
h
es
o
l
u
t
i
o
nt
e
c
h
n
i
q
u
ei
sa
l
s
oa
v
a
i
l
a
b
l
e[
1
1
]
.
l
eQPs
u
b
p
r
o
b
l
e
memployedi
nt
h
i
ss
u
b
s
e
c
t
i
o
ni
sb
a
s
e
done
x
p
a
n
d
i
n
gt
h
c
ec
o
n
s
t
r
a
i
n
t
sa
r
e
o
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
nq
u
a
d
r
a
t
i
c
a
l
l
ya
b
o
u
tt
h
ec
u
r
r
e
n
td
e
s
i
g
n
. l
e
x
p
a
n
d
e
dl
i
n
e
a
r
l
ya
si
nSLP[
8
]
. i
si
sc
a
l
l
e
ds
e
q
u
e
n
t
i
a
lq
u
a
d
r
a
t
i
cprogramming
(
S
Q
P
)
.

3
2

Minim

zo

dX

~J{.X): g
j
(X
j
)+V
gJ
{Xj
)
Td
.
XS0
; j=1.2.....m (
7.
33
)
j
s
x
}
o
wSsx
Sx~P;

3
4

-4

-3

-2

o
6x

F
l
g
u
7
.
9 l
Iu
s
t
r
a
t
l
o
no
fmoveI
I
m
l
t
s
.
s
t
e
p
s
i
z
ed
e
t
e
r
m
i
n
a
t
i
o
ns
h
o
u
l
dt
h
e
na
c
c
o
u
n
tf
o
rt
h
ea
c
t
u c
h
a
n
g
ei
nt
h
ed
e
s
i
g
n
v
e
c
t
o
r
. The p
r
o
b
l
e
me
x
p
r
e
s
s
e
d by Eq (
7.
31
)(
7.
35
)i
saq
u
a
d
m
t
i
c
programming(
Q
P
)s
u
b
p
r
o
b
l
e
m
.Ani
t
e
m
t
i
v
es
o
l
u
t
i
o
no
ft
h
i
ss
u
b
p
r
o
b
l
e
mc
b
eone
formo
f SQPme
o
d
.T
h
i
s odi
sp
o
p
u
l
a
randi
su
s
e
dw
i
d
e
l
y
.Whilet
h
eQP
i
n
t
s
.i
ti
ss
t
i
l
la
s
u
b
p
r
o
g
r
a
mi
s as
i
m
p
l
en
o
n
l
i
n
e
a
rp
r
o
b
l
e
mw
i
t
hl
i
n
e
a
rc
o
n
s
m
u
l
t
i
v
a
r
i
a
b
l
cn
o
n
l
i
n
e
a
rp
r
o
b
l
e
m
.F
o
r
t
u
n
a
t
e
l
yQP p
r
o
b
l
e
m
sa
p
p
e
a
ri
n many
d
i
s
c
i
p
l
i
n
e
sa
n
dt
h
e
r
ea
r
ee
f
f
i
c
i
e
n
tp
r
o
c
e
d
u
r
e
sanda
l
g
o
r
i
t
h
m
st
oa
d
d
r
e
s
st
h
e
s
ep
r
o
b
l
e
r
n
s
[
6
1
0
]
.Oneo
ft
h
s
em
e
t
h
o
d
si
sa
ne
x
t
e
n
s
i
o
no
ft
h
eS
i
m
p
l
e
xmethodo
fL
in
e
a
r
P
r
o
g
r
a
m
m
i
n
g
.
l
es
o
l
u
t
i
o
nt
o eQPi
sw
e
l
ld
o
c
u
m
e
n
t
e
d
.I
twouldl
r
u
i
r
emuchd
i
s
c
u
s
s
i
o
na
n
d
t
h
e
r
e
f
o
r
ei
sn
o
ti
n
t
r
o
d
u
c
e
dh
c
r
e
.R
e
a
d
e
r
sa
r
er
e
f
e
r
r
e
dt
ot
h
ev
a
r
i
o
u
sr
e
f
e
r
e
n
c
e
s
.T
h
i
s
i
sa
l
s
o ep
r
i
m
a
r
yr
e
a
s
o
nwhyt
h
eQPmethodi
sa
v
a
i
l
a
b
l
ei
nt
h
eO
p
t
i
m
i
z
a
t
i
o
n
o
r
m
a
lm
e
t
h
o
d
sf
o
rQPp
r
o
b
l
e
m
sa
r
eb
a
s
e
done
f
f
i
c
i
e
n
c
yi
n
T
o
o
l
b
o
xi
nMATLAs.Thef
t
h
ev
a
r
i
o
u
ss
t
a
g
e
so
ft
h
ei
m
p
l
e
m
e
n
t
a
t
i
o
n
.Ani
n
e
f
f
i
c
i
e
n
tb
u
ts
u
c
c
e
s
s
f
u
ls
o
l
u
t
i
o
n1
0t
h
e
p
r
o
b
l
e
mc
a
nbeo
b
t
a
i
n
e
dbyu
s
i
n
gI
h
eALMo
ft
h
ep
r
e
v
i
o
u
ss
e
c
t
i
o
nt
os
o
l
v
et
h
eQP
s
u
b
p
r
o
b
l
e
m
.Af
t
e
ra
l
l
.c
o
m
p
u
t
i
n
gr
e
s
o
u
r
c
e
sa
r
en
o
tac
o
n
s
t
r
a
i
n
tt
o
d
a
y
.I
nl
i
e
uo
ff
o
r
m
a
l
u
m
e
r
i
c
a
lmethodi
sa
t
t
e
m
p
di
nt
h
i
s
d
i
s
c
u
s
s
i
o
n
.a
ni
n
t
u
i
t
i
v
eu
n
d
e
r
s
t
a
n
d
i
n
gofen
et
r
a
n
s
l
a
t
i
o
no
fSQPi
n
t
oc
o
d
ei
sn
o
tt
r
i
v
i
a
la
n
dt
h
emethodi
su
s
u
a
l
l
yt
h
eb
a
s
i
s
b
o
o
k
.Th
o
fc
o
m
m
e
r
c
i
a
ls
o
f
t
w
a
r
e
.F
o
re
x
a
m
p
l
e
SQPi
st
h
eo
n
l
ya
l
g
o
r
i
t
h
mf
o
rc
o
n
s
t
r
a
i
n
e
d

r
1
(
d
.
X
)=
f
'
)+Vf(Xi
d
.
X+
TV2fl

)+
S
u
b
j l
t
o
: ():h
k
(
X
VhI(X)=0
; k=1
.
2
.
.
.
.
.
1(
7.
32
)
j

-1

5
;

D
I
A
E
C
TMETHODSFOAC
O
N
S
T
A
A
I
N
E
DOPT
lM
I
Z
A
T
I
O
N 2
9
1

i=1
.
2.

.
.n

(
7
.
3
4
)

I
nE
q
u
a
t
i
o
n(
7
.
3
6
)V(X
)i
st
h
eH
e
s
s
i
a
nm
a
t
r
i
x
.I
na
c
t
u
a
li
m
p
l
e
m
e
n
t
a
t
i
o
nt
h
er
e
a
l
H
e
s
s
i
a
nw
i
l
ln
o
tb
eu
s
e
d
.I
n
s
t
e
a
dam
e
t
r
i
c] i
supda
dw
i
e
a
c
hi
t
e
r
a
t
i
o
ni
s
sb
a
s
e
don s
u
c
c
e
s
so
ft
h
eV
a
r
i
a
b
l
eMe
cM
ethods(V

.
f
)o
ft
h
e
u
s
e
d
. si
p
r
e
v
i
o
u
sc
h
a
p
t
e
r
.S
e
v
e
r
a
lr
e
s
e
a
r
c
h
e
r
s[
6
.
9
]h
a
v
eshown BFGSu
p
d
a
t
ef
o
rt
h
e
i
c
i
e
n
ti
m
p
l
e
m
e
n
t
a
t
i
o
no
ft
h
eSQPm
e
t
h
o
d
.T
h
i
sQPi
sw
e
l
l
H
e
s
8
i
a
np
r
o
v
i
d
e
sa
ne
p
o
s
e
da
n
dc
o
n
v
e
xands
h
o
u
l
dy
i
e
l
das
o
l
u
t
i
o
n
.S
o
l
u
t
i
o
nt
ot
h
i
ss
u
b
p
r
o
b
l
e
mi
sa
t
h
e
a
r
to
feSQPm
e
t
h
o
d
.A c
o
n
s
i
d
e
r
a
b
l
eamounto
fr
e
s
e
a
r
c
hh
a
sb
e
e
ni
n
v
e
s
t
e
di
n
s s
u
b
p
r
o
b
l
e
m
. I
n a f
o
r
m
a
l
d
e
v
e
l
o
p
i
n
ge
f
f
i
c
i
e
n
t t
e
c
h
n
i
q
u
e
s t
oh
a
n
d
l
e i
nE
q
u
a
t
i
o
n
s(
7.
36
) d(
7.
32
)(
7
.
3
4
)m
u
s
lbe
i
m
p
l
e
m
e
n
t
a
t
i
o
no
f SQPd.Xi
p
l
a
c
e
dbyt
h
es
e
a
r
c
hd
i
r
e
c
t
i
o
nS
.TheQPf
o
rSa
l
8
0m
o
d
i
f
i
e
st
h
ec
o
n
s
t
r
a
i
n
te
q
u
a
t
i
o
n
s
e
s
p
e
c
t
t
oec
u
r
r
e
n
ta
c
t
i
v
ec
o
n
s
t
r
a
i
n
t
s
.
8
0t
h
a
taf
e
a
s
i
b
l
ed
i
r
e
c
t
i
o
nc
a
nbef
o
u
n
dwi r
Sec
oncema
b
o
u
tmovingl
i
n
e
a
r
l
yt
oap
o
i
n
tona
na
c
t
i
v
e
F
i
g
u
r
e7
.
1
0i
l
l
u
s
a
i
nt
.XO
t
h
ec
u
r
r
e
n
td
e
s
i
g
ni
sona
na
c
t
i
v
ec
o
n
s
t
r
a
i
n
.
tI
ft
h
es
e
a
r
c
hd
i
r
e
c
t
i
o
nS
I
c
o
n
8
f
o
l
l
o
w
s l
i
n
e
a
r
i
z
e
df
u
n
c
t
i
o
na
tXO(
t
h
et
a
n
g
e
n
tt
ot
h
ef
u
n
c
t
i
o
na
tt
h
a
tp
o
i
n
t
)
.a
n
y
hd
i
r
e
c
t
i
o
n
.howevers
m
a
l
lw
i
l
lc
a
u
s
et
h
ec
o
n
s
t
r
a
i
n
tt
ob
e
s
t
e
p
s
i
z
e
.a
l
o
n
g E
v
i
o
l
a
t
e
d
.I
no
r
d
e
rt
od
e
t
e
r
m
i
n
ean
e
i
g
h
b
o
r
i
n
gp
o
i
n
tt
h
a
tw
i
l
ls
t
i

ls
a
t
i
s
f
yt
h
ec
o
n
s
t
m
i
n
t
s
.
as
e
a
r
c
hd
i
r
e
c
t
i
o
ns
l
i
g
h
t
l
yl
e
s
st
h et
a
n
g
e
n
ti
su
s
e
d
.Thi
si
n
t
r
o
d
u
c
e
sal
a
c
ko
f
c
o
n
s
i
s
t
e
n
c
ya
st
h
ed
e
v
i
a
t
i
o
nfromt
a
n
g
e
n
c
ybecomesam
a
t
t
e
ro
fi
n
d
i
v
i
d
u
a
le
x
p
e
r
i
e
n
c
e
.E
x
p
e
r
i
m
e
n
t
sh
a
v
es
u
g
g
e
s
t
e
da
t90-95%o
ft
h
et
a
n
g
e
n
ti
sau
s
e
f
u
lf
i
g
u
r
e
a
n
d
p
r
a
c
t
i
a
l
t
h
o
u
g
hmakingi
ta
sc
l
o
s
et
o100%i
srecommended[
2
]
.

b
es
e
a
r
c
hd
i
r
e
c
t
i
o
nf
i
n
d
i
n
g
QPs
u
b
p
r
o
b
l
e
mi
s

M
i
n
i
m
i
z
e 7
(
S
)=f(X
f
(
X
j
)
TS
T
[
H]S
)+V
1
S
u
b
j
e
c
l

(S):cMXi)+VhI(X
)S=O;

(
7
.
3
7
)

k=1
.
2
.
.
.
.
1

(
7
.
3
8
)

(
S
)
:c
g
i
<
X
i
)+V8
.

x
)
'
fSS0
; j=1
.2
.
.
m

(
7.
39
)

s
}
OWSSjSS
r
"
; i=I.2.....n

(
7
.
4
0
)

292

CONSTRA1NEDOPT1M
lZA
TION
NUMERICALTECHN1QUESFO

7
.
3 DIRECTMTHODSFORCONSTRAINEDOPTIMIZATION

x
j
=
x
;
'1+S

(
7.
41
)

nu

m
A

r
e

.4
2
)

nu

Rur

mTL-H m

Jr
A

FJ

t
n

.
r

''E
J

AV

AV

M
i
n
i
m
i
z
e
:

n
i
m
i
z
e
:
Mi

293

(
7.
4
3
)

E
q
u
a
t
i
o
n(
7.
4
2
)o
r(
7
.
4
3
)c
a
nbeu
s
e
dt
oo
b
t
a
i
n ev
a
l
u
egs
t
e
p
s
i
z
e

.Th
ef
u
n
c
t
i
o
ni
n
43
)r
e
q
u
i
r
e
sa
ni
m
p
o
r
t
a
n
to
b
s
e
r
v
a
t
i
o
n
.S
h
o
u
l
dt
h
eL
a
g
r
a
n
g
em
u
l
t
i
p
l
i
e
r
s
E
q
u
a
t
i
o
n(
7.
bec
a
l
c
u
l
a
t
e
df
o
re
a
c
hv
a
l
u
eof?Im
a
g
i
n
eu
s
i
n
gt
h
eg
o
l
d
e
ns
e
c
t
i
o
nt
os
o
l
v
eE
q
u
a
t
i
o
n
4
3
)w
i
t
hem
u
l
t
i
p
l
i
e
r
sc
a
l
c
u
l
a
t
e
domt
h
es
a
f
a
c
t
i
o
no
f eKubnTuc
k
e
r
(
7.
c
o
n
d
i
t
i
o
n
sf
o
re
a
c
hnewv
a
l
u
eo
fX
.S
e
v
e
r
a
lu
p
d
a
t
i
n
gs
t
r
a
t
e
g
i
e
sf
o
r m
u
l
t
i
p
l
i
e
r
si
n
Eq
u
a
t
i
o
n(
7
.
4
3
)c
a
nbef
o
u
n
di
ne
l
i
s
t
d
r
e
f
e
r
e
n
c
s
.As
i
m
p
l
eonei
st
oh
o
l
dt
h
ev
a
l
u
e
s
r
e
v
i
o
u
si
t
e
r
a
t
i
o
nv
a
l
u
ea
n
dc
h
a
n
g
ethema
teend
o
ft
h
em
u
l
t
i
p
i
e
r
sc
o
n
s
t
a
n
t ep
o
ft
h
ei
t
e
r
a
t
i
o
n
.

>
<
"1

F
l
g
u
r
e
7
.
1
0M
o
v
l
n
gl
i
n
e
a
rt
oa
na
c
t
i
v
ec
o
n
s
t
r
a
i
n
t
.
Thes
i
d
ec
o
n
s
t
r
a
i
n
t
sa
r
en
e
c
e
s
s
a
r
y i
si
saw
e
l
l
p
o
s
e
dp
o
tn
ef
a
c
t
o
rci
r
o
b
l
e
m
. l
s
t
h
ea
d
j
u
s
t
m
e
n
tf
o
rmovingt
a
n
g
e
n
tt
t
.
l
ot
h
ec
o
n
s
t
r
a
i
n
es
u
g
g
e
s
t
e
dv
a
l
u
ef
o
rci
s

ev
o
rt
h
ei
n
e
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
nt
0
.
9
0
.
9
5
.F
a
l
u
eo
fci
s1i
ft
h
ec
o
n
s
t
r
a
i
n
ti
ss
a
t
i
s
f
i
e
d
(
s
h
o
u
l
db
ea
b
l
et
omovet
a
n
g
e
n
tt
ot
h
ec
o
n
s
t
r
a
i
n
t
)
.

S
t
e
p
s
i
z
eC
a
l
c
u
l
a
t
i
o
n
:On es
e
a
r
c
hd
i
r
e
c
t
i
o
nh
a
sb
e
e
nd
e
t
e
r
m
i
n
e
d
ti
i
s
Yt
n
e
c
e
s
s
oc
a
l
c
u
l
a
t
et
h
es
t
e
p
s
i
z
e
.S
i
n
c
et
h
emethodso
h
a
p
t
e
rh
a
v
es
f sc
e
e
na
n
i
nt
h
enumbero
fc
a
l
c
u
l
a
t
i
o
n
s
enormousi
n
c
r
e
a
s
t
h
es
t
e
p
s
i
z
ec
a
l
c
u
l
a
t
i
o
n
sa
r
eb
a
s
e
d
e
c
l
'
s
lUl
t
a
n
e
o
u
s
l
yd
i
n
geo
i
v
ea
sw
b
j t
e
l
la
ons
m
p
r
o
v
i
n
g c
si
o
n
s
t
r
a
i
n
t
r
d
e
rt
oa
s
a
t
i
s
f
a
c
t
i
o
n
.I
no
c
c
o
m
p
l
i
s
ht
o
a
las
u
i
t
a
b
l
ef
h
i
sg
u
n
c
t
i
o
nmustbed
n
d
e
v
e
l
o
p
e
da
e
t
e
r
m
i
n
a
t
i
o
nb
es
t
e
p
s
i
z
ed
a
s
e
donm
i
n
i
m
i
z
i
n
gt
h
i
su
n
c
o
n
s
t
r
a
i
n
e
df
u
n
c
t
i
o
n
.S
e
v
e
r
a
l
e
f
e
rt
oi
ta
st
h
d
f
l
l
n
c
t
i
o
n
.T
h
ed
e
s
c
e
n
t
e
s
c
r
i
p
t
i
o
no
a
u
t
h
o
r
sr
h
i
sf
ft
u
n
c
t
i
o
nv
r
o
m
a
r
i
e
sf
h
eLa
or
e
s
e
a
r
c
h
e
r
.Somed
sa
st
r
e
s
e
a
r
c
h
e
rt
e
f
i
n
ei
g
r
a
n
g
i
a
nf
a
s
e
u
n
c
t
i
o
ni
h
i
c
bc
nw
u
l
t
i
p
l
i
e
r
s must be d
e
t
e
r
m
i
n
e
d
. Many o
L
a
g
r
a
n
g
em
t
h
e
r
si
n
c
l
u
d
eo
n
l
yt
h
e
d
o
p
t EPFformf
i
t
hap
e
n
a
l
t
ym
c
o
n
s
t
r
a
i
n
t
sw
n
da
o
r f
u
n
c
t
i
o
n
.S
t
i
l
l
u
l
t
i
p
l
i
e
ra
n
l
yt
n
c
l
u
d
eo
h
emaximumv
o
n
s
t
r
a
i
n

z
o
t
h
e
r
si
t
.Twof
o
r
m
so
ft
h
ef
u
n
c
t
i
o
na
i
o
l
a
t
e
dc
t
e
db
e
l
o
w
:

is

C
a
l
c
u
l
a
t
i
n
gt
h
eM
u
l
f
l
p
t
t
e
r
s
:Thepopularity oft
h
e SQPmodi
st
h
a
ti
t
e
s
s
i
a
nb
a
s
e
d on t
h
eL
a
g
r
a
n
g
i
a
no
ft
h
ep
r
o
b
l
e
m
.
a
t
t
e
m
p
t
st
ou
p
d
a
t
e e H
C
a
l
c
u
l
a
t
i
n
gt
h
em
u
l
t
i
p
l
i
e
r
si
sn
e
c
e
s
s
a
r
yt
ov
e
r
i
f
yt
h
eKuhn-Tuckerc
o
n
d
i
t
i
o
n
sa
t
t
h s
o
l
u
t
i
o
n
.C
a
l
c
u
l
a
t
i
n
gt
h
em
u
l
t
i
p
l
i
e
r
s
f
o
rt
h
ec
u
r
r
e
n
tv
a
l
u
eo
ft
h
ed
e
s
i
g
nv
e
c
t
o
r
i
sn
e
c
e
s
s
a
r
yf
o
ra
d
d
i
t
i
o
n
a
lo
p
e
r
a
t
i
o
n
s
i
n
c
l
u
d
i
n
gs
e
t
t
i
n
gupt
h
ed
e
s
c
e
n
tf
u
n
c
t
i
o
ni
n
E
q
u
a
t
i
o
n(
7.
43
)a
sw
e
l
la
su
p
d
a
t
i
n
gt
h
em
e
t
r
i
cr
e
p
l
a
c
e
m
e
n
tf
o
rt
h
eH
e
s
s
i
a
n
(
f
o
l
l
o
w
snex
t
)
.Thec
a
l
c
u
l
a
t
i
o
nu
s
e
st
hFOCb
a
s
e
dont
h
eL
a
g
r
a
n
g
i
a
n
.Thev
a
l
u
e
s
o
ft
h
eg
r
a
d
i
e
n
t
so
ft
h
ev
a
r
i
o
u
sf
u
n
c
t
i
o
n
sa
r
ecomputeda
tt
h
ec
u
r
r
e
n
td
e
s
i
g
n
.The
h
e
no
b
t
a
i
n
e
da
sas
o
l
u
t
i
o
nt
oas
e
tofl
i
n
e
a
re
q
u
a
t
i
o
n
s
.Th
e
m
u
l
t
i
p
l
i
e
r
s t
s
o
l
u
t
i
o
n
st
oQPandSQPp
r
o
b
l
e
m
su
s
u
a
l
l
yt
a
k
ea
d
v
a
n
t
a
g
eoft
h
a
c
t
i
v
ec
o
n
s
t
r
a
i
n
t
s
e
t
.Thisi
n
c
l
u
d
e
sa
l
le
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
t
sando
n
l
yt
h
o
s
ei
n
e
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
t
st
h
a
t
t
a
r
ta
taf
e
a
s
i
b
l
ep
o
i
n
t
.The
a
r
ea
c
t
i
v
e
.T
h
i
si
m
p
l
i
e
st
h
a
t SQP i
t
e
r
a
t
i
o
n
ss
m
u
l
t
i
p
l
i
e
r
sc
o
r
r
e
s
p
o
n
d
i
n
gt
oi
n
a
c
t
i
v
ec
o
n
s
t
r
a
i
n
t
sa
r
es
e
tt
oz
e
r
o
.Ther
e
m
a
i
n
i
n
g
m
u
l
t
i
p
l
i
e
r
sc
a
nbee
s
t
a
b
l
i
s
h
e
dc
o
n
s
i
d
e
r
i
n
gas
e
tofa
na
p
p
r
o
p
r
i
a
t
enumbero
fl
i
n
e
a
r
i
n
d
e
p
e
n
d
e
n
te
q
u
a
t
i
o
n
sfromE
q
u
a
t
i
o
n(
7
.
1
2
)
:

J
e
F e
J
f "d
h
oh n e
}
g1.
.n e
}
g
+ +L-+PI+ +~n !!!.=O;
a
x
i axiftaxitaxif l axiiaaxi

= +

i=1

2
.
.
.n
(
7.
12
)

Rep
c
i
n
gHessianw
i
t
hBFGSU
p
d
a
t
e
:l
nE
ql
n (
7.
37
)c
a
l
c
u
l
a
t
i
n
gt
h
e
o
n
v
e
r
g
e
st
ot
h
e
a
c
t
u
a
lH
e
s
s
i
a
ni
ss
t
r
o
n
g
l
yd
i
s
c
o
u
r
a
g
e
d
.S
i
n
c
et
h
eBFGSmodc

t
h
eBFGSm
e
t
r
i
ci
sc
o
n
s
i
d
e
r
e
dt
obea
ne
x
c
e
l
1
e
n
tr
e
p
l
a
c
e
m
e
n
to
ft
h
eH
e
s
s
i

H
e
s
s
i
a
n
[
9
]
.Theme
ci
sb
a
s
e
dont
h
eL
a
g
r
a
n
g
i
a
no
fef
u
n
c
t
i
o
n[
12
]
:

7
.
3

2
9
4 N
U
M
E
R
I
C
A
LT
E
C
H
N
I
O
U
E
SF
O
RC
O
N
S
T
R
A
lN
E
DO
P
T
I
M
I
Z
A
T
I
O
N

A
p
p
l
i
c
a
t
i
o
n01SQP:Th
S
QPt
e
c
h
n
i
q
u
ei
sa
p
p
l
i
e
dt
oE
x
a
r
n
p
l
e7
.1
.T
h
e
r
ea
r
e
s
e
v
e
r
a
lw
a
y
st
oa
c
c
o
m
p
l
i
s
ht
h
i
s
.Thea
p
p
l
i
c
a
t
i
o
nc
a
nu
s
eg
r
a
p
h
i
c
a
ls
u
p
p
o
r
ta
sshown
w
i
t
ht
h
eSLP(
t
h
i
si
sa
ne
x
e
r
c
i
s
e
)
.Thea
p
p
l
i
c
a
t
i
o
nc
a
nu
s
et
h
eQPp
r
o
g
r
a
mf
r
o
m
sb
o
o
k QPp
r
o
b
l
e
mi
ss
o
l
v
e
db
ya
p
p
l
y
i
n
geKT
MATLAB(
a
l
s
oa
ne
x
e
r
c
i
s
e
)
.
l
ni
r
a
l
lf
o
ri
ss
p
e
c
i
f
i
ce
x
a
m
p
l
e
c
o
n
d
i
t
i
o
n
s
w
b
i
c
hi
st
h
ef
e
a
t
u
r
eo
f
t
h
i
ss
u
b
s
e
c
t
i
o
n
.A
t
i
ef
u
n
c
t
i
o
n
sa
r
ea
n
a
l
y
t
i
c
a
l
l
yd
e
f
i
n
e
da
n
de
i
r
s
t
r
u
c
i
ss
i
m
p
l
e
.Wec
a
nh
a
m
e
s
st
h
e
t
oc
a
l
c
u
l
a
t
eed
e
r
i
v
a
t
i
v
e
sa
n
dea
c
t
u
a
lH
e
s
s
i
a
n
.As
s
y
m
b
o
l
i
co
p
e
r
a
t
i
o
n
s MATL
i
nec
a
s
eo
fSLP
t
h
ea
c
t
u
a
lc
h
a
n
g
i
nt
l
Xi
sc
o
m
p
u
t
e
d[
E
q
u
a
t
i
o
n
s(
7
.
3
6
)
.
(
7
.
3
2
)
(
7
.
3
4
)
]r
a
t
h
e
rt
h
a
nt
h
es
e
a
r
c
hd
i
r
e
c
t
i
o
nSf
o
rt
h
eQPs
u
b
p
r
o
b
l
e
m
.
o
ds
y
m
b
o
l
i
c
a
l
l
yt
oE
x
a
m
p
l
e7
.1
.l
ti
d
e
n
t
i
f
i
e
s
Sec7
_3_2.ma
p
p
l
i
e
st
h
eSQPme
t
h
es
o
l
u
t
i
o
nu
s
i
n
geLag
r
a
n
g
em
e
t
h
o
d(
C
h
a
p
t
e
r5
)
.I
tp
romp
f
o
rt
h
ec
u
r
r
e
n
td
e
s
i
g
n
v
e
c
t
o
r
.I
ts
o
l
v
e
st
h
eQPa
n
a
l
y
t
i
c
a
l
l
ya
p
p
l
y
i
n
gt
h
eKTc
o
n
d
i
t
i
o
n
sa
n
dp
r
o
v
i
d
i
n
ge
o
r
r
e
s
p
o
n
d
st
os=0(
gm
u
s
tb
el
e
s
sa
nz
e
r
o
)
.C
a
s
e
r
e
s
u
l
t
sf
o
rt
h
et
w
oc
a
s
e
s
.C eac
s>0
)
.Theo
u
t
p
u
tf
r
o
mr
u
n
n
i
n
gt
h
ep
r
o
g
r
a
mi
sr
e
c
o
r
d
e
db
e
l
o
w
bc
o
r
r
e
s
p
o
n
d
st
og=0(
f
o
rs
e
v
e
r
a
li
t
e
r
a
t
i
o
n
si
ni
si
m
p
l
e
m
e
n
t
a
t
i
o
no
fS
Q
P
.Ana
1
y
t
i
c
a
ls
o
l
u
t
i
o
nw
i
t
hu
s
e
r
.S
ec7
_3_2.mm
u
s
tb
er
u
nf
o
re
a
c
hi
t
e
r
a
t
i
o
n
.Yous
h
o
u
l
db
ea
b
l
e
i
n
t
e
r
v
e
n
t
i
o
ni
su
s
t
oc
o
p
ya
n
dp
a
s
t
d
e
s
i v
e
c
t
o
ri
n commandw
i
n
d
o
w
.

)+LAkVhk(X 1
)+ s
j
V
C
j
(
X
'

'
)
p=Vf(X
1
<
=1
j
=
I
V
j
e
J
xq

k'Vhk{X~ +LsVg
i
)

-Vf
(X
q
)+

(
7.
4
4
)

j
=l
l
t
j
e
J

1
<
=1

D
I
R
E
C
TM
E
T
H
O
D
SF
O
RC
O
N
S
T
R
A
I
N
E
DO
P
T
I
M
I
Z
A
T
I
O
N 295

q
y=X
q+I_X
+
l
]= J
+
[
H
q
p
'
y y[
H
q
]
y

jEJi
mp
I
ie
st
h
es
e
to
fa
c
t
i
v
ei
n
e
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
t
s
.I
ti
sr
e
c
o
m
m
e
n
d
e
d me
c

b
ek
e
p
tp
o
s
i
t
i
v
ed
e
f
i
n
i
t
ea
smucha
sp
o
s
s
i
b
l
ee
v
e
ni
fi
ti
sl
i
k
e
l
y a
ti
twi
I
ln
o
lb
e
p
o
s
i
t
i
v
ed
e
f
i
n
i
t
ea
tt
h
es
o
l
u
t
i
o
n
.Onewayo
fe
n
s
u
r
i
n
gi
si
st
ov
e
r
i
f
yt
h
a
t
pTy>O

A'
o
r
l
t
h
m
:SequentialQuadraticProgrammlng(SQ
7
.

S
t
e
p1
.C
h
o
o
s
eXI
1
(
n
o
.o
fi
t
e
r
a
t
i
o
n
s
)
.
E
/
S(
f
o
rc
o
n
v
e
r
g
e
n
c
ea
n
ds
t
o
p
p
i
n
g
)
q=I(
i
t
e
r
a
t
i
o
nc
o
u
n
l
e
r
)
S
t
e
p2
.C
a
l
lQPt
oo
p
t
i
m
i
z
e(
7.
37
)(
7.
4
0
)(X
q
)
1=x
O
u
t
p
u
t
:S
a
l
c
u
l
a
t
i
o
n
Useac
o
n
s
t
r
a
i
n
e
d c
AX="s
Xq+1=Xq+AX

S
t
e
p3
.C
o
n
v
e
r
g
e
n
c
ef
o
rSQP
fork=1
.
2

.
.
.
.
l
;
Ifhk=O
~O f
o
r
j
=1
.
2

.
.
.
m;
I
fKTc
o
n
d
i
t
i
o
n
sa
r
es
a
t
i
s
f
i
e
d
nC
o
n
v
e
r
g
e
d
.S
t
o
p
S
t
o
p
p
i
n
gC
r
i
t
e
r
i
a
:
AX=xq+l_xq
T
I
fAX
AX~ E
I
:S
t
o
p(
d
e
s
i
g
nn
o
tc
h
a
n
g
i
n
g
)
t
o
p(m
umi
t
e
r
a
t
i
o
n
sr
e
a
c
h
e
d
)
I
fq=Ns:S
C
o
n
t
i
n
u
e
UpdateMe
c[

qq+I

nnr
nS
t
p
n

XI=[3

I
t
e
r
a
t
i
o
n1
:

f
o
re
a
c
hi
t
e
r
a
t
i
o
n
.

l
2

=
[

M
1
;

Vh=

[~Jg=42

d
.
xl
=
1
.0
5
9
1
;

1
.
16
1
3

=0
; g
=-0
.
8
2
2
5
;
X1=1
.
9
4
0
9
;
i

I
t
e
r
a
t
i
o
n2
:

=0.8387

1
.9
4
0
90
.
8
3
8
7
]
X2=[

3.
12
3
; = U Z ;

7
0
5
; Vh=[

1222

g=0.4693; =11

Thes
o
I
u
t
i
o
nt
ot
h
eQPp
r
o
b
l
e
mi
sC
a
s
ea
:
=0
d
.
xl
.
6
1
8
6
;d
.
x
O.
0
0
4
1
2=h=O;

g=1.8281;

2
9
6 N
U
M
E
R
I
C
AlT
E
C
H
N
I
Q
U
E
SF
O
RC
O
N
S
T
R
A
I
N
E
DO
P
T
I
M
I
Z
A
T
I
O
N
1=1
.
32
2
2
;

7
.
3

=0.7975

7
.
3
.
3 G
e
n
e
r
a
l
l
z
e
dReducedG
r
a
d
l
e
n
t(GRG)Method

1
.3
2
2
20
.
7
9
7
5
)
X3=[

I
t
e
r
a
t
i
o
n3
:

D
I
R
E
C
TM
E
T
H
O
D
SF
O
RC
O
N
S
T
R
A
I
N
E
DO
P
T
I
M
I
Z
A
T
I
O
N 2
9
7

1
GRGm
e
t
h
o
di
sa
n
o
t
h
c
rp
o
p
u
l
a
rt
e
c
h
n
i
q
u
ef
o
rc
o
n
s
t
r
a
i
n
e
do
p
t
i
m
i
z
a
t
i
o
n
.l
e

1m
e
t
h
o
d
t
h
eR
e
d
u
c
e
dG
r
a
d
i
e
n
tM
e
t
h
o
d[
1
3
]h
a
ss
e
e
ns
e
v
e
r
a
1d
i
f
f
e
r
e
n
t
o
r
i
g
i
na
r
a
1r
e
s
e
a
r
c
h
e
r
s[
1
4
)
.L
ik
e SQPi
si
sa
oa
c
t
i
v
es
e
t
c
u
s
t
o
m
i
z
a
t
i
O
D
Sd
u
et
os
e
v
e
m
e
t
h
o
d
-d
e
a
l
sw
i
t
ha
c
t
i
v
ei
n
e
q
u
a
l
i
t
i
e
s
h
e
n
c
et
h
em
e
t
h
o
di
oc
Iu
d
e
s u
a
l
i
t
y
i
n
t
so
n
l
y
.Thei
n
u
a
l
i
t
yc
o
n
s
i
n
t
s E s
f
o
r
m
e
dt
oe
q
u
a
1
i
t
yc
o
a
i
n
t
s
c
o
n
s
u
s
i
o
gal
i
n
e
a
rs
l
a
c
kv
a
r
i
a
b
l
eo
ft
h
et
y
p
eu
s
e
di
nLPp
r
o
b
l
e
m
s
.

n
eg
e
n
e
r
a
1o
p
t
i
m
i
z
a
t
i
o
n
p
r
o
b
l
e
mi
sr
e
s
t
r
u
c
t
u
r
e
da
s

2281
=0
=!?
;
;
l

f=4.2126;

0
8
5
8
; Vg

TheS
O
I
U
t
i
O
Dt
ot
h
eQPp
r
o
b
l
e
mi
sC
a
s
ea
:
dxl=0
.
2
6
8
5
; dx
.
2
0
4
2
2=+0
h=0
;

g=0
.
0
1
9
0
;

X1=1
.
0
5
3
7
; =1
.0
0
1
8

I
t
e
r
a
t
i
o
n4
:

X =[
1
.0
5
3
7 1
.
0
0
1
8
]

f=3
.
0
6
8
5
;

Vf=~J 22Ul

;vh=I2124;g=0 =lTZl
I!~

(
7
.
4
5
)

S
u
b
j
e
c
tt
o
: h
t
(
X
I
'x
2
'
x

)=0
k=1

2
.
.
.
1

(
7.
46
)

8
J
(
XI
oX
2
'...
x

)+x

t
+
j=0
j=1

2 .
.

4
7
)
(
7.

xlsx;sxr i
=1

.
.
.
n

(
7.
48
)

X
n
+
j~ 0 j=1

2
.
.
m

(
7.
4
9
)

Thet
o
t
a
ln
u
m
b
e
ro
fv
a
r
i
a
b
l
e
si
sn+m
.T
hea
c
t
i
v
e!
:
e
t!
:
t
r
n
t
e
g
yao
;
s
u
m
e
st
h
a
ti
ti
s
p
o
s
s
i
b
l
e
d
e
f
i
n
eo
rc
h
o
o
s
ens
u
b
s
e
to
f(
1
1-1
)i
n
d
e
p
e
n
d
e
n
tv
a
r
i
a
b
l
e
somt
h
es
e
to
f
(
l+m)e
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
t
s
.Oncet
h
e
s
ev
a
r
i
n
b
l
e
sa
r
ee
s
t
a
b
I
is
h
e
dl
h
ed
e
p
e
n
d
e
n
t
l+m
)c
a
nb
er
e
c
o
v
e
r
e
dt
h
r
o
u
g
h c
o
n
sL
r
a
i
n
t
s
.Then
u
m
b
e
ro
fi
n
d
e
p
e
n
d
e
n
t
v
a
r
i
a
b
l
e
s(
v
a
r
i
a
b
l
e
si
sb
a
s
e
do
nco
r
i
g
i
n
n
ld
e
s
i
g
nv
a
r
i
n
b
l
c
sa
n
dt
h
eo
r
i
g
i
n
a
le
q
ua
1i
t
yc
o
n
s
t
r
a
i
n
t
s
o
ft
h
ep
r
o
b
l
e
m
.Thef
o
1o
l
w
i
n
gd
e
v
e
l
o
p
m
e
n
ti
sn
d
a
p
t
e
df
r
o
mR
e
f
e
r
e
n
c
e2
.
X
]i
sp
a
r
t
i
t
i
o
n
e
di
n
t
oe
i
n
d
e
p
e
n
d
e
n
t
s
e
t
[
Z
] d
Thes
e
to
ft
h
ed
e
s
i
g
nv
a
r
i
a
b
l
e
s[
1
1
.
d
e
p
e
n
d
e
n
ts
e
t[

1
.=1
.0
0
1
4
; =1.
;

E E -EJ

rE E -EEah

=0; g=O;

ZY

0.9275

d
x
.=-0
.
0
5
2
3
; L
1
x
z
=
O
.
0
0
1
7
;

M
i
n
i
m
i
z
e f
(
X
I
'X
2
.

Xn
)

Thec
o
n
s
t
r
a
i
n
t
sa
r
ea
c
c
u
m
u
l
a
t
e
da
s

f
=
3
.
0
0
1
4
; h=O. 2
7
; g=0
.
0
0
0
6

H=[:]
Thes
o
l
u
t
i
o
ni
sa
1
m
o
s
tc
o
n
v
e
r
g
e
da
n
df
u
r
t
h
c
ri
t
e
r
a
t
i
o
n
sa
r
el
e
f
tt
ot
h
er
e
a
d
e
r
.The
n
e
x
ti
t
e
r
a
t
i
o
ns
h
o
u
l
dp
r
o
d
u
c
ev
a
1
u
e
sf
o
rc
o
n
v
c
r
g
e
n
c
e
.
Th
eSQPh
e
r
et
a
k
e
sf
i
v
ei
t
e
r
a
t
i
o
n
st
oc
o
n
v
e
r
g
e
.Thi
sw
a
sb
a
s
e
do
nc
o
m
p
u
t
i
n
gt
h
e
a
c
t
ua
1H
e
s
s
i
a
n
.TheSLPt
o
o
kf
o
u
r
.TheALMt
o
o
kf
i
v
e
.F
o
rE
x
a
m
p
l
e7.1eSQP
oh
a
v
e ys
i
g
n
i
f
i
c
a
o
ta
d
v
a
o
t
a
g
ew
h
i
c
hi
su
n
e
x
p
e
c
t
e
d
.TheALM
d
o
e
sn
o
ta
p
p rt
t
t
r
a
c
t
i
v
e
. The SQP i
n Sec7
_3_2.m i
si
m
p
l
e
m
e
n
t
e
d
s
l
a
r
t
st
oa
p
p
e
a
rq
u
i a
a
n
a
l
y
t
i
c
a
1
l
yb
a
s
e
do
nKTc
o
n
d
i
t
i
o
n
s
.I
ti
sv
e
r
yd
i
f
f
e
r
e
n
tt
h
a
nA
l
g
o
r
i
t
h
m(A
7
.
5
)w
h
i
c
h
s
e
n
t
san
u
m
e
r
i
c
a
li
m
p
l
m
e
n
t
a
t
i
o
n
.
r
e
p

1
a
t
i
o
n
s(
7.
4
5
)(
7
.
4
9
)c
a
nb
er
e
c
a
s
la
s
Re
(

ZY
)
M
i
n
i
m
i
z
e f

(
7
.
5
0
)

S
u
b
j
e
c
t1
0
: H
(

ZY
)=0

(
7
.
5
1
)

z
!
Sz
;
S
z
f i=1

.
.
.
n-l

(
7
.
5
2
)

'
1
1
1
1
1

298 N
U
M
E
R
I
C
A
LT
E
C
H
N
I
Q
U
E
SF
O
RC
O
N
S
T
A
A
I
N
E
DO
P
T
I
MlZAT
I
O
N
a
li
m
p
l
m
e
n
t
a
l
i
o
nt
h
e
As
i
m
i
l
a
rs
e
to
fs
i
d
ec
o
n
s
t
r
a
i
n
t
sf
o
rYc
a
nbee
x
p
r
e
s
s
e
d
.1
na
c
i
n
a
c
t
i
v
ei
n
u
a
i

t
yc
o
n
s
t
r
a
i
n
t
sa
r
en
o
ti
n
c
1udedi
n p
r
o
b
l
e
m
.A
l
s
o
t
h
es
l
a
c
k
ot
h
ei
n
d
e
p
e
n
d
e
n
t . I
nGRG l
i
n z
e
d0
1
vei
s
v
a
r
i
a
b
l
e
sc
a
nb ngt
m
i
n
i
m
i
z
t
ds
u
h
d
e
c
tt
ot
h
e
l
i
n
e
a
r
i
z
c
dc
o
n
s
t
r
a
i
n
t
sa
b
o
u
tt
h
ecumntd
e
s
i
g
nr
.

i
i
Minimize]
(
l
:
J
.
z
'l
:
J
.
Y)=f(X
f
(X
i?d
.
Z+Vyf(X?l
:
J
.
Y
)+Vz

(
7
.
5
3
)

i
S
u
b
j
e
c
tt
o
:
(
!
:
J
.
Z
l
:
J
.
Y)=H
.
(X
)+Vz
H
.
(
X
i
)
Td.Z+V
.
;
.
(
X
i
)
T
l
:
J
.
Y=0

(
7
.
5
4
)

l
ft
h
eGRGb
e
g
a
na
taf
e
a
s
i
b
l
ep
o
i
n
t
.
eB(Xi)=0
el
i
n
a
r
i
z
e
dc
o
n
s
t
r
a
i
n
t
sa
r
ea
s
o
z
e
r
o

n i
b
l
e
a
r
i
z
e
d
e
q
u
a
t
i
o
n
si
n]d
e
t
e
n
n
i
n
eas
e
to
fl
i
n
e
a
re
q
u
a
t
i
o
n
si
nl
:
J
.
Zand
d
.
Z
.C
o
n
s
i
d
e
r
i
n
g
v
e
c
t
o
r[
H
]t
h
el
i
n
e
a
re
q
u
a
t
i
o
nc
a
nbew
r
i
t
t
e
n
[
A
]d
.
Z
+

]
l
:
J
.
y=o

(
7
.
5
5
)

1v~1i!" 1

where

sc
h
o
i
c
e
.Thei
t
e
r
a
t
i
o
n
ss
t
a
r
tw
i
t
haf
e
a
s
i
b
l
ec
u
r
r
e
n
t
c
o
n
v
e
r
g
e
n
c
ea
s
s
o
c
i
a
t
e
dw
i i
X
i
]=[Z
Y
d
e
s
i
g
nv
e
c
t
o
r[

S
t
e
p
s
l
z
eComputatlon:A
rlthm(GRG)
S
t
e
p1
.F
i
n
dGR
S=-G
e
l
e
c
ta
R; S
S
t
e
p
2
. q=1

d
.
Z=S; Z=Z
i+
!
:
J
.
Z
;
I
l
:
J
.yq=[Br[
A
]d
.
Z
S
t
e
p3
.yq+l=Yq+l
:
J
.
yq

Z
I

x
1
+
1
=
1
Iyq

I
f[H(X)]=0
;S
t
o
p
C
onverged

GoToS
t
e
p3

a&

1V

"-1

[
A
]=
1.
--1
;[B]=
1

D
I
R
E
C
TMETHODSFORCONSTR
A
lN
EDO
P
T
I
M
I
Z
A
T
I
O
N 299

seqq+1
l
:
J
.Y9=r1[
H
(
X
i
+
I
)
]

r
r
HT
IV~HnI
IV
v
y
11
I
YZ
I
I
J

7
.
3

I
Vz
V
)

mI I
FromE
q
t
i
o
n(
7
.
5
5
)t
h
ec
h
a
n
g
ei
ned
e
p
c
n
d
e
n
tv
a
r
i
a
b
l
el
:
J
.
Yc
a
nb
ee
x
p
r
e
s
s
e
di
n
l
e
r
m
so
ft
h
ec
h
a
n
g
e
si
nt
h
ci
n
d
e
p
e
n
d
e
n
tv
a
r
i
a
b
l
e
sa
s

(
7
.
5
6
)

l
:
J
.
Y =[Brl[
A
]d
.
Z

S
u
b
s
t
i
t
u
t
i
n
gE
q
u
a
t
i
o
n(
7
.
5
6
)i
n
t
oE
q
u
a
t
i
o
n(
7
.
5
3
)r
e
d
u
c
e
sitanu
n
c
o
n
s
t
r
a
i
n
e
d
p
r
o
b
l
m i
nd
.
Z
:

Evea
g
o
r
i
m
n
e
i Z[
A
][
B
J
noris n
g
e
d
.Thcs
i
n
g
u
l
a
r
i
t
yo
f[
B
]
i
saconcemandc
a
nb
econ
O
I
l
e
dbyt
h
E
s
e
l
e
c
t
i
o
n
o
f
i
n
d
e
p
e
n
d
e
n
tv
a
r
i
a
b
l
e
s
.
I
t
a
p
P
E

atazz
sr
e
s
p
o
n
s
i
b
l
ef
o
rd
e
c
r
E
a
s
e
i
nt
h
Ef
u
n
c
t
i
o
nw
h
i
l
eAYi
sr
E
s
p
o
n
s
i
b
l
E
f
o
r
f
e
a
s
i
bi
1
it
y
.Thea
b
o
v
ea
l
g
o
r
i
t
h
mi
su
s
u
a
l
1yx
e
c
u
t
e
df
o
rlwova
u
e
sof.Aq
u
a
d
r
a
t
i
c
i
n
t
e
r
p
o
l
a
t
i
o
nb
e
t
w
e
e
nt
h
et
h
r
p
o
i
n
t
s(=0i
sa
v
a
i
l
a
b
J
efromt
h
ep
r
e
v
i
o
u
si
t
e
r
a
t
i
o
n
)
i
s
todetennineeoptimum psize.One
f
i
n

li

e
r

a oni
sp
e
r
f
o
r
m
e
df
o
r
optimums
t
e
p
s
i
z
e
.

A01 m
:Genel
e
!
l
l
z
e
dReducedGradient(GR

S
t
e
p1
. ChooseXI(
m
u
s
tb
ef
e
a
s
i
b
l
e
)
N.(
n
o
.o
fi
t
e
r
a
t
i
o
n
s
)

E
;
'
S(
f
o
rc
o
n
v
e
r
g
e
n
c
eands
t
o
p
p
i
n
g
)
i
]()=f(X
f
(
X
i
)
Td
.
Z-V
y
f
(
X
i
)
T
[
B
r
l
[
A
]d
.
Z
)+Vz

p=1(
i
t
e
r
a
t
i
o
nc
o
u
n
t
e
r
)

S
t
e
p2
.I
d
e
n
t
i
f
yZ
Y
)+{
=f(Xi
v
zf(X

T-Vyfl

T[Brl[A])l
:
J
.
Z

(
7
.
5
7
)

=
X+{Vzf(Xi
)([l[Alfvdxt)

=
f(XI
G
R
]
Tl
:
J
.
Z
)+[
e
d
u
c
e
dg
r
a
d
i
e
t
J
to
ft
h
ef
u
n
c
t
i
o
n
j
{
X)
I
tp
r
o
v
i
d
e
st
h
es
e
a
r
c
hd
i
r
e
c
t
i
o
n
.The
[G
]i
ser
R
izeisf
o
u
n
dbyJ1 i
r
i
n
g t ng
s
i
nl
:
J
.
Y

J
1
J
1
e
5
p
o
n
d
i
n
g1
0 c
h e
s
i
nhz(S)d
e
t
e
r
m
i
n
eal
f
s
i
b
l
ed g
n
.T
h
i
si
si
m
p
l
e
m
e
n
t
e
dr
o
u
g
hNewton's
themi
snoq
u
a
d
r
a
t
i
c
methoda
sd
e
s
c
r
i
b
e
dbelow.SincethehnctionsMEl
i
n
e
a
r

Ca
c
u
l
a
t
e[
A
]
[
B
]
Ca

c
u
J
a
t
e[GR]
C
a
l
c
u
l
a
t
eOptimumS
t
e
p
s
i
z
e

(
s
e
ea
l
g
o
r
i
t
h
m
)
C
a
l
c
u
l
a
t
eXP+l
S
t
e
p3
.C
o
n
v
e
r
g
e
n
c
ef
o
rGRG

f
o
rk=1

2
.
.
.

1
;
I
fh
k=0
I
f8
jS
;0

f
o
rj=1

2
.
.
.

m;

I
fKTc
o
n
d
i
t
i
o
n
sa
r
es
a
t
i
s
f
i
e
d
Th
e
nConverged

S
l
O
p

300 N
U
M
E
R
I
C
A
LT
E
C
H
N
I
Q
U
E
SF
O
RC
O
N
S
T
R
A
I
N
E
DO
P
T
I
M
I
Z
A
T
I
O
N

7
.
3 DlR
E
C
TMETHODSF
O
RCONST
RAlN
E
DO
P
T
I
M
I
Z
A
T
l
ON 3
0
1
A

S
t
o
p
p
i
n
gC
r
i
t
e
r
i
a
:
L
l
X Xp+1- XP
T
I
fL
l
X
L
l
X~ E
I
:S
t
o
p(
d
e
s
i
g
nn
o
tc
h
a
n
g
i
n
g
)
I
fp N
.
:S
t
o
p(maximumi
t
e
r
a
t
i
o
n
sr
e
a
c
h
e
d
)
C
o
n
t
i
n
u
e
pp+l
Got
oS
t
e
p2

2.4000e+000
6.0000e-001
B
1.4967e+000 0

1.1225e+000 1.0000e+000

-6.0183e+000

'
0

A
p
p
l
i
c
a
t
l
o
n t
h
eGRGMethod:TheGRGa
l
g
o
r
i
t
h
mi
sa
p
p
l
i
e
d1
0Ex
n
p
l
e
7
.
1
.Oncea
g
a
i
nt
h
ei
I
lu
s
t
r
a
t
i
o
ni
st
h
r
o
u
g
has
e
to
fc
a
l
c
u
l
a
t
i
o
n
sg
e
n
e
r
n
t
e
du
s
i
n
g
MA
1LA
B
.F
o
ra5
t
a
r
tExample7.
1i
sr
e
w
r
i
t
t
e
nu
s
i
n
gas
l
a
c
kv
a
r
i
a
b
l
ef
o
rt
h
i
n
e
q
u
a
l
i
t
y
t
.
c
o
n
s
t
r
a
m
M
i
n
i
m
i
z
e f
(
X
hx
2

X
3
)
: x1-2
xr
X
2+x
t+XIX~ -2xl
+4
Su
t

H
1
(
x
l
X
2

X
3
)
: xi+x~-2=O

== =

O~x2~5;

X3~O

Input the stepsize for evaluation.


(
Prompt for first stepsize)
0.1
alpha
1.0000e-001
No. of iterations and constraint error: 29
1.0000e-001 8.2187e-009

function and constraints (


f h1 h2)
3.3548e+000 7.2526e-005 5.4394e-005

s"
'

F
o
ri
sp
r
o
b
l
e
mn 2
1 1

m 1

n+m 3

"-l l

l+m 2
.
Example7
.
1i
sn
o
tagoodp
r
o
b
l
e
mt
oi
I
lu
s
t
r
a
t
et
h
ew
o
r
k
i
n
go
ft
h
eGRGm
e
t
h
o
d
b
e
c
a
u
s
e o
n
l
yf
e
a
s
i
b
l
ep
o
i
n
tw
h
e
r
egi
sa
c
t
i
v
ei
sa
l
s
oes
o
l
u
t
i
o
n
.T
h
i
smakesZ
e
f
o
r
eGR as
c
a
l . Thee
x
a
m
p
l
ei
sr
e
t
a
i
n
e
db
e
c
a
u
s
eo
fu
n
i
f
o
r
m
i
t
y
.I
n
a
n
dr
i
m
p
l
e
m
e
n
t
a
t
i
o
nb
e
l
o
w

Z=[
X
I
]a
n
dY=[
X
2X
3
)
T
.

Sec7
_3_3.m: lIsmf
i
Ie h
a
n
d
l
e
s Example 7
.
1t
h
r
o
u
g
h e GRG method
s
y
m
b
o
l
i
c
a
l
l
y
.I
tr
e
q
u
e
s
t
sas
t
a
r
t
i
n
gd
e
s
i
g
nw
h
i
c
hs
h
o
u
l
db
ef
e
a
s
i
b
l
e
.I
ts
e
t
supeZ
a
n
dYa
u
t
o
m
n
t
i
c
a
l
l
ya
n
dc
o
m
p
u
t
e
s[
A
)nnd[
8
]
.Experimentingwitht
h
ec
o
d
ei
n
d
i
c
a
t
e
s
t
h
a
tn
o
ta
l
lf
e
a
s
i
b
l
ep
o
i
n
t
sp
r
o
v
i
d
ean
o
n
s
i
n
g
u
l
a
r[
8
]
.I
tt
h
e
np
r
o
m
p
t
st
h
eu
s
e
rf
o
rt
w
o
vuofwhichi
tu
s
e
st
op
e
r
f
o
r
maq
u
a
d
r
a
t
i
ci
n
t
e
o
l
a
t
i
o
n
.I
tc
o
m
p
u
t
e
sa
noptimum
s
l
e
p
s
i
z
ea
n
dc
o
m
p
l
e
t
e
st
h
ei
t
e
r
a
t
i
o
nf
o
rd
.
Y
.Thi
sm
p
l
e
l
e
so
n
ei
t
e
r
a
t
i
o
no
f GRG.
E
x
p
e
r
i
m
e
n
t
i
n
gw
i
t
h ec
o
d
ei
I
1
u
s
t
r
a
t
e
sa
tt
h
ec
h
o
i
c
e
so
fneedsomet
h
o
u
g
h
tf
o
r
t
h
ei
t
e
r
a
t
i
o
n1
0s
u
c
c
e
e
d
.Thef
o
l
l
o
w
i
n
gi
st
r
a
n
s
c
r
i
b
e
df
r
o
mt
h
eCommandwindowf
o
r
t
w
oi
t
e
r
a
t
i
o
n
s
.N
o
t
eep
r
o
m
p
t
sf
o
ra
l
p
h
a
:

l
t
e
r
a on1
sec7_3_3
Input the feasible design chosen for evaluation.
[1.2 0.74833]
% This is x1

The design vector [1.2000 0.7483 0.2200]


function and constraints(f h1 h2):
3.6304e+000 -2.2111e-006 0

(ReducedGradient)

design vector: 5.9817e-001 1.2815e+000 -3.2110e-001

H
2
(
X
h
X
2
)
: O.25xi+o.75x ~-1 +X3=O
O~xl~5;

-6.0183e+000

Input the stepsize for evaluation.


(
Prompt for second stepsize)
0.05
alpha
5.0000e-002
No. of iterations and constraint error: 12
5.0000e-002 6.2635e-009
design vector: 8.990ge001 1.0916e+000 -9.5823e-002
function and constraints (
f h1 h2)
2.9702e+000 6.3314e-005 4.7485e-005
alpha
S56595e 002
(
quadraticallyinterpolated}

-6.0183e+000
No. of iterations and Constraint error: 14
5.6595e-002 4.3237e-009
design vector: 8.5940e001 1.1231e+000 -1.3072e-001
function and constraints (
f h1 h2)
2.9903e+000 5.2604e-005 -3.9453e-005

302

NUMERICALTECHNIQUESFORCONSTRAINEDOPTIMIZATION

7
.
3 D
lRECTMETHODSFORCONSTRAINEDOPTIMIZATION

T
h
i
smple
o
n
e
i
t
e
r
a
t
i
o
n
.T
h
i
si
sf
a
rfroms
a
t
i
s
f
a
c
t
o
r
ya
sX3
t
h
es
l
a
c
kv
a
r
i
a
b
l
e
i
s
n
e
g
a
t
i
v
e
.Thes
t
e
p
s
i
z
ea
n
dd
e
s
i
g
nc
h
a
n
g
ec
a
l
c
u
l
a
t
i
o
nmustbemad
s
e
n
s
i
t
i
v
et
oe
s
i
d
ec
o
n
s
t
r
i
n
t
s
.I
na
c
t
u
a
li
m
p
l
e
m
e
n
t
a
t
i
o
no
ft
h
eGRG
s
i
n
c
eg
(X)Wasi
n
a
c
t
i
v
ea
te
s
t
a
r
t

i
twouldn
o
lh
a
v
eb
e
e
ni
nc
1udedi
nt
h
ec
a
l
c
u
l
a
t
i
o
n
sa
ta
l
l
.U
s
i
n
gt
h
es
o
l
u
t
i
o
nt
o
s
tt
h
en
e
x
ta
d
d
i
t
i
o
n
a
li
t
e
r
a
t
i
o
n
sc
a
nb
eo
b
t
a
i
n
e
dl
e
a
d
i
n
g1
0t
h
es
o
l
u
t
i
o
n
.

303

I
n SGRAo
n
l
ya
c
t
i
v
ei
n
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
t
sa
r
eo
fi
n
t
e
r
e
st
.E
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
t
sa
r
e
1udev
i
o
l
a
t
e
dc
o
n
s
t
r
a
i
n
t
s
.I
f1
1
a
l
w
a
y
sa
c
t
i
v
e
.A
c
t
i
v
ei
n
e
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
l
sa
l
s
oi
nc
i
n
d
i
c
a
t
e
st
h
es
e
to
fa
c
t
i
v
ec
o
n
s
t
r
a
i
n
t
su
a
t
i
o
n
s(
7
.
6
) d(
7
.
7
)a
r
ecombinedi
n
t
o
av
e
c
l
o
ro
fa
c
t
i
v
ec
o
n
s
i
n
t
s(

)
:

l
t
e
r
a
t
i
o
n2

(
7
.
5
8
)

(X)=[
R
e
r
u
nSec7
_3_3.mw
i
t
hX2=[
8
.
5
9
4
0
e
0
0
11
.
1
2
3
1
e
)
0
0
]

.TheLa
g
r
a
n
g
i
a
nf
o
rt
h
ep
r
o
b
l
e
m
Thenumbero
fa
c
t
i
v
ei
n
e
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
t
si
sv-l
c
t
i
v
ec
o
n
s
i
a
l
o
n
e[
s
i
n
c
e m
u
l
t
i
p
l
i
e
r
sf
o
re
c
a
nb
ee
x
p
r
e
s
s
e
di
nl
e
r
m
so
f a
g
J
{
X
)<0w
i
l
lbes
e
tt
oz
e
r
oa
sp
a
r
to
fKTc
o
n
d
i
t
i
o
n
s
]
. l
eKTc
o
n
d
i
t
i
o
n
sa
t
h
e
n
e
x
p
r
e
s
s
e
da
s

Ther
e
s
to
fei
t
e
r
a
t
i
o
n
sa
r
el
e
f
ta
sa
ne
x
e
r
c
i
s
ef
o
r r
e
a
d
e
r
.Toc
o
m
p
l
e
t
e
s
i
n
c
e
Sec7
_3_3.mt
h
ec
a
l
c
u
l
a
t
i
o
n
sf
o
rKTc
o
n
d
i
t
i
o
n
smustbei
n
c
l
u
d
e
d
.I
nt
h
ea
b
o
v
e
t
h
eKTc
o
n
d
i
t
i
o
n
sw
e
r
en
o
tc
o
m
p
u
t
e
d
.Thi
si
sc
u
r
r
e
n
t
l
yoneo
ft
h
e
g
(
X
)i
si
nv
i
o
l
a
t
i
o
n
morer
o
b
u
s
tm
e
t
h
o
d
s
.Fort
h
i
spa
c
u
l
a
re
x
a
m
p
l
et
h
ep
r
o
g
r
e
s
so
fs
o
l
u
t
i
o
ni
sq
u
i
l
e
s
e
n
s
i
t
i
v
et
oes
t
e
p
s
i
z
ec
h
o
s
e
nf
o
ri
n
t
e
r
p
o
l
a
t
i
o
n
.

VxF(
X
) Vx
f
.
(
X
)+[V~P'"

(
7
.
5
9
a
)

7
.
3
.
4 SequentlalGradlentRestoratlonAlgorlthm(SGRA)
TheSGRAi
sat
w
o
p
h
a
s
ea
p
p
r
o
a
c
ht
oes
o
l
u
t
i
o
no
fNLPp
r
o
b
l
e
m
s
.S
t
a
r
t
i
n
gata

t
h
eG
r
a
d
i
e
n
tp
h
a
s
ed
e
c
r
e
a
s
e
st
h
ev
a
l
u
eo
ft
h
eo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
nw
h
i
l
e
f
e
a
s
i
b
l
ed
e
s
i
g
n
s
a
t
i
sI
I
n
g l
i
n
e
a
r
i
z
e
da
c
t
i
v
ec
o
n
s
t
r
a
i
n
t
s
.Th
i
sw
i
l
lc
a
u
s
ec
o
n
s
a
i
n
tv
i
o
l
a
t
i
o
nf
o
r
n
o
n
l
i
n
e
a
ra
c
t
i
v
ec
o
n
s
i
n
t
s
.The R
e
s
t
o
r
a
t
i
o
np
h
a
s
eb
r
i
n
g
sb
a
c
kt
h
ed
e
s
i
g
n1
0
i
f
f
e
r
e
n
ls
e
to
fa
c
t
i
v
ec
o
n
s
t
r
a
i
n
t
s
.Thi
sc
y
c
l
e
f
e
a
s
i
b
i
l
i
t
yw
h
i
c
hmaye
s
t
a
b
l
i
s
hanew dd
o
ftwop
h
a
s
e
si
s p
e
a
l
e
du
n
t
i
lt
h
eoptimumi
sf
o
u
n
d[
15
1
6
]
.Th
eme
o
di
n
c
O
J
p
o
r
a
t
e
sa
e
n
tp
r
o
p
e
r
t
yw
i
t
h c
hc
y
c
l
e
.L
ikeo
t
h
e
ra
l
g
o
r
i
t
h
m
si
ni
ss

t
i
o
n

i
tu
s
e
san8Ct
i
v
e
d c
n
s
t
r

i
n
ts
t
r
a
t
e
g
y
.I
nc
o
m
p
a
r
i
s
t
h
eGRGc
o
m
b
i
n
e
st
h
et
w
ophest
o
g
e
t
h
.
m
e
t
h
o
dh
a
sb
e
e
nf
a
v
o
r
a
b
l
yc
o
m
p
8
1
t
oGRGa
n
dt
h
eG
r
a
d
i
e
n
tP
r
o
j
e
c
t
i
o
nmethod[
1
7
]
.
Th
eSGRAa
so
r
i
g
i
n
a
l
l
yi
n
d
ucedinel
i
s
t
e
dr
e
f
e
r
e
n
c
e
sa
s
s
u
m
e
st
h
ei
n
e
q
u
a
l
i
t
y
c
o
n
s
a
i
n
t
so
ft
h
ef
o
r
m

E
F

where[V
X
+
1
=[V
X
+I Vx
z

(X)=O

(
7
.
5
9
b
)

(
7
.
5
9
c
)

Vx
'
!
>v
]
; =l1

.
J
T

GrBdientPhBse: Givenaf
e
a
s
i
b
l
ed
e
s
i
g
nX
'

f
i
n
dt
h
en
e
i
g
h
b
o
r
i
n
gp
o
i
n
t

s=Xi+AX
s
u
c
ht
h
a
t5
j<0 =O
.Imposingaq
u
a
d
r
a
t
i
cc
o
n
s
t
r
a
i
n
tont
h
ed
i
s
p
l
a
c
e
m
e
n
tdX
[
16
]ep
r
o
b
l
e
mc
a
nbes
e
tupa
sa
no
p
t
i
m
i
z
a
t
i
o
ns
u
b
p
r
o
b
l
e
mwhoseKTc
o
n
d
i
t
i
o
n
s
d
e
t
e
r
m
i
n
e

g
j
(
X
)
0
; j=1

2 m

Tob
ec
o
n
s
i
s
t
e
n
tw
ied
i
s
c
u
s
s
i
o
ni
nt
h
ebooki
ti
sc
h
a
n
g
e
dt
ot
h
es
t
a
n
d
a
r
df
o
r
mo
f
e
f
i
n
e
di
ni
st
e
xt
.I
nt
h
ef
o
l
l
o
w
i
n
gd
e
v
e
l
o
p
m
e
n
to
ft
h
ea
l
g
o
r
i
t
h
m
NLPp
r
o
b
l
e
m
s d
o
n
l
yt
h
es
a
l
i
e
n
tf
e
a
t
u
r
e
sa
r
ei
n
d
i
c
a
t
e
d
.Theya
r
ea
l
s
ot
r
a
n
s
c
r
i
b
e
dc
o
n
s
i
s
t
e
n
tw
i
t
he
o
t
h
e
ra
l
g
o
r
i
t
h
m
si
nt
h
i
ss
e
c
t
i
o
n
.I
ti
s commended t
h
er
e
a
d
e
rr
e
v
i
e
wt
h
e
r
e
f
e
r
e
n
c
e
sf
o
rad
e
t
a
i

e
dd
e
s
c
r
i
p
t
i
o
n
.Th
eg
e
n
e
r
a
lp
r
o
b
l
e
mi
s
M
i
n
i
m
i
z
e

X]
n
f
(
X
)
[

t
l
X=-aVF
x
(
XIA
.
=S

(
7
.
6
0
)

Th
ei
n
t
e
r
p
r
e
t
a
t
i
o
no
ft
h
es
e
a
r
c
hd
i
r
e
c
t
i
o
nS i
su
s
e
dt
oc
o
n
n
e
c
tt
oeo
t
h
e
r
sb
o
o
k
.I
ti
sn
o
ti
nt
h
eo
r
i
g
i
n
a
ld
e
v
e
l
o
p
m
e
nt
.Thes
e
a
r
c
hd
i
r
e
c
t
i
o
ni
s
a
l
g
o
r
i
t
h
m
si
ni
r
a
d
i
e
n
to
ft
h
eLa
g
r
a
n
g
i
a
n
whichi
san
o
v
e
lf
e
a
t
.T
ocomputee
o
p
p
o
s
i
t
et
oeg
l
h
eL
a
g
r
a
n
g
em
u
l
t
i
p
l
i
e
r
sh
a
v
et
ob
ec
a
l
c
u
l
a
t
e
d
.T
h
i
si
sdonebys
o
l
v
i
n
ga
L
a
g
r
a
n
g
i
a
n
in
e
a
re
q
u
a
t
i
o
n
s
s
y
s
t
e
mo
f
I

(
7
.
5
)

S
u
b
j
e
c
tt
o
: [
h
(
X
)
]
1=0

(
7
.
6
)

[
g
(
X
)
]" ~ 0

(
7
.
7
)

X10wsXSXUP

(
7
.
8
)

I
j
(
X
V~]T[V~(X')] ." =0
)+[
[V~(XIW Vx

(
7
.
61
)

Steps;zef
o
rGradlentPhBse:Th
es
t
e
p
s
i
z
e

c
a
l
c
u
l
a
t
i
o
ni
sb
a
s
e
dond
r
i
v
i
n
gE
o
p
t
i
m
a
l
i
t
yc
o
n
d
i
t
i
o
n
si
nE
qu
a
t
i
o
n(
7
.
5
9
a
)t
oz
e
r
o
.Th
e
r
e
f
o
i
f

304

7
.:
3 DIRECTMETHODSFORCONSTRAINEDOPT
l
MIZATION

NUMERICALTECHNIQUESFORCONSTRAINEDOPTIMIZATION

=Xi

TheR
e
s
t
o
r
a
t
i
o
np
h
a
s
ei
si
t
e
r
a
t
i
v
e
l
ya
p
p
l
i
e
du
n
t
i
l

(Xii

.
v
)

-(J.
V

l
C
T
(

)S 81

t
h
c
no
p
t
i
m
u
m

i
sf
o
u
n
db
yc
u
b
i
ci
n
t
e
r
p
o
l
a
t
i
o
nw
h
i
l
et
r
y
i
n
g
s
a
t
i
s
f
y

VxF(
'
)
.
.
l
'lVxF

i
.
.
I
')=o

(
7
.
6
3
)

R
e
f
e
r
e
n
c
e1
6s
u
g
g
e
s
t
s tPm
i
sr
e
l
a
t
e
dt
oa
n
o
e
rp
e
r
f
o
r
m
a
n
c
ei
n
d
e
xQ
w
h
i
c
hi
s
ee
r
r
o
ri
nt
h
eo
p
t
i
m
a
l
i
t
yc
o
n
d
i
t
i
o
n
s
:
Q=VFxCXi I')TVFxCXi 1')

(
7
.
6
4
)

R
e
s
t
o
r
a
t
i
o
nPhase:I
ti
se
x
p
e
c
dt
h
a
t
a
tn
do
ft
h
eg
r
a
d
i
e
n
tp
h
a
s
et
h
ef
u
n
c
t
i
o
n
o
u
l
db
esomec
o
n
s
t
r
a
i
n
td
i
s
s
a
t
i
s
f
a
c
t
i
o
n(
a
s
s
u
r
n
i
n
g
w
i
l
lh
a
v
ed
e
c
r
e
a
s
e
db
u
t w
t
h
e
r
ew
a
sa
tl
e
a
s
to
n
en
o
n
l
i
n
e
a
ra
c
t
i
v
ec
o
n
s
t
r
a
i
n
ta
tt
h
eb
e
g
i
n
n
i
n
go
ft
h
eg
r
a
d
i
e
n
t
p
h
a
s
e
)
.Ther
e
s
t
o
r
a
t
i
o
np
h
a
s
ee
s
t
a
b
l
i
s
h
e
san
e
i
g
h
b
o
r
i
n
gf
e
a
s
i
b
l
es
o
l
u
t
i
o
n
.1
td
o
e
st
h
i
s
In
e
a
r
i
z
e
dc
o
n
s
t
r
a
i
n
t
sa
r
ef
e
a
s
i
b
l
e
.P
r
i
o
rt
oi
s
.ea
c
t
i
v
e
b
ye
n
s
u
r
i
n
gt
h
a
tt
h
ei
c
o
n
s
t
r
a
i
n
ts
e
th
a
st
ob
eu
p
d
a
t
e
d(
i
i
)s
i
n
c
ep
r
e
v
i
o
u
s
l
yf
e
a
s
i
b
l
ec
o
n
s
t
r
a
i
n
t
sc
o
u
l
dh
a
v
e
b
e
C
9
m
ei
n
f
e
a
s
i
b
l
e(
a
n
dp
r
e
v
i
o
u
s
l
yi
n
f
e
a
s
i
b
l
ec
o
n
s
t
r
a
i
n
t
sc
o
u
l
dh
a
v
eb
e
c
o
m
ef
e
a
s
i
b
l
e
)
e
s
i
g
nc
h
a
n
g
ec
a
u
s
e
db
yeg
r
a
d
i
e
n
tp
h
a
s
e
.Thed
e
s
i
g
nv
e
c

ra
n
dt
h
ed
e
s
i
g
n
w
i
t
h d
c
h
a
n
g
e
sf
o
ri
sp
h
a
s
ec
a
nb
ew
r
i
t
t
e
na
s
X
=Xg+d
X

(
7
.
6
5
)

X
i
so
b
t
a
i
n
e
da
saI
s
ts
q
u
a
r
e
e
.

'
o
ri
nL
h
ed
e
s
i
g
n
Thed
e
s
i
g
nc
h
a
n
g
ef
o
rt
h
i
sp
h
a
s
ed
c
h
a
n
g
e
ss
u
b
j
e
c
tt
oes
a
f
a
c
t
i
o
no
ft
h
el
i
n
rc
o
n
s
a
i
n
t
s
.S
e
t
t
i
n
gu
p NLP
e
s
i
g
nc
h
a
n
g
e
si
s
s
u
b
p
r
o
b
l
e
mt
h
i
sc
a
l
c
u
l
a
t
i
o
nf
o
red
dXr=-V
x
C
T
(X

g)

(
7
.
6
6
)

H
e
r
e
"
i
st
h
e

v
e
c
t
o
rL
a
g
r
a
n
g
em
u
l
t
i
p
l
i
e
ro
ft
h
eq
u
a
d
r
a
t
i
cs
u
b
p
r
o
b
l
e
m
.Thev
a
l
u
e
s
f
o
rt
h
em
u
1
t
i
p
l
i
e
r
sa
r
ee
s
b
l
i
s
h
e
dt
h
r
o
u
g
ht
h
el
i
n
e
a
re
q
u
a
t
i
o
n
s
(g)-VAds)TVJdE)=0

(
7
.
6
7
)

T
h
e
f
a
c
t
o
r
i
sau
s
e
r
c
o
n
t
r
o
l
l
e
dp
'
a
m
e
t
e
rt
od
i
s
c
o
u
r
a
g
el
a
r
g
ed
e
s
i
g
nc
h
a
n
g
e
s
.

(
7
.
6
8
)

w
h
e
r
eEIi
sas
m
a
l
ln
u
m
b
e
r
.
o
n
s
t
r
a
i
n
t
sa
r
ef
e
a
s
i
b
l
ea
n
dt
h
en
e
x
t
Att
h
ec
o
n
c
l
u
s
i
o
no
ft
h
er
e
s
t
o
r
a
t
i
o
np
h
a
s
eec
e
c
b
ea
p
p
l
i
e
da
g
a
i
n
.Asummaryo
ft
h
ea
b
o
v
e
c
y
c
l
eo
fG
r
a
d
i
e
n
t
R
e
s
t
o
r
a
t
i
o
np
h
l
g
o
r
i
t
h
mg
i
v
e
nb
e
l
o
w
.
s
t
e
p
si
sc
a
p
t
u
r
e
di
nea

(
7
.
6
2
)

C
a
r
em
u
s
tb
et
a
k
e
n t
h
i
ss
t
e
p
s
i
z
ed
o
e
sn
o
tc
a
u
s
es
i
g
n
i
f
i
c
a
n
tc
o
n
s
t
r
a
i
n
tv
i
o
l
a
t
i
o
n
.
T
h
i
si
sn
f
o
l
"
c
e
db
yc
a
p
p
i
n
gt
h
es
q
u
a
r
e
de
r
r
o
ri
nt
h
ec
o
n
s
t
r
a
i
n
t
sb
yas
u
i
t
a
b
l
eu
p
p
e
r
b
o
u
n
dw
J
u
c
hi
ss
e
tu
p

(
g)T()SPm

3
0
5

Aorithm:SequentialGradientRestorationA
'
r
i
t
h
m(SGRA)(
A
7
.
S
t
e
p]
.C
h
o
o
s
eXJ(
m
u
s
tb
ef
e
a
s
i
b
l
e
)
N.(
n
o
.o
fi
t
e
r
a
t
i
o
n
s
)
.
8
j
'
S(
f
o
rc
o
n
v
e
r
g
e
n
c
ea
n
ds
t
o
p
p
i
n
g
)
p=
1(
i
t
e
r
a
t
i
o
nc
o
u
n
t
e
r
)
S
t
e
p2
.E
x
e
c
u
t
eG
r
a
d
i
e
n
tP
h
a
s
e
o
l
a
t
i
o
n
C
a
l
c
u
l
a
t
eS
t
e
p
s
i
z
eu
s
i
n
gc
u
b
i
ci
n
t
e
l
C
a
l
c
u
l
a
t
e
S
t
e
p3
.E
x
e
c
u
t
eR
e
s
t
o
r
a
t
i
o
nP
h
a
s
e

x
r

x
P
+J

S
t
e
p4
.C
o
n
v
e
r
g
e
n
c
ef
o
rSGRA
I
fhk=
O
.f
o
rk=
1
.2
..
.
.
1
;
I
fg
jS0

f
o
rj=
1
.2
..
.
.
/
1
1
;
I
fKTc
o
n
d
i
t
i
o
n
sa
r
es
a
t
i
s
f
i
e
d
ThenC
o
n
v
e
r
g
e
d

S
t
o
p
S
t
o
p
p
i
n
gC
r
i
t
e
r
i
a
:
!
'
aX=XJ*I- X
Tx
IfaX
t
o
p(
d
s
i
g
nn
o
tc
h
a
n
g
i
n
g
)
SEI:S
N
.
:S
t
o
p(maximumi
t
e
r
a
t
i
o
n
sr
e
a
c
h
e
d
)
I
fp=
C
o
n
t
i
n
u
e
pp+l
Got
o
S
t
e
p
2
A
p
p
l
l
c
a
t
i
o
noftheSGRA7
.
:Example7
.
1i
su
s
e
dt
ot
e
s
tt
h
ea
p
p
l
i
c
a
t
i
o
no
f
t
h
eSGRA.Themf
i
l
eSec7
_3_
4.mc
o
n
t
a
i
n
st
h
ec
o
d
et
h
a
ta
p
p
l
i
e
st
h
eSGRAt
h
r
o
u
g
h
r
e
v
i
o
u
se
x n
p
l
e
si
t
ac
o
m
b
i
n
a
t
i
o
no
fs
y
m
b
o
l
i
ca
n
dn
u
m
e
r
i
cc
a
l
c
u
l
a
t
i
o
n
s
.Asi
nep
a
1
i
t
ya
n
do
n
ei
n
e
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
nt
.U
n
l
i
k
et
h
e
i
sc
u
s
t
o
m
i
z
e
df
o
rt
w
ov
a
r
i
a
b
l
e
s
.o
n
ee
q
u
GRGi
m
p
l
e
m
e
n
t
a
t
i
o
ni
ti
n
c
o
r
p
o
r
a
t
e
sea
c
t
i
v
ec
o
n
s
t
r
a
i
n
ts
t
r
a
t
e
g
y
.I
te
x
e
c
u
t
e
so
n
e
r
a
d
i
e
n
tp
h
a
s
ea
n
ds
e
v
e
r
a
l
c
y
c
l
eo
ft
h
eSGRAw
h
i
c
hh
a
sas
i
n
g
l
ei
t
e
r
a
t
i
o
ni
n G
e
s
t
o
r
a
t
i
o
np
h
a
s
e
.I
nt
h
eg
r
a
d
i
e
n
tp
h
a
s
et
h
ec
u
b
i
ci
n
t
e
r
p
o
l
a
t
i
o
no
ft
h
e
i
t
e
r
a
t
i
o
n
si
neR
l
a
c
e
db
yt
h
eq
u
a
d
r
a
t
i
ci
n
t
e
r
p
o
l
a
t
i
o
nu
s
e
de
v
e
r
y
w
h
e
r
ei
nt
h
e
o
r
i
g
i
n
a
lt
e
c
h
n
i
q
u
ei
s p
r
a
d
i
e
n
tp
h
a
s
et
h
em
e
t
h
o
d
b
o
o
k
.Thes
t
a
r
t
i
n
gi
n
p
u
ti
saf
e
a
s
i
b
l
ed
e
s
i
g
nv
e
c
t
o
r
.I
neg

3
0
6 N
U
M
E
R
I
C
A
LT
E
C
H
N
I
Q
U
E
SF
O
RC
O
N
S
T
R
A
I
N
E
DO
P
T
I
M
I
Z
A
T
I
O
N
av
a
l
u
ef
o
rt
h
es
t
e
p
s
i
z
et
os
c
a
nf
o
rt
h
eoptimum u
s
i
n
gq
u
a
d
r
a
t
i
c
r
e
q
u
e
s
o
l
a
t
i
o
n
.T
h
i
si
si
m
p
o
r
t
a
n
tf
o
rs
u
c
c
e
s
s
i
v
ec
y
c
l
e
ss
t
a
r
ta
tt
h
eoptimumv
a
l
u
eo
ft
h
e
i
n
t
.T
hev
a
l
u
eofinE
q
u
a
t
i
o
n(
7
.
6
7
)i
ss
e
ta
t0
.
5
.P
r
i
n
t
i
n
g
s
t
e
p
s
i
z
ef
o
rt
h
ep
r
e
v
i
o
u
sc
y
c
l
i
sc
o
n
t
r
ol
Ie
d r
o
u
g
ht
h
es
e
m
i
c
o
l
o
na
n
d er
e
a
d
e
ri
se
n
c
o
u
r
a
g
e
d

f
o
l
l
o
wd
e
t
a
i
l
sb
y
f
l
ovmg1
1
.
I
nt
h
ef
o
l
Io
w
i
n
gt
h
ep
r
o
g
r
a
mi
sr
u
nw
i
es
t
a
r
t
i
n
gd
e
s
i
g
nu
s
e
df
o
rt
h
eGRG
m
e
t
h
o
d
.1
o
u
t
p
u
tf
r
o
mt
h
eCommandwindowi
sp
a
s
t
e
db
e
l
o
ww
i
t
hsomeb
o
l
d
a
n
n
o
l
a
t
i
o
n
s1
0d
r
a
wa
u
e
n
t
i
o
n
.

>
>

7
.
4

Sec7 3 4

npu
he feasible design chosen for evaluation.

[1.2 0
.74833]

(feasible design vector input)

The design vector [1.2000 0.7483]


function and constraints(f h g):
3.6304e+000 -2.2111e-006 -2.2000e-001

*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
*
Gradient Phase

The Lagrange multipliers


-1.0812e+000
Input the stepsize for evaluation.
1
(alpha used to start scanning process)
alpha for quadratic interpolation
o 1.2500e-001 2.5000e001
(alpha that brackets the minimum)
function values for quadratic interpolation
1.0142e+001 6.6291e+000 1.1300e+001
(corresponding function value)
optimum alpha: 1.1615e-001
The design vector [1.0043 1.0622]
function and constraints(f h1 h2):
3.0077e+000 1.3681e-001 9.8328e-002
(both constraints violated)
The performance indices QP
6.3181e+000 1.8718e-002
(
P measures the squared error in the constraints)

**********************

3.0000e+000 6.712ge-005 4.6741e-005


(Constraints satisfied - 9 active)
The performance index P
6.6911e-009
t
e
r
a
t
i
o
n
.TheSGRAi
so
ft
h
esamec
a
l
i
b
e
ra
st
h

Thes
o
l
u
t
i
o
nh
a
sc
o
n
v
e
r
g
e
d onei
GRG(
s
o
m
ec
o
m
p
a
r
i
s
o
n
sshowi
tt
ob
eb
e
t
t
e
r
)
.I
ti
sd
i
f
f
i
c
u
l
tt
oc
o
m
p
a
r
ed
i
f
f
e
r
e
n
t
a
l
g
o
r
i
t
h
m
sb
e
c
a
u
s
eo
fd
i
f
f
e
r
e
n
lc
o
d
i
n
gs
t
r
u
c
l
u
r
e
sa
n
dn
o
n
s
t
a
n
d
a
r
di
m
p
l
e
m
e
n
t
a
t
i
o
n
d
e
t
a
i
l
s
.N
e
v
e
r
t
h
e
l
e
s
st
h
eSGRAh
a
sc
o
n
s
i
d
e
r
a
b
l
em
e
r
it
.I
ti
sd
e
f
i
n
i
t
e
l
yb
e
t
t
e
rt
h
a
nt
h
e
GRGa
sf
a
ra
sExa
m
p
l
e7
.
1i
sconcemed(
S
e
c
t
i
o
n7
.
3
.
3
)
.T
h
i
sc
o
m
p
a
r
i
s
o
nmayn
o
tb
e
f
a
i
rs~nce t
h
ea
c
t
i
v
ec
o
n
s
t
r
a
i
n
ts
t
r
a
t
e
g
ywasn
o
tu
s
e
df
o
rt
h
eGRGe
v
e
nt
h
o
u
g
hi
ti
sp
a
r
t
o
ft
h
eGRG.T
h
i
sd
e
m
o
n
s
t
r
a
t
e
st
h
en
e
e
df
o
rs
u
c
has
t
r
a
t
e
g
yi
na
l
la
l
g
o
r
i
t
h
m
s a
t
d
e
a
l
c
o
n
s
i
n
e
do
p
t
i
m
i
z
a
t
i
o
n
.
wi

7
.
4 ADDITIONALEXAMPLES

7
.
4
.
1 Example7
.
2
agpoleProblem

Restoration Phase
*********************

The design vector [1.0000 1.0000]

function and constraints(f h1 h2):

Thr
e
ea
d
d
i
t
i
o
n
a
le
x
a
m
p
l
e
sa
r
ei
n
t
r
o
d
u
c
e
dh
e
r
e1
0i
l
Iu
s
t
r
a
t
ea
n
de
x
p
a
n
dt
h
e eofs
e
ewasas
i
g
n
i
f
i
c
a
n
t
t
e
c
h
n
i
q
u
e
s
.D
u
r
i
n
gt
h
ee
a
r
l
yd
e
v
e
l
o
p
m
e
n
to
ft
h
e
s
et
e
c
h
n
i
q
u
r
e
m
p
h
a
s
i
sone
f
f
i
c
i
e
n
c
y
r
o
b
u
s
t
n
e
s
s

a
n
dk
e
e
p
i
n
gc
o
m
p
u
t
a
t
i
o
n
a
lr
e
s
o
u
r
c
e
sa
sl
o
wa
s
p
o
s
s
i
b
l
e
.Mosto
fet
e
c
h
n
i
q
u
e
sw
e
r
ed
e
v
e
l
o
p
e
do
nm
a
i
n
f
r
a
m
e
su
s
i
n
gF
O
R
T
R
A
N
.
a
ta
r
el
a
r
g
e
l
yh
e
u
r
i
s
t
i
c
)
c
o
u
p
l
e
dw
i
t
hs
i
m
p
l
e
TodayI
h
es
e
a
r
c
hf
o
rg
l
o
b
a
ls
o
l
u
l
i
o
n
s(
c
a
l
c
u
l
a
t
i
o
n
sr
e
p
l
e
da
l
m
o
s
te
n
d
l
e
s
s
l
yp
e
r
f
o
r
m
e
dondktopPCsw
i i
n
c
r
e
d
i
b
l
e
c
o
m
p
u
t
i
n
gpower

I
h
r
o
u
g
hs
o
f
t
w
r
u
s
y
s
t
e
m
s a
lp
r
o
v
i
d
eaw
i
d
er
a
n
g
eo
fr
e
s
o
u
r
c
e

h
a
ss
h
i
f
t
e
dt
h
ef
o
c
u
st
oe
x
p
e
r
i
m
e
n
t
a
t
i
o
na
n
dc
r
e
a
t
i
v
i
t
y
.Thea
u
t
h
o
rf
e
e
l
ss
u
c
ha
ne
f
f
o
r
t
m
u
s
tb
ec
o
u
p
l
e
dw
i
t
hk
n
o
w
l
e
d
g
ea
n
di
n
s
i
g
h
tg
a
i
n
e
dt
h
r
o
u
g
haw
i
d
es
e
l
e
c
t
i
o
no
f
n
d
e
r
s
t
a
n
d
i
n
gep
r
o
g
r
e
s
so
fn
u
m
e
r
i
c
a
lc
a
l
c
u
l
a
t
i
o
n
.I
ns
p
i
t
eo
fa
l
l
a
p
p
l
i
c
a
t
i
o
n
s du
ee
x
p
o
s
es
of
a
ri
nt
h
e f
o
rmanyr
e
a
ld
e
s
i
g
np
r
o
b
l
e
m
st
h
es
o
l
u
t
i
o
ni
sn
o
t
i
n
gw
i
t
hn
o
n
l
i
n
e
a
re
q
u
a
t
i
o
n
si
sac
h
a
l
l
e
n
g
ea
n
do
f
t
e
ns
u
r
p
r
i
s
i
n
g
.
a
u
t
o
m
a
t
i
c
.D l
T
r
y
i
n
g1
0u
n
d
e
r
s
t
a
n
dwhyt
h
el
e
c
h
n
i
q
u
ed
o
e
sn
o
tworki
sa
smuchal
e
a
r
n
i
n
g
e
x
p
e
l
"
ncear
o
u
t
i
n
es
o
l
u
t
i
o
nt
ot
h
ep
r
o
b
l
e
m
.S
i
n
c
el
h
e
s
ee
x
a
m
p
l
e
su
s
et
h
em
e
l
h
o
d
s
o
fep
v
i
o
u
ss
e
c
t
i
o
nw
i
t
hd
i
f
f
e
r
n
tf
u
n
c
t
i
o
n
sa
l
lo
ft
h
ec
o
d
ed
e
v
e
l
o
p
e
db
e
f
o
r
ei
s
r
e
v
i
s
e
dw
i
t
ha
n
dt
h
ea
p
p
r
o
p
r
i
a
t
ec
h
a
n
g
e
sa
n
dr
e
n
a
m
e
d
.s
em
o
d
i
f
i
c
a
t
i
o
n
s q
u
i
r
e
i
sa
l
l
o
w
st
h
eu
s
e
rt
omake
t
h
a
tf
e
wl
i
n
e
sbec
h
a
n
g
e
df
r
o
mt
h
eo
r
i
g
i
n
a
lc
o
d
e
.l
i
n
c
r
e
m
e
n
t
a
lc
h
a
n
g
e
sa
sn
e
e
d
e
d
.T
h
i
si
sn
o
ta
ne
f
f
i
c
i
e
n
ta
p
p
r
o
a
c
hb
U
lc
a
p
i
t
a
l
i
z
e
sont
h
e
u
n
l
i
m
i
t
e
dh
a
r
dd
r
i
v
es
p
a
c
ea
v
a
i
l
a
b
l
et
o
d
a
y
.S
t
u
d
e
n
t
sa
r
eu
r
g
e
dt
ou
n
d
e
r
s
t
a
n
dt
h
e
a
l
g
o
r
i
t
h
m
t
h
ef
l
o
wo
ft
h
ec
o
d
e
a
n
der
e
s
u
l
tb
e
i
n
gs
o
u
g
h.
t

*********************

Number of Restoration iterations: 21

A
D
D
I
T
I
O
N
A
LE
X
A
M
P
L
E
S 307

T
h
i
si
st
h
esamep
r
o
b
l
e
mw
h
i
c
hwass
o
l
v
e
dg
r
a
p
h
i
c
al
Iyi
nC
h
a
p
t
e
r2a
n
a
l
y
t
i
c
a
l
l
y
i
nC
h
a
p
t
e
r4
a
n
dnowc
o
n
s
i
d
e
r
e
df
o
rn
u
m
e
r
i
c
a
ls
o
l
u
t
i
o
n
.I
ni
ss
u
b
s
e
c
l
i
o
nt
h
e
s
c
a
l
e
dv
e
r
s
i
o
no
fI
h
ep
r
o
b
l
e
mi
ss
o
l
v
e
d
.FromC
h
a
p
t
e
r4Example4
.
3i
ti
s
r
e
i
n
t
r
o
d
u
c
e
da
s

7
.
4A
O
O
l
n
O
N
A
LE
X
A
M
P
L
E
S 309

308 N
U
M
E
R
I
C
A
LT
E
C
H
N
I
Q
U
E
SF
O
RC
O
N
S
T
R
A
I
N
E
DOP
T
lM
I
Z
A
T
I
O
N
Example2

M
i
n
i
m
i
z
e f= xT-~

(
7
.
6
9
)

g
l
:2.
5
5+
O.
l3
6
l
x1- 33.
148(x1-x~) ~ 0

(
7
.
7
0

g
2
:(
0
.
8
8+1.4619x 1 )( +x+~ 166.5614(x~ -x~) ~ 0

(
7
.
7
0
b
)

1.0
86
8x1+0
.
3
482-28-~) ~O

(
7
.
7
O
c
)

g
4
:x
0
.
0
0
1~O
2-x
1+

(
7
.
7
0
d
)

0
.
0
2~ x1~ 1
.0
;

0
.
0
2~ x2~ 1
.0

2 3 ...]

[0.75 0.55]

sm-f
i
Ieu
s
e
st
h
eALMm
e
t
h
o
d1
0s
o
l
v
eep
r
o
b
l
e
m
.l
no
r
d
e
r
ALM_7
_4_1.m: Thi
t
op
r
e
s
e
r
v
et
h
eo
l
dALM.mt
o f
l
e
c
tt
h
ed
i
s
c
u
s
s
i
o
ni
nS
e
c
t
i
o
n7
.
2
.
2ac
o
p
yo
ft
h
ef
i
l
e
i
sm
a
d
e
.Ofun341.m

Hfun341.m
Gfun341.m
a
n
dFALM341.mr
e
f
l
e
c
te
f
u
n
c
t
i
o
n
sf
r
o
mExample7
.
2
.A1s
oALM_7_4_1.mi
sa
d
j
u
s
t
e
davoidh
a
n
d
l
i
n
g
e
q
u
a
l
i
t
yc
o
n
s
a
i
n
t
s dp
r
o
v
i
d
esomep
r
i
n
t
i
n
gs
u
p
p
o
r
t
.0e
r
w
i
s
ei
ti
se
s
s
e
n
t
i
a
l
l
yt
h
e
t
e
db
e
l
o
wa
n
d
s
a
m
e
.Oncea
g
a
i
nt
h
eo
u
t
p
u
tf
r
o
mt
h
eCommandwindowi
sp
a
n
n
o
t
a
t
e
dt
od
r
a
wa
t
t
e
n
t
i
o
nt
osomei
s
s
u
e
s
.O
n
l
y i
n
i
t
i
a
lp
r
o
m
p
t
sw
i t
h
ei
n
p
u
t
s
t
h
ei
n
i
t
i
a
li
t
e
r
a
t
i
o
n
a
n
dt
h
ec
o
n
s
o
l
i
d
a
t
e
di
n
f
o
r
m
a
t
i
o
na
f
t
e
rt
h
ef
i
n
a
li
t
e
m
t
i
o
n c
o
p
i
e
d
b
e
l
o
w
.
ALM_7_4_1
(invoke program)
Input the starting design vector
This is mandatory as there is no default vector setup

(initial designvector)

The initial design vector:


7.5000e-001 5.5000e-001
Input the minimum values for the designvector.
These are input as a vector. The defa~lt values are
3 *start value below the start value unless it is
zero.
In that case it is -5:
[0.02 0.02] (input lower bound not currently used)

(
7
.
7
1
)

Ins
htaboutNumericalS
o
l
u
t
i
o
n
:I
nC
h
a
p
t
e
r4i
twasi
n
v
e
s
t
i
g
a
t
e
d e
s
o
l
u
t
i
o
ni
sn
o
tar
e
g
u
l
a
rp
o
i
n
t
.
s
o
I
u
t
i
o
nd
o
e
sn
o
ts
a
t
i
s
f
yt
h
eKTc
o
n
d
i
t
i
o
n
sb
e
c
a
u
s
et
h
I
tw
asa
s
c
e
r
t
a
i
n
e
d teg
r
a
d
i
e
n
t
so
fet
w
oa
c
t
i
v
ec
o
n
s
t
r
a
i
n
t
s td
e
t
e
r
m
i
n
ee
s
o
l
u
t
i
o
n(
8
1
0g3)a
r
en
e
a
r
l
yp
a
r
a
l
l
e
li
nt
h
en
e
i
g
h
b
o
r
h
o
o
do
f
t
h
es
o
l
u
t
i
o
n
.Howw
i
l
lt
h
i
s
a
f
f
e
c
tt
h
en
u
m
e
r
i
c
a
ls
e
a
r
c
hf
o
rt
h
es
o
I
u
t
i
o
n
?
T
y
p
i
ca
Jl
yw
h
a
tan
u
m
e
r
i
c
a
lt
e
c
h
n
i
q
u
ed
o
e
si
sp
i
c
kau
s
e
f
u
ld
i
r
e
c
t
i
o
na
n
ds
e
ehow
b
e
s
ti
tc
a
nmovea
l
o
n
gi
t
.S
u
p
p
o
s
et
h
et
e
c
h
n
i
q
u
ed
e
c
i
d
e
st
omovea
l
o
n
gt
h
eg
r
a
d
i
e
n
t
o
f8
3t
od
e
v
e
l
o
pi
ti
n
t
oa
na
c
t
i
v
ec
o
n
s
t
r
a
i
n
.
tT
h
c
r
ei
so
b
v
i
o
u
s
l
yg
o
i
n
gt
ob
eap
r
o
b
l
e
m
obecomea
c
t
i
v
e
.P
u
s
h
i
n
gac
o
n
s
t
r
a
i
n
tt
ob
ea
c
t
i
v
ei
sab
a
s
i
cwayt
o
i
np
u
s
h
i
n
gglt
o
s
t
r
e
d
u
c
et
h
eo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
ni
nw
e
l
l
f
o
r
m
u
l
a
t
e
dNLPp
r
o
b
l
e
m
.T
h
e
r
e
f
o
r
em
a
l
g
o
r
i
t
h
m
s(
u
n
l
e
s
se
ya
r
et
r
a
i
n
e
dt
oc
o
m
p
r
o
m
i
s
ei
nsomes
i
g
n
i
f
i
c
a
n
tw
a
y
)wi
l
I
u
n
d
e
r
a
c
h
i
e
v
eb
yo
n
l
yp
u
s
h
i
n
go
n
eo
f
t
h
ec
o
n
s
t
r
a
i
n
t
st
ob
ea
c
t
i
v
ea
n
de
n
s
u
r
i
n
gt
h
eo
t
h
e
r
i
sf
e
a
s
i
b
l
e
.

The length of your vector indicates the number of


unknowns (
n
)
Please enter it now and hit return for example [1

The minimum values for the design variables are


2.0000e-002 2.0000e-002

:~.Ui

Input the maximum values for the design vector.


These are input as a vector. The defau1t va1ues are
3 *start value above the start value.
f start value i
s 0 then it is +5:
[
11
]
(input upper bound -currently not used)
The maximum values for the design variables are

11

Number of equality constraints [


0
]
(default value of 0 is ok)
Number of inequality constraints [
0
]
4
Initial values for beta
[
1111
]
(input number of inequality constraints and initial
beta)
ALM iteration number: 0
Design Vector (X)
7.5000e-001 5.5000e001
Objective function 2.6000e-001
(In the following violated constraints are shown in
bold)
Inequality constraints -5.9180e+000 -3.5997e+001
5.2383e+000 -1.9900e-001
Square Error in constraints(hg
)
00
Lagrange Multipliers (lambda beta): 0 1 1 1 1
Penalty Multipliers (rh rg): 1 1

310 N
U
M
E
R
I
C
A
LT
E
C
H
N
I
Q
U
E
SF
O
RC
O
N
S
T
R
A
I
N
E
DO
P
T
I
M
I
Z
A
T
I
O
N
ALM iteration number: 4
Design Vector (X)
6.4726e001 6.1013e-OOl
Objective function: 4.6695e-002
)
0 2.1366e009
Error in constraints (
hg
Lagrange Multipliers (lambda beta): 0 0 4.8227e002

Penalty Multipliers (rh rg): 1000 1000

The values for x and f and 9 are


7.5000e-001 5.5000e-001 2.6000e-001 0
6.5647e-00l 6.0173e-001 6.8871e-002 0
6.4726e-001 6.1017e-001 4.6638e-002 1.5525e-006
6.4726e-001 6.1013e001 4.6695e-002 2.1366e-009
6.4726e-001 6.1013e-001 4.6695e-002 2.1366e009
(Note design and function are not changing hence
stopped)
The values for beta are
1.0000e+000 0 0 0
1.0000e+000 0 0 0
1.0000e+000 1.4062e-002 3.B9B3e-002 4.8227e-002
1.0000e+000 8.9252e001 1.7073e-001 0
(Each column represents values of beta for an
2 and 4 are zero and
iteration -beta for constraints 1
it can be seen below they are feasib1e but inactive.
Only the third constraint is active)
Inequality constraints 6.7858e-002 -4.9734e+OOO
4.6223e-005 -3.6136e-002
(Note third constraint is active. The first one is
almost activel
Thea
n
a
l
y
t
i
c
a
ls
o
l
u
t
i
o
nf
o
ri
sp
r
o
b
l
e
mi
s

X
;
=
O
7
7
4
; =0
.
6
4
4
5
; I=4.
3577E-0
2
;
g
;
O
.
O
; g;=-4.
9
8
3
7
; g
;
=
O
.
O
; g;=-O
.
3
1
9 -02
I
nc
o
m
p
a
r
i
s
o
n
t
h
eALMh
a
sd
o
n
ead
e
c
e
n
1
j
o
bo
fg
e
u
i
n
gc
Jo
s
e1
0t
h
es
o
l
u
t
i
o
nb
e
f
o

s
t
a
I
l
in
g
.Someo
fi
sc
a
nb
et
r
a
c
e
d1
0 d
i
s
c
u
s
s
i
o
np
r
i
o
r1
0r
u
n
n
i
n
gep
r
o
g
r
a
m
.

7
.
4A
D
D
I
T
I
O
N
A
lE
X
A
M
P
L
E
S 3
1
1

S
e
c
t
i
o
n7.
3.
2
. s
e
c
t
i
o
nc
o
m
b
i
n
e
st
h
es
y
m
b
o
l
i
cc
a
p
a
b
i
l
i
t
yofMA
t
h
e
r
e
s
o
u

a
v
a
i
l
a
b
l
ei
nt
h
eO
p
t
i
m
i
z
a
t
i
o
nT
o
o
l
b
o
xt
oi
m
p
l
e
m
e
n
ta
no
r
i
g
i
na
1i
n
t
e
r
p
r
e
t
a
t
i
o
no
ft
h
eSQP.
e
x
a
m
p
l
ei
sed
e
s
i
g
no
fa
nI
b
e
a
mf
o
ru
s
ei
nap
a
r
c
a
r
s
c
t
u
r
a
1p
r
o
b
l
e
m
. i
s
a d
i
t
i
o
n
a
1 m
h
a
n
i
c
a
V
c
i
v
i
le
n
g
i
n
e
e
r
i
n
gs
t
r
u
c
r
a
1o
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
e
m
.I
tw
as
i
n
t
r
o
d
u
c
e
di
nC
h
a
p
t
e
r1f
o
ri
l
Iu
s
t
i
o
n
.I
ta
1
s
oa
p
p
e
a
r
si
nCh
a
p
t
e
r
s8 dI
Of
o
ri
I
lu
s
t
i
o
n
o
fa
d
d
i
t
i
o
n
a
1i
s
s
u
e
s
.I
n
ee
x
a
m
p
l
eh
e
r
e

e
I
b
mw
i
I
1
r
r
yaw
e
i
g
h
ta
t R
r
.l
e
n
c
l
o
a
d
g de
d
e
s
i
v
a
r
i
a
b
l
e
sf
o
rep
r
o
b
l
e
ma
r
ed
e
f
u
:
t
e
dFig
u
r
e8
.
7(
C
h
a
p

8
)
.l
ef
i
g
u
r
ei
sn
o
tr
e
p
r
o
d
u
c
e
dh
e
b
u
tXI
i
st
h
ed
e
p
t
ho
ft
h
eb
e
a
m
X2i
st
h
ef
l
a
n
g
ew
i
d
t
h
X3i
s
t
h
e
w
i
c
k
n
. dX4ise
f
1a
n
g
et
J
uc
k
n

s
.O
t
h
e
rp
a
r
a
m
e
t
e
r
s(
g
W.E

p
y
)
a
r
e
n
o
t
r
e
f
i
n
e
d
h
e
.
l
em
a
t
h
e
m
a
t
i
ca
1m
o
d
e
li
s
M
i
n
i
r
n
i
z
eJ
(
X
)=p
L
A
c
3
-1
l
(
X
)
:WL
2
E
I~ 0
.
2
5 (
d
e
t
i
o
nc
o
n
s
t
r
a
i
n
t
)
S
u
b
j
e
c
1t
o
:g

/
1
xI-8
y~ 0 (
n
o
r
m
a
Is
t
r
e
s
sc
o
n
s
t
r
a
i
n
t
)
8
2
(
X
)
:WL

g
3
(
X
)
:WO
c-21x3

y
::
S
;0 (
s
h

Someg
e
o
m
e
t
r
i
cc
o
n
s
1
r
a
i
n
t
st
or
e
l
a
t
et
h
ev
a
r
i
o
u
sd
e
s
i
g
nv
a
r
i
a
b
l
e
sa
r
e

g
4
(
X
)
:XI- 3X2~ 0
8
'
(
X
)
:2
x
2- XI~ 0

g
6
(
X
)
:X3- 1
.
5
x
4~ 0
g
7
(
X
)
:0.
5
x
4-X3~ 0
Thes
i
d
ec
o
n
s
t
r
a
i
n
t
soned
e
s
i
g
nv
a
r
i
a
b
l
e
sa
r
e
S
;x
0 2~X2 :S; 1
5 0
.
1
2
5:
S
;x3~ 0
.
7
5 0
.
2
5:
S
;x4:
S
;1
.2
5
3:
1~ 2

Th
ef
o
l
l
o
w
i
n
gr
e
l
a
t
i
o
n
sd
e
f
i
n
eec
r
o
s
s
s
e
cl
Io
n
a
la
r
e
a(
A
)

momento
fi
n
e
r
(d
c
m
omento
fa
r
e
aa
b
o
u
1t
h
ec
e
n
t
r
o
i
d(
Q
c
)i
nt
e
r
m
so
ft
h
ed
e
s
i
g
nv
a
r
i
a
b
l
e
s
:

Ac=2
X4+x
x
x
x
1
3-2
4
3

x
f

-x~(xl
(
I

1
2

7
.
4
.
2 Example7
.
BeamD
eslgn
T
h
i
ss
e
c
1
i
o
nr
u
i
r
e
seu
s
eo
feO
p
t
i
m
i
z
a
t
i
o
nT
o
o
l
b
o
xf
r
o
mM
A
T
L
A
B
.More
J
ly

i
tu
s
e
s1
h
eQPf
u
n
c
t
i
o
nsolve1
h
eQPs
u
b
p
r
o
b
l
e
mi
nt
h
eSQPme
d
.
s
p
e
c
i
f
i
c
a
Thep
r
o
b
l
e
mi
ss
o
l
v
e
du
s
i
n
gt
h
ed
i
r
e
c
tp
r
o
c
e
d ei
l
Iu
s
t
r
a
l
e
dt
h
r
o
u
g
ht
h
ee
x
a
m
p
l
ei
n

s
sc
o
t
)

-2
x
3
)
3

1
2

x
Qc=0.5x
J+0
.5
3
(
x1-xJ2

I-x
2
Thed
e
s
i
g
np
a
r
a
m
e
t
e
r
sf
o
rt
h
ep
r
o
b
l
e
ma
r
eW=1
.
4x2
0
0
0(
1.
4i
s f
a
c
t
o
ro
fs
a
f
e
t
y
)
L=9
6
E=2
9E+06p=0
.
2
8
4
y
=36E+03
:=21E+
0
3
.

312 N
U
M
E
R
I
C
A
LT
E
C
H
N
I
Q
U
E
SF
O
RC
O
N
S
T
R
A
I
N
E
DO
P
T
I
M
I
Z
A
T
I
O
N

f
i
l
ea
lw
i
l
li
m
p
l
e
m
e
n
tt
h
eS
o
l
u
l
i
o
nl
h
r
o
u
g
ht
h
eSQP.
EX343.m: T
h
i
si
se m
Whilet
h
i
si
sat
r
a
d
i
t
i
o
n
a
Idesignproblem.i
ti
sa
Is
on
o
lt
r
i
v
i
a
.
lI
ti
sr
a
r
et
of
i
n
da
p
u
b
l
i
s
h
e
dd
e
t
a
i
l
e
dp
r
o
c
e
d
u

f
o
rs
o
l
u
t
i
o
n
. The s
o
l
u
t
i
o
ni
su
s
ua
Il
yo
b
t
a
i
n
e
d
n
u
m
e
r
i
c
a
l
l
y
.I
nanye
v
e
n
t
.a
na
I
y
t
i
ca
Ia
p
p
l
i
c
a
t
i
o
no
ft
h
eKTc
o
n
d
i
t
i
o
n
si
su
s
ua
I
l
yn
o
t.

(fore7c
o
n
s
t
r
a
i
n
t
s
)e
v
e
ni
fmanyo
ft
h
o
s
ec
a
s
e
s
c
o
n
t
e
m
p
l
a
t
e
db
e
c
a
u
s
eo
f27c
a
r
et
r
i
v
i
a
.
lI
n
c
Iu
s
i
o
noft
h
es
i
d
ec
o
n
s
t
r
a
i
n
t
sw
i
l
lp
r
o
v
i
d
ea
na
d
d
i
t
i
o
na
I8l
i
n
e
a
r
(
a
g
r
a
n
dt
o
o
f1
5c
o
n
s
t
r
a
i
n
t
s
)o
ft
h
eform(
i
I
Iu
s
t
e
df
o
rX)o
n
l
y)
c
o
n
s

g
s(
x
)
:-x1+3S0

ample7
.
4-0
ptlmalControl
7
.
4
.
3 Ex

Th
i
ss
e
c
t
i
o
ni
n
t
r
o
d
u
c
e
sa
n
o
t
h
e
rc
Ia
s
so
fp
r
o
b
l
e
m
s C
b
ea
d
d
r
e
s
s
e
dbyd
e
s
i
g
n
o
p
t
i
r
n
i
z
a
t
i
o
nt
e
c
h
n
i
q
u
e
s
.I
ti
st
h
eb
a
s
i
ct
r
a
j
e
c

o
r
yd
e
t
e
r
m
i
n
a
t
i
o
nproblemi
no
p
t
i
ma
I
c
o
n
t
r
o
.
lGiventwop
o
i
n
t
si
nap
l
a
n
ef
i
n
dap
a
a
twi
1
Ibet
r
v
e
r
s
e
dbetweenthem
u
s
i
n
gas
t
e
e
r
i
n
gv
t
o
ro
fc
o
n
s
t
a
n
tm
a
g
n
i
t
u
d
e
.Th
ep
a
t
hh
a
st
obea
v
e
r
s
e
di
nu
n
i
t
mc
b
ed
e
s
c
r
i
b
e
du
s
i
n
gF
i
g
u
r
e7
.
1
1
.I
ft
h
ep
r
o
b
l
e
m nbed
e
s
c
r
i
b
e
d
t
i
m
e
.Thep
r
o
b
l
i
nep
h
y
s
i
c
a
lp
l
a
n
e
.t
h
es
l
r
i
n
gv
e
c
t
o
rc
o
n
s
i
d
e
r
e
da
sv
e
l
o
c
i
t
y
.e
nt
h
ek
i
n
e
m
a
t
i
c
u
a
t
i
o
n
sf
o
rap
a
r
t
ic
Iemovingonepa
c
a
nb
ed
e
s
c
r
i
b
e
dbyt
h
ef
o
l
l
o
w
i
n
g
i
o
p
t
i
m
i
z
a
t
i
o
np
r
e
f
e
r
st
od
e
a
lw
i
t
ha
1
g
e
b
r
a
i
cf
u
n
c
t
i
o
n
s
)
.
d
i
f
f
e
r
e
n
t
i
a
le
q
u
a
t
i
o
n
s(
d
whe
X
.
Yl
o
c
a
t
eep
r
o
b
l
e
mi
np
h
y
s
i
c
a
ls
p
a
c
e
:
=Vc
o
s8
(
t
)
; 0StS1

g
I
2(
X
)
:xl-20S0

=Vs
i
n8(
1
)
;
o
d
ei
nt
h
em-f
i
Ief
o
re
f
o
l
l
o
w
i
n
g
Ther
e
a
d
e
ri
ss
t
r
o
n
g
l
yrecommended1
0r
e
v
i
e
w c
r
e
a
s
o
n
s
:

The power o
fs
y
m
b
o
l
i
cm
a
n
i
p
u
l
a
t
i
o
ni
s p
l
o
i
t
e
df
u
r
t
h
e
ru
s
i
n
gv
e
c
t
o
r
d
e
f
i
n
i
t
i
o
n
s1
0maket
h
ei
m
p
l
e
m
e
n
t
a
t
i
o
nf
a
i
r
l
ye
a
s
y
.Thes
y
m
b
o
l
i
cc
a
l
c
u
l
a
t
i
o
n
s
a
r
eu
s
e
dt
ogenemteed
a
l
af
o
rt
h
eQPs
u
b
p
r
o
b
l
e
m
.
y
m
b
o
l
i
cc
a
l
c
u
l
a
t
i
o
n
sa
r
ee
f
f
o
r
t
l
e
s
s
l
yi
n
l
e
g
r
a
l
e
d w
i
t
h n
u
m
e
r
i
c
a
l
Th s
c
o
m
p
u
t
a
t
i
o
n
s(
t
h
s
o
l
uL
iont
ot
h
eQPi
sn
u
m
e
r
i
c
a
l
)
.
ss
i
g
h
t
f
o
r
w
a
r
dandd
i
r
e
c
t(
a
n
ds
i
m
p
l
e
)
.
Thei
m
p
l
e
m
e
n
t
a
t
i
o
no
f SQPi

'hesolutionissi ificantlybetter anthenumericalsolutionbyMAT


L

ownSQPprogram(Cha
p
1
0
)
.
Theprogramr
e
q
u
i
r
e
st
h
s
t
a
r
t
i
n
gg
u
e
s
sf
o
rt
h
ed
e
s
i
g
nv
e
c
t
o
r
.Thef
u
n
c
t
i
o
n
sf
o
rt
h
e
examplea
r
h
a
r
dc
o
d
e
d
.Th
ef
i
n
a
ls
o
l
u
t
i
o
ni
sp
r
i
n
t
e
dt
ot
h
eCommandwindow:

0StS1

Thei
n
i
t
ia
Iandf
i
n
a
lc
o
n
d
i
t
i
o
n
sa
r
e(
t
1
1
e
s
eweret
h
ev
a
l
u
e
sc
h
o
s
e
nf
o
rt
h
ee
x
a
m
p
l
e
)

=
=
x
(
l
)=
3
; y(
l
)=
2

(
0
) 0
; y
(
O
) 0

H
e
r
e
.8
(
t
)i
sr
e
g
a
r
d
da
sac
o
n
t
r
o
lv
a
r
i
a
b
l
e
.I
Ii
sac
o
n
t
i
n
u
o
u
sf
u
n
c
t
i
o
no
ftb
e
t
w
e
e
n0S
c1ef
rom i
n
i
t
i
a
lp
o
i
n
t1
0t
h
ef
i
n
a
lp
o
i
n.
tI
nc
o
n
t
r
a
s
t
.
tS1t
h
a
tw
i
l
lt
r
a
n
s
f
e
rt
h
epa
d
e
s
i
g
no
p
t
i
m
i
z
a
t
i
o
n(
i
.e
.
.
h
i
sb
o
o
k
)d
e
t
e
r
m
i
n
e
ss
c
a
l
a
rv
a
l
u
e
so
ft
h
ed
e
s
i
g
nv
a
r
i
a
b
l
e
s
.
Theset
y
p
e
so
fp
r
o
b
l
e
m
sb
e
l
o
n
gt
ot
h
ea
r
e
ao
fo
p
t
i
m
a
lc
o
n
t
r
o
.
lTh
a
r
es
p
e
c
i
a
l
n
u
m
e
r
i
c
a
lt
e
c
h
n
i
q
u
e
sa
la
r
ea
v
a
i
l
a
b
l
e1
0s
o
l
v
eI
h
e
s
ec
a
t
e
g
o
r
i
e
so
fp
r
o
b
l
e
m
s
.One
a
p
p
r
o
a
c
ht
ot
h
e
s
ep
r
o
b
l
e
m
si
st
ou
s
ed
e
s
i
g
no
p
t
i
r
n
i
z
a
t
i
o
nt
e
c
h
n
i
q
u
e
s
.Th
i
si
sa
i
d
e
dby
u
s
i
n
gaparame
i
cr
e
p
r
e
s
e
n
t
a
t
i
o
nf
o
rt
h
ec
o
n
t
r
o
lf
u
n
c
t
i
o
n
.Manyd
i
f
f
e
r
e
n
tp
a
r
n
e
c
p
s
e
n
t
a
t
i
o
n
sf
o
rt
h
ec
o
n
t
r
o
lh
a
v
eb
e
e
nu
s
e
d
.s
u
c
ha
sc
u
b
i
cs
p
l
i
n
e
sandH
e
r
m
i
t
e
p
o
l
y
n
o
r
n
i
a
l
s
.I
nt
h
i
ss

u
d
yaB
e
z
i
e
rc
u
r
v
ci
su
s
e
dt
or
e
p
s
e
n
tt
h
ec
o
n
t
r
o
l
.The
d
i
f
f
e
r
e
n
t
i
a
le
q
u
a
t
i
o
n
sa
r
e
nn
u
m
e
r
i
c
a
l
l
yi
n
l
e
g
r
a
t
e
d1
0e
s
b
l
i
s
ht
h
j
e
c
l
o
r
y
.

nunu

'anu

zJ'i

nuJ

nU 4

nunU

nuzutJ
n
u
U06

o
n
ur
o

m5

I
n
i
t
i
a
lG
u
e
s
s
:
F
i
n
a
lD
e
s
i
g
n
:
O
b
j
e
c
t
i
v
e
:

A
D
D
I
T
I
O
N
A
LE
X
A
M
P
L
E
S 313

7.
4

Al
Ic
o
n
s
t
r
a
i
n
t
sa
r
s
a
t
i
s
f
i
e
d
:
A
c

i
v
ec
o
n
s
t
r
a
i
n
t
s
: 1(
d
e
f
l
e
c
t
i
o
n
)
.7.9(X2 l
o
w
e
r
b
o
u
n
d
)
.
1
0
(
.
l
owerbo
d
)
.
1
1(X4 l
o
w
e
rb
o
u
n
d
)
i
ss
i
g
n
i
f
i
c
a
n
t
1
y
Ther
e
a
d
e
ri
se
n
c
o
u
r
a
g
e
dt
oe
x
p
l
o
r
ewhyt
h
es
o
l
u
t
i
o
no
b
t
a
i
n
e
dh
e
b
e
t
t
e
r t
h
eoneu
s
i
n
gt
h
eSQPfromt
h
et
o
o
l
b
o
x
.g
r
a
n
t
e
dt
h
a
tt
h
ei
m
p
l
e
m
e
n
t
a
t
i
o
n
s
a
r
en
o
tt
h
es
a
m
e
.Somet
h
o
u
g
h
ts
h
o
u
l
db
eg
i
v
e
nt
o penormanceo
fq
u
a
s
i
N
e
w
t
o
n
(
m
e

r
i
c
)methodsandt
h
e
i
ra
d
a
p
t
a
t
i
o
n
0c
o
n
s
t
r
a
i
n
e
dl
e
c
h
n
i
q
usi
ft
h
eH
e
s
s
i
a
ni
sn
o

p
o
s
i
t
i
v
ed
e
f
i
n
i
t
ewhichi
nt
h
i
sc
a
s
ei
ti
s
.

x
F
I
g
u
r
e
7

D
e
s
c
r
i
p
t
l
o
no
fI
r
a
j
e
c
l
o
r
yn
l
r
o
l
.

314

NUMERICALTECHNIQUESFORCONSTRAINEDOPTIMIZATION

7
.
4 ADDITIONAlEXAMPLES

S
o
l
u
t
l
o
nTechnique (
a
d
o
p
t
e
di
ni
ss
e
c
t
i
o
n
)
UseaB
e
z
i
e
rc
u
r
v
et
op
a
r
a
m
e
t
r
i
z
et
h
ec
o
n
t
r
o
lf
u
n
c
t
i
o
n
.
I
n
t
e
g
r
a
t
et
h
ed
i
f
f
e
r
e
n
t
i
a
le
q
u
a
t
i
o
n
sf
o
r
w
a
r
du
s
i
n
g aR
u
n
g
e
K
u
t
t
a method
45
.
m
)
(
o
de
Use u
n
c
o
n
s
i
n
e
dl
e
c
h
n
i
q
u
e(
f
m
i
n
u
n
c
.
m
)m
i
n
i
m
i
z
e
rt
oi
v
ees
q
u
a
r
eo
f
e
e
r
r
o
r
i
nef
i
n
a
lc
o
n
d
i
t
oz
e
r
o0
s
ts
q
u
a
r
e
r
r
o
r
)
.
o
o
l
b
o
xt
oe
x
e
c
u
t
ef
m
i
n
u
n
c
.
m
.A
l
t
e
m
a
t
e
l
y
.e
T
h
i
sexamplendst
h
eM ATI.AO t
u
n
c
o
n
s
t
r
a
i
n
e
dm
i
n
i
m
i
z
e
rc
a
nb
eamongt
h
o
s
ep
r
e
v
i
o
u
s
l
yd
e
v
e
l
o
p
e
d
. i
st
r
a
n
s
l
a
t
e
s
i
n
l
oa
nu
n
c
o
n
s
i
n
e
dminimump
r
o
b
l
e
m
w
h
i
c
ht
h
e
r
e
f
o
r
es
h
o
u
l
dh
a
v
eb
e
e
nh
a
n
d
l
e
d
i
nC
h
a
p
t
e
r6
.I
ti
si
nc
udedh
e
r
eb
e
c
a
u
s
eed
i
f
f
e
r
e
n
t
i
a
le
q
u
a
t
i
o
n
sa
r
ea
c
t
u
a
l
l
y
r
e
n
t
i
a
lc
o
n
s
l
r
a
i
n
t
s
.Theo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
nt
h
e
r
e
f
o
r
ei
s
c
o
n
s
i
d
e
r
e
dd
M z
e

f
{X}=
['
f
i
n
n
l
x
(
l)
f+[
Y
f
i
n
a
l
-y
(l
)J
2

Ex744.mw
i
l
le
x
e
c
u
t
et
h
ee
x
a
m
p
l
ea
n
dp
r
o
d
u
c
ep
l
o
t
so
ft
h
ec
o
n
t
r
o
Ia
n
dt
h
et
r
j
t
o
r
y
.
I
tu
s
e
stwos
t
a
n
d
a
r
dMATLAOf
u
n
c
t
i
o
n
sm
e
n
t
i
o
n
e
db
e
f
o
r
e(
o
d
c
4
5
.
ma
n
df
m
i
n
u
n
c
.
m
)
.
u
i
r
e
sc
o
e
f
f
.
m
.Comblnatlon.m.F
a
c
t
o
r
l
a
l
.
m
.l
e
TheB
e
z
i
e
rc
u
r
v
eg
e
n
e
r
a
t
i
o
nr
o
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
ni
sc
o
n
s
t
r
u
c
t
e
di
nobLoptcont.m.I
t q
u
i st
h
eu
s
eo
fo
de4
5
.
m
.

315

which r
e
q
u
i
r
e
s brachi.m t
or
e
t
u
mt
h
es
t
a
t
es
p
a
c
ed
e
f
i
n
i
t
i
o
no
f p
r
o
b
l
e
m
.
brachi.ma
l
s
oi
n
c
l
u
d
e
st
h
eB
e
z
i
e
ri
n
t
e
r
p
r
e
t
a
t
i
o
nc
o
d
e
.
F
i
g
u
r
e7
.
1
2d
e
s
c
r
i
b
e
seB
e
z
i
e
rv
e
i
c
e
so
ft
h
ec
o
n
t
r
o
lf
u
n
c
t
i
o
na
tt
h
es
t
a
nand
f
i
n
a
li
t
e
r
a
t
i
o
n
s
.Thef
i
n
a
lc
o
n
t
i
n
u
o
u
sc
o
n
t
r
o
li
sa
l
s
op
l
o
t
t
e
dont
h
ef
i
g
u
r
e
.Thes
t
a
r
t
a
i
g
h
tl
i
n
et
h
r
o
u
g
ht
h
ev
e
r
t
i
c
e
sandi
sn
o
ts
h
o
w
n
.F
i
g
u
r
e7
.
1
3showse
c
u
r
v
ei
sas
s
t
a
r
tt
r
a
j
e
c
t
o
r
y d ef
i
n
a
l
j
e
c
t
o
r
y
.I
ti
sc
l
r
frome
f
i
g
u
r
e ei
n
i
t
i
a
lande
j
e
c
t
o
r
y
.S
i
n
c
eee
q
u
a
t
i
o
n
swe

f
i
n
a
lc
o
n
d
i
t
i
o
n
sa
r
es
a
t
i
s
f
i
e
df
o
rt
h
ef
i
n
a
lt
r
i
n
t
e
g
r
a
t
e
d
.t
h
ed
i
f
f
e
r
e
n
t
i
a
lc
o
n
st
r
a
In
t
sa
r
et
h
e
r
e
f
o
r
ec
o
m
p
l
e
t
e
l
ys
a
t
i
s
f
i
e
d
.
Thef
i
n
a
lt
r
e
c
t
o
r
yshowni
so
n
l
yf
e
a
s
i
b
l
e
i
ti
soneo
fmany
canp

b
e
t
w
nt
h
etwop
o
i
n
t
s
.T
h
i
so
p
e
n
st
h
ep
o
s
s
i
b
i
l
i
t
yataf
u
r
t
h
e
rop
t
i
o
nc
b
e
p
e
c
tt
oj
e
c
t
o
r
y
.F
o
r
e
x
a
m
p
l
e
.f
i
n
des
h
o
r
t
e
s
tp
a
t
hi
ne
t
i
m
e
p
e
r
f
o
r
m
e
dw
i
t
h s
p
r
o
v
i
d
e
d
.Not
i
ti
sn
o
tes
t
r
a
i
g
h
tp
a
s
i
n
c
ef
o
rag
i
v
e
nv
e
l
o
c
i
t
ym
a
g
n
i
t
u
d
et
h
e
d
e
s
t
i
n
a
t
i
o
nmayb
e a
c
h
e
ds
o
o
n
e
ranp
s
c
r
i
b
e
d
.Thec
a
s
s
i
c
a
lp
r
o
b
l
e
mi
no
p
t
i
m
a
l
ntroli
seb
r
a
c
h
i
s
l
o
c
h
r
o
n
ep
r
o
b
l
e
m
.l
ti
seminimumt
i
m
ef
o
rt
h
ep l
e1
0
v
e
lb
e
t
w
e
e
ntwop
r
e
s
c
r
i
b
e
dp
o
i
n
t
sd
y
n
a
m
i
c
a
l
l
y(
o
b
e
y
i
n
gp
r
i
n
c
i
p
l
e
so
fd
y
n
a
m
i
c
s
)
.
t
r
e
c
h
n
i
q
u
ea
n
dec
o
d
et
oe
x
p
l
o
r
es
e
Ther
e
a
d
e
ri
sc
n
c
o
u
r
a
g
e
dt
oe
x
p
a
n
d t
a
d
d
i
t
i
o
na
i
n
t
e
r
e
s
t
i
n
gp
r
o
b
l
e
m
s
.

e
r
t
l
c
e
sa
n
d0(
1
)
Thev

F
e
a
s
l
b
l
et
r
o
J
e
c
t
o
r
y
3
.
5

l
l
n
a
lp
o
l
n
l

'

0
.
5

2
.
5

>
- 1.
5

dy

0.
5

.
.
.

a
'

a'

1
.
5

0
.
1

0.
2

0
.
3

0
.
4

0
.
5

0
.
6

t
l
m
e
o
n
t
r
o
lf
u
n
c
l
i
o
n0
(
/
)
F
l
g
u 7
.
1
2C

0
.
7

O
.

0
.
9

-0
.
5
0

0
.
5

1
.
5

2
.
5

F
l
g
u 7
.
1
3T
r
a
j
e
c
l
o
r
i
e
s
t
a
r
t n
a
.
l

3
.
5

PROBLEMS 317

316 N
U
M
E
R
I
C
A
LT
E
C
H
N
I
Q
U
E
SF
O
RC
O
N
S
T
R
A
I
N
E
DO
P
T
I
M
I
Z
A
T
I
O
N
REFERENCES
1
.F
i
a
c
c
o A
.V
. a
n
d McC
o
n
n
i
c
k G
.P
.Nonlinear P
r
o
g
r
a
l
l
l
m
i
J
l
g
Sequential
U
n
c
o
n
s
t
r
a
i
l
l
e
dM
i
n
i
m
i
z
o
t
i
o
l
lT
e
c
h
n
i
q
u
e
s

W
i
l
e
yNewY
o
r
k
1
9
6
8
.
p
l

G
.N
.NumericalOptimizotion TechlliquesforEngineering D
e
s
i
g
l
l
2
.V r
McGraw-Hi
INewYork1
9
8
4
.
An
s
y
sI
n
c
.P
i
t
t
s
b
u
r
g
hP
A
.
3
.ANSYSS
o
f
t
w
a
r
e
D
.P
.
C
o
n
s
t
r
a
i
n
e
dO
p
t
i
m
i
z
o
t
i
o
nandLa
grangeMethods

A
c
a
d
e
m
i
cP
r
e
s
s

New
4
.B
e
r
t
e
k
a
s
Y
o
r
k
1
9
8
2
.
l
e
y
J
.E
.
TheC
u
t
t
i
n
gP
l
a
n
Me
o
d
JoumalofSIAM
Vo
l
.8

p
p
.7
0
2
7
1
21
9
6
0
.
5
.KeJ
6
.G
i
l
l
P
.E
.
M
u
r
r
a
y
.W.
a
n
dW
r
i
g
h
tM.H
.
P
r
a
c
t
i
c
a
lO
p
t
i
m
i
z
o
t
i
o
n
A
c
a
d
e
m
i
cP
r
e
s
s

New
1
9
81
.
Y
o
r
k
7
.Boo

1J
.C
.G
.Q
u
a
d
r
a
u
cP
r
o
g
r
a
m
m
i
n
g
Sludiesi
t
lM
a
t
h
e
m
a
t
i
c
a
landMQlla
g
e
r
i
a
l
Economics

H
.T
h
e
i
l
(d
V
o
l
.2

N
o
r
t
h
H
o
l
l
a
n
dAm
s
t
e
r
d
n
1
9
6
4
.
n
e
dM
i
n
i
m
i
z
a
t
i
o
nU
s
i
n
gR
e
c
u
r
s
i
v
eQ
u
a
d
r
a
u
cP
r
o
g
r
a
m
m
i
n
g
:Some
8
.B
i
g
g
s

M.C
.
C
o
n
s
t
r
TowardsG
l
o
b
a
lOpt
I
I
z
o
t
i
o
n
L
.C
.W.D
i
x
o
na
n
d
A
l
t
e
m
a
t
eS
u
b
p
r
o
b
l
e
mF
o
n
n
u
l
a
u
o
n
s
p
p
.3
4
1
3
4
9N
o
r
t
h
H
o
l
l
a
n
d
A
m
s
t
e
r
d
w
n1
9
7
5
.
G
.P
.S
z
e
g
o(
e
d
s
.
)
M.J
.D
.AF
a
s
tA
l
g
o
r
i
t
h
mf
o
rN
o
n
l
i
n
e
a
rC
o
n
s
l
r
a
i
n
e
dO
p
t
i
m
i
z
a
u
o
nCa
Ic
u
l
a
l
i
o
n
s
9
.P
o
w
e
l
l
N
o
.DAMPrP77INA2

U
n
i
v
e
r
s
i
t
yo
fC
a
m
b
r
i
d
g
eE
n
g
l
a
n
d1
9
7
7
.
S
.P
.A G10
b
a
l
l
yC
o
n
v
e
r
g
e
n
tM
e
t
h
o
df
o
rN
o
n
l
i
n
e
a
rP
r
o
g
r
a
m
m
i
n
gJoumal
1
0
.Han
O
p
t
i
m
i
z
o
t
i
o
nT
l
r
e
o
r
yandA
p
p
l
i
c
a
t
i
o
n
s

Vo
l
.2
2p
.2
9
71
9
7
7
.
1
1
.A
r
o
r
a
.J
.S
.
.I
n
t
r
n
d
u
c
t
i
o
l
l
l
oOplimwnD
e
.
r
i
s
n
McGrnw-Hi
l
I
NewY
o
r
k1
9
8
9
.
1
2
.B
r
a
n
c
h
M.A
.
a
n
dG
r
a
c
e

A
.
O
p
t
i
m
i
z
o
t
i
o
t
lToo
'
o
x

U
s
e
r
'
sG
u
i
d
e

M
a
t
h
W
o
r
k
sl
n
c
.1
9
9
6
.
P
.Mo
d
so
fN
o
n
l
i
n
e
a
rP
r
o
g
r
a
m
m
i
n
g
RecenlAdv
Qll
c
e
si
/
lM
a
t
h
e
m
a
l
i
c
a
l
1
3
.W
o
l
f
e
R
.L
.G
r
a
v
e
sa
n
dP
.W
o
l
f
e(
e
d
s
.
)McGrnw-Hi
I
l
NewY
o
r
k1
9
6
3
.
Programming

G
.A
.
D
n
dR
a
g
s
d
e

J
lK
.M
.
l
eG
c
n
e
r
o
l
i
z
e
dG
r
a
d
l
c
n
tM
e
t
h
o
d
:AR
e
l
i
a
b
l
eT
o
o
lf
o
r
1
4
.G
u
b
r
i
e
l
e
aID
e
s
i
g
n
.ASMEJoumalofEn
g
i
n
e
e
r
i
n
gandI
n
d
J
"S
e
T
B
Vo
.
l9
9

May1
9

.
Op
AH
u
a
n
g
.
H
.
Y
.
d
H
e
i
d
e
m
a
n
J
.
C
.
S u
e
n
t
ia
IG
r
a
d
i
e
n
tR
e
s
t
o
r
a
u
o
nA
I
g
o
r
i
t
h
m
1
5
.M
i
e
l
e
f
o
rt
h
eMini
m
i
z
a
u
o
no
fC
o
n
s
t
r
a
i
n
e
dF
u
n
c
t
i
o
n
s
Ord
i
n
a
r
ya
n
dC
o
n
j
u
g
a
t
eGra
d
i
e
n
t
Joumalo
f
O
p
t
i
m
i
z
o
t
i
o
nTheo
andA
p
p
l
i
c
a
t
i
o
n
s
Vo
l
.4
N
o
.41
9
6
9
.
V
e
r
s
i
o
n
s
ev
yA
.V
.a
n
d Gomez S
.S
e
q
u
e
n
t
ia
IG
r
a
d
i
e
n
t
R
e
s
t
o
r
a
t
i
o
nA
I
g
o
r
i
t
h
mf
o
rt
h
e
1
6
.L
O
p
t
i
m
i
z
a
u
o
no
faN
o
n
l
i
n

rC
o
n
s
t
r
a
i
n
e
dF
u
n
c
t
i
o
n
.Joumal'
h
eA
s
t
r
o
n
a
u
t
i
c
a
lS
c
i
e
t
l
c
e
s

S
p
e
c
i
a
l
l
s
s
u
eo
nNumic
Me
d
si
nO
p
t
i
m
i
z
a
u
o
n
D
e
d
i
c
a
t
e
dt
oA
n
g
e
l
oM
i
e
l
e
K
.H
.
W
e
l
l(
s
p
.e
d
.
)

Vo
.
lXXX
N
o
.2
.1
9
8
2
.
.B
.T
h
eG
r
a
d
i
e
n
tP
r
o
j
e
c
t
i
o
nM
e
t
h
o
df
o
rN
o
n
l
i
n
e
a
rP
r
o
g
r
a
m
m
i
n
g
P
a
r
tD
:
1
7
.R
o
s
e
nJ

SIAMJoumalof
A
p
p
l
i
e
dMathematics

Vol
.9

N
o
.41
9
6
1
.
N
o
n
l
i
n
e
a
rC
o
n
s
t
r
a
i
n
t
s

PROBLEMS
or
e
v
i
e
wc
o
n
t
o
u
rv
a
l
u
e
sandp
l
o
ta
u
t
o
m
a
t
i
c
a
l
l
y
7
.
1 ModifySec7_2_1.J)loLmt
s
e
l
c
c
t
e
dc
o
n
t
o
u
r
s
.

2 ModifySec7_2_1_calc.mt
oa
u
t
o
m
a
t
i
c
a
l
l
yc
a
l
c
u
l
a
t
et
h
ei
n
i
t
i
a
lm
u
l
t
i
p
l
i
e
r
s
7.
l
o
o
po
v
e
rp
e
n
a
1
t
ym
u
l
t
i
p
l
i
e
r
s
.andt
oa
u
t
o
m
a
t
i
c
a
l
l
yp
i
c
kt
h
ed
e
s
i
g
nv
e
c
t
o
r
s
a
t
i
s
f
i
e
st
h
es
i
d
ec
o
n
s
t
r
a
i
n
t
s
.

7~

(
u
s
e
sO
p
t
i
m
i
z
a
t
i
o
nT
o
o
l
b
o
x
)DevelopeSLPbyu
s
i
n
gt
h
eLPs
o
l
v
e
romt
h
e
.
O
p
t
i
m
i
z
a
t
i
o
nToolboxa
n
da
p
p
l
yt
oExample7
.1

7
.
4 S
o
l
v
eExampl
7
.
1bys
o
l
v
i
n
geQPp
r
o
b
l
e
mg
r
a
p
h
i
c
a
1l
y
.
7.
5 Dcvelop codet
of
i
n
daf
e
a
s
i
b
l
es
o
l
u
t
i
o
nfromag
i
v
e
ns
l
a
r
t
i
n
gv
e
c
t
o
rf
o
r
n
s
l
t
i
n
e
dp
r
o
b
l
e
m
s
.
7
.
6 D
e
v
e
l
o
paprogramt
oc
a
l
c
u
l
a
t
et
h
em
u
l
t
i
p
l
i
e
r
sf
o
rt
h
eNLPproblema
tag
i
v
e
n
d
e
s
i
g
n
.
7
.
7 D
e
v
e
l
o
paprogramf
o
rac
o
n
s
t
r
a
i
n
e
ds
t
e
p
s
i
z
ec
a
l
c
u
l
a
t
i
o
np
r
o
c
e
d
u
r
e
.
7
.
8 (
u
s
e
sO
p
t
i
m
i
z
a
t
i
o
nT
o
o
l
b
o
x
)Uset
h
eQPprogramfromt
h
et
o
o
l
b
o
xandd
e
v
e
l
o
p
y
o
u
rownSQPi
m
p
l
e
m
e
n
t
a
t
i
o
n
.
r
o
mMATLABt
os
o
l
v
e
7
.
9 (
u
s
e
sO
p
t
i
m
i
z
a
t
i
o
nT
o
o
l
b
o
x
)UseeSQPprogramf
Example7
.
1
.
7
.
1
0 Develop Sec7
_3_2.m t
oa
u
t
o
m
a
t
i
ca
11
ye
x
e
c
u
t
es
e
v
e
r
a
1i
t
e
r
a
t
i
o
n
st
o
c
o
n
v
e
r
g
e
n
c
e
.
7
.
1
1I
m
p
l
e
m
e
n
tanewv
e
r
s
i
o
no
fSQPwheret
h
eH
e
s
s
i
a
nm
a
t
r
i
xi
sm
a
i
n
t
a
i
n
e
da
st
h
e
i
d
e
n
t
i
t
ym
a
t
r
i
xf
o
ra
1
1i
t
e
r
a
t
i
o
n
s
.
7
.
1
2F
i
n
i
s
hs
o
l
v
i
n
gExample7
.
1u
s
i
n
gt
h
eGRGme
o
d
.

X
2
]andY=[
x x
)
]
T
.
7
.
1
3S
o
l
v
eExample7
.
1w
i
t
hZ=[
7
.
1
4 Modify t
h
eGRGc
o
d
et
oi
n
c
l
u
d
ec
o
n
s
i
d
e
r
a
t
i
o
no
fa
c
t
i
v
ec
o
n
s
t
r
a
i
n
t
sa
n
d
comp
ep
e
r
f
o
r
m
c
e
.
7
.
1
5B
u
i
l
di
nKTc
o
n
d
i
t
i
o
nc
a
l
c
u
l
a
t
i
o
ni
n
t
ot
h
eSGRA.
7
.
1
6S
o
l
v
et
h
eB
r
a
c
h
i
s
t
o
c
h
r
o
n
eo
p
t
i
m
a
lc
o
n
t
r
o
lp
r
o
b
l
e
m
.

D
I
S
C
R
E
T
EO
P
T
I
M
I
Z
A
T
I
O
N

8
DISCRETEOPTIMIZATION

T
h
i
sc
h
a
p
t
e
ri
m
r
o
d
u
c
e
ssomec
o
n
c
e
p
t
sm
e
t
h
o
d
sa
n
da
l
g
o
r
i
t
h
m
sa
s
s
o
c
i
a
t
e
dw
i
d
i
s
c
r
e
t
eo
p
t
i z
a
t
i
o
n
.I
ti
sn
o
tp
o
s
s
i
b
l
et
oc
o
m
p
r
e
s
s s
u
to
fD
i
s
c
r
e
t
e
o
p
t
i
m
i
z
a
u
o
nt
oas
i
n
g
l
ec
h
a
p
t
e
ra
st
h
et
o
p
i
ci
t
s
e
l
fi
sc
a
p
a
b
l
eo
fs
p
a
w
n
i
n
gs
e
v
e
r
a
l
d
i
f
f
i
c
u
l
t
d
i
v
e
r
s
e
andc
o
n
t
i
n
u
e
st
o
c
o
u
r
s
e
s
.D
i
s
c
r
e
t
eo
p
t
i
m
i
z
a
t
i
o
ni
sv
e
r
yd
i
f
f
e
r
e
n
t
d
e
v
e
l
o
pe
v
e
n t
o
d
a
y
.T
h
e
s
ep
r
o
b
l
e
m
sa
r
el
a
r
g
e
l
yc
o
m
b
i
n
a
t
o
r
i
a
la
n
da
r
e
o
na
1l
ymoret
i
m
ei
n
t
e
n
s
i
v
e c
o
r
r
e
s
p
o
n
d
i
n
gc
o
n
t
i
n
u
o
u
sp
r
o
b
l
e
m
s
.
compu
o
nr
e
s
e
a
r
c
hcommunitya
r
em
a
i
n
l
y
Thep
r
o
b
l
e
m
sa
d
d
r
e
s
s
e
dbyt
h
ed
i
s
c
r
e
t
eo
p
t
i
m
i
z
i
n e
a
ofo
p
e
r
a
t
i
o

s
e
c
h
ch

a
c

r
i
z
u
s
u
a
l
l
ybyl
i

n
e

a
r
mo
o
p
t
i
m
i
z
a
t
i
on

m
o
s
t
l
yi
n
c

o
r

p
o
r
a
t

.
e
s
n

o
n
l
i
n
e
a
r
r

e
l

t
i

o
n
I

S
.Fromar
e
a
lp
e
r
s
p
e
c
t
iv
e

d
i
s
c
r
e
t
e
e
sI
l
m
d
e
s
i
g
nv
a
I
t
a
b
l
e
s
M
B
f
u
n
d
a
m
e
n
t
a
li
n
e
n
g
i
n
e
e
d
n
go
p
t
i
m
i
z
a
t
i
o
n
.
hebeamd
p
l
m i
n Chap
r 1
a p
i
c
a
ls
o
l
u
t
i
o
ns
h
o
u
l
di
n
v
o
l
v
i
d
e
n
t
i
f
y
i
n
ga
n

o
f
f
t
h
e
s
h
e
l
f
'beama
ser
o
l
l
i
n
gr
n
i
l
lw
i
l
lp
r
o
b
a
b
l
ymakei
tp
r
o
h
i
b
i
t
i
v
e
l
ye
x
p
e
n
s
i
v
e
n
i
q
u
ebeam wasi
d
e
n
t
i
f
i
e
d t
h
es
o
l
u
t
i
o
nt
oe
f
o
ral
i
m
i
t
e
dp
r
o
d
u
c
t
i
o
no
f u
fd
i
f
f
e
r
e
n
t
c
o
n
t
i
n
u
o
u
so
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
e
m
.I
nt
h
ep
r
o
b
l
e
mr
e
g
a
r
d
i
n
genumbero
d
i
s
c
r
e
t
ev
a
l
u
e
)
.
p
l
a
c
e
m
e
n
tm
a
c
h
i
n
e
st
h
es
o
l
u
t
i
o
n was e
x
p
e
c
t
e
dt
ob
ei
n
t
e
g
e
r
s(
rnilarlchoiceofdian r
s
l
e
n
g
t
h
sw
a
s
h
e
r
s
v
a
l
v
e
s
mencomponentsck
oona
r
eu
s
u
a
l
l
yg
u
i
d
e
dbya
v
o
i
d
i
n
gh
i
g
hp
r
o
c
u
r
e
m
e
n
tc
o
s
t
sa
s
s
o
c
i
a
t
e
d
S
l
z
e
SG s
w
i n
o
n
s
d
a
r
dc
o
m
p
o
n
e
n
t
su
n
l
e
s
sc
o
s
tr
e
c
o
v
e
r
yi
sp
o
s
s
i
b
l
eb
e
c
a
u
s
eo
fl
a
r
g
e

n
u

t
i
I
Il
O
l
v
o
l
u
m
e
s
.C
o
i
t
e
mst
ob
ec
o
m
p
l
e
t
e
l
ym
a
n
u
f
a
c
t
u
r
e
d
i
n

"

h
o

u
s
(
n
oo
f
l
l
fcom
n
e
n
li
s
n
e
c
e
s
s
a
r
y
)
P
r
a
c
t
i
c
a
le
n
g
i
n
e
r
i
n
gd
e
s
i
g
n q
u
i
r
l
a
lsomed
e
s
i
g
nv
a
r
i
a
b
l
e
sb
e
l
o
n
g(0 a
n
o
r
d
e
r
e
ds
e
to
fv
a
l
u
e
s
.
d
i
s
c
r
e
t
ev
a
r
i
a
b
l
e
s
.
T
h
i
smakesi
tad
i
s
c
r
e
t
eo
p
t
i
m
i
z
a
t
i
o
n
318

319

p
r
o
b
l
e
m
.However
therea
r
ev
e
r
yfewe
n
g
i
n
e
e
r
i
n
gp
r
o
b
l
e
m
s a
r
ec
u
r
r
e
n
t
l
y
s
o
l
v
e
da
sad
i
s
c
r
e
t
eo
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
e
m
.F
i
r
s
t

d
i
s
c
r
e
t
eo
p
t
i
m
i
z
a
t
i
o
na
l
g
o
r
i
t
h
m
s
h
e
ya
r
et
i
m
ec
o
n
s
u
m
i
n
g
.T
h
i
r
dmostd
i
s
c
r
e
t
e
a
r
ed
i
f
f
i
c
u
l
lt
oa
p
p
l
y
.Secondt
a
l
g
o
r
i
t
h
m
sa
n
dc
o
d
ea
d
d
r
e
s
sl
i
n
e
a
rm
a
t
h
e
m
a
t
i
c
a
lm
o
d
e
l
s
.D
i
s
c
r
e
t
eo
p
t
i
m
i
z
a
t
i
o
ni
n
l
1n
e
c
e
s
s
a
r
i
l
yi
n
v
o
l
v
ea
d
a
p
t
a
t
i
o
n
sfrom ec
u
r
r
e
n
t
l
ya
v
a
i
l
a
b
l
e
e
n
g
i
n
e
e
r
i
n
gwi

t
e
c
h
n
i
q
u
e
su
s
e
d by t
h
ed
e
c
i
s
i
o
n
m
a
k
i
n
g communily. Such a
d
a
p
t
a
t
i
o
n
sa
r
uncommon.I
ti
sa
l
s
or
a
r
et
of
i
n
d ybookond
e
s
i
g
no
p
t
i
m
i
z
a
t
i
o
n a
d
d
r
e
s
s
e
s
n
o
t
w
i
t
h
s
t
a
n
d
i
n
gt
h
ef
a
c
t
o
ri
n
c
l
u
d
e
st
h
es
u
b
j
e
c
tofn
o
n
l
i
n
e
a
rd
i
s
c
r
e
t
eo
p
t
i
m
i
z
a
t
i
o
n
a
ti
ti
senormousi
ne
x
t
e
nt
.Am
o
d
e
s
te
f
f
o
r
ti
sb
e
i
n
gmadei
nt
h
i
sbookt
oa
c
q
u
a
i
n
t
ta
r
e
a
.
t
h
er
e
a
d
e
rw
i
t
ht
h
es
u
b
jc
l
et
y
p
i
c
a
1a
p
p
r
o
a
c
hf
o
ri
n
c
o
r
p
o
r
a
t
i
n
gd
i
s
c
r
e
t
ed
e
s
i
g
nv
a
r
i
a
b
l
e
si
ne
n
g
i
n
e
e
r
i
n
gi
s
t
os
o
l
v
et
h
ec
o
r
r
e
s
p
o
n
d
i
n
gc
o
n
t
i
n
u
o
u
so
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
e
ma
n
da
d
j
u
s
tt
h
eo
p
t
i
m
a
l
e
a
r
e
s
td
i
s
c
r
e
t
ev
a
l
u
e
s(
i
si
ss
i
r
n
i
l
a
rt
ot
h
er
o
u
n
d
i
n
gp
r
o
c
e
s
st
oa
r
r
i
v
e
d
e
s
i
g
nt
o n
n
t
e
g
e
rn
u
m
b
e
r
)
.A
c
c
o
r
d
i
n
gt
oF
l
e
t
c
h
e
r[
1
]
t
h
e
r
ei
snog
u
a
r
a
n
t
e
i
s
p
r
o
c
e
s
s
a
ta
ni
o
l
u
t
i
o
nc
a
nbeo
b
t
a
i
n
e
di
sw
a
y
.Veryo
f
t
e
ni
sr
o
u
n
d
i
n
g
i
sc
o
r
r
e
c
to
r agoods
mayr
e
s
u
l
ti
ni
n
f
e
a
s
i
b
l
ed
e
s
i
g
n
s
.G
i
v
e
n t
h
ea
l
t
e
m
a
t
i
v
ei
st
os
o
l
v
et
h
ed
i
s
c
r
e
t
e
as
y
s
t
e
m
a
t
i
ca
p
p
r
o
a
c
ht
oi
sr
o
u
n
d
i
n
gp
r
o
c
e
s
sh
a
s
o
p
t
i
r
n
i
z
a
t
i
o
np
r
o
b
l
e
mi
t
s
e
l
f
sb
a
s
e
donm
a
i
n
t
a
i
n
i
n
gaf
e
a
s
i
b
l
ed
e
s
i
g
n
becomea
c
c
e
p
t
a
b
l
e
.Ar
o
u
n
d
i
n
gp
r
o
c
e
s
s i
i
sa
v
a
i
l
a
b
l
ei
nR
e
f
e
r
e
n
c
e2
.

'
h
i
sc
h
a
p
t
e
ra
n
dt
h
en
e
x
tw
i
l
ld
i
f
f
e
rs
i
g
n
i
f
i
c
a
n
t
l
yomt
h
ep
r
e
v
i
o
u
sc
h
a
p
t
e
r
si
nb
o
t
h
c
o
n
t
e
n
ta
n
do
r
g
a
n
i
z
a
t
i
o
n
.T
h
e
i
rp
r
i
m
a
r
yf
o
c
u
si
sp
r
e
s
e
n
t
i
n
gnewi
d
e
a
sa
n
d
c
o
n
t
e
n
tr
a
t
h
e
r a
nd
e
v
e
l
o
p
i
n
gat
e
c
h
n
i
q
u
eo
ra
s
s
i
s
t
i
n
gi
nc
o
d
ed
e
v
l
o
p
m
e
n
.
tD
is
c
r
e
t
e
r
a
d
i
t
i
o
n
a
lc
o
n
t
i
n
u
o
u
s
p
r
o
b
l
e
m
sr
e
q
u
i
r
es
i
g
n
i
f
i
c
a
n
t
l
yd
i
f
f
e
r
e
n
tc
o
n
c
e
p
t
s et
m
a
t
h
e
m
a
t
i
c
a
lm
o
d
e
l
so
ft
h
ee
a
r
l
i
e
rc
h
a
p
t
e
r
s
.S
i
m
p
l
ee
x
a
m
p
l
e
sa
r
eu
s
e
dt
ob
r
i
n
go
u
t
ed
i
f
f
e
r
e
n
c
e
.F
o
re
x
a
m
p
l
ed
e
r
i
v
a
t
i
v
e
s
g
r
a
d
i
e
n
t
sa
n
dH
e
s
s
i
a
ndon
o
ta
p
p
l
yt
o
e
a
r
c
hd
i
r
e
c
t
i
o
n
sa
n
do
n
e
d
i
m
e
n
s
i
o
n
a
ls
t
e
p
s
i
z
e
d
i
s
c
r
e
t
ep
r
o
b
l
e
m
s
.Bye
x
t
e
n
s
i
o
ns
r
ea
r
es
v
e
r
a
le
x
c
e
l
1
e
n
tr
e
f
e
r
e
n
c
e
son
c
o
m
p
u
t
a
t
i
o
nl
o
s
et
h
e
i
rr
e
l
e
v
a
n
c
e
.W
h
i
l
ee
a
l
m
o
s
ta
l
lo
fthemd
e
a
lw
i
t
hl
i
n
e
a
rm
a
t
h
e
m
a
t
i
c
a
lm
o
d
e
l
s
.T
h
e
r
e
d
i
s
c
r
e
t
eo
p
t
i
m
i
z
a
t
i
o
n
fa
n
ya
b
o
u
tn
o
n
l
i
n
e
a
rd
i
s
c
r
e
t
eo
p
t
i
m
i
z
a
t
i
o
na
p
p
l
i
e
dt
oe
n
g
i
n
e
e
r
i
n
g
a
r
efewi
p
r
o
b
l
e
m
s
.
D
is
c
r
e
t
eo
p
t
i
r
n
i
z
a
t
i
o
np
r
o
b
l
e
m
s(
D
P
)a
r
ei
m
p
l
i
e
dwhent
h
ed
e
s
i
g
nv
a
r
i
a
b
l
e
sa
r
en
o
t
o
n
c
e
p
to
ft
h
es
i
d
ec
o
n
s
t
r
a
i
n
t
s
t
h
e
s
ed
e
s
i
g
nv
a
r
i
a
b
l
c
o
n
t
i
n
u
o
u
s
.E
x
t
e
n
d
i
n
g c

t
e
g
e
rv
a
l
u
eo
f8=>b
i
n
a
r
yv
a
l
u
eo
f1
0
0
0

320 D
I
S
C
R
E
T
EO
P
T
I
M
I
Z
A
T
I
O
N

.
1

i
fed
e
s
i
g
nv
a
r
i
a
b
l
exi
sr
e
s
c
t
e
dt
oi
n
t
e
g
e
r
sb
e
t
w
e
e
n0S
;xS
;1
6

t
h
e
nx
T
h
e
r
e
f
o
r
e
c
a
nb
e p
l
a
c
e
dbyf
i
v
e0-1d
e
s
i
g
nv
a
r
i
a
b
l
e
s[
y
.
.Y
2
Y
3
Y
4Y
s
Jfromwhichxc
b

a
s
s
e
m
b
l
e
da
s

0
.
2
20
.
7
51
x
.73 2
.
2
42
.
7
8
]EX3d
3[

eprogrammingp
r
o
b
l
e
m
s
I
nt
h
eo
p
e
n
i
n
gd
i
s
c
u
s
s
i
o
n somed
e
f
m
i
t
i
o
n
so
fd
i
s
c t
r
e
g
a
r
d
i
n
gt
h
en
a
t
u
r
eo
fd
i
s
c
r
e
t
eo
p
t
i
m
i
z
a
t
i
o
nw
e
r
ei
n
o
d
u
c
e
d
.
l
ni
ss
e
c
t
i
o
nt
h
e
e
s
u
l
t

c
o
n
c
e
p
t
sa
r
ed
e
t
a
i
l
e
du
s
i
n
gas
i
m
p
l
eu
n
c
o
n
s
t
r
a
i
n
e
dm
i
n
i
m
i
z
a
t
i
o
ne
x
a
m
p
l
e
.Asar
eo
p
t
i
m
a
l
i
t
yi
s
s
u
d
on
o
ti
n
d
eonp
r
e
s
e
n
t
i
n
gt
h
ei
m
p
o
r
t
a
n
ti
d
e
a
si
nd
i
s
c
r
e
t
e
o
p
t
i
m
i
z
a
t
i
o
ni
n
c
l
u
d
i
n
gt
h
et
r
e
a
t
m
e
n
to
ff
i
n
d
i
n
gac
o
n
t
i
n
u
o
u
ss
o
l
u
t
i
o
nf
o
l
l
o
w
e
dby
a
d
j
u
s
t
i
n
gt
h
ed
i
s
c
r
e
t
ed
e
s
i
g
nv
a
r
i
a
b
l
e
st
on
e
i
g
h
b
o
r
i
n
gd
i
s
c
r
e
t
ev
a
l
u
e
s
.
Example8
.
1 Mi
n
i
m
i
z
e o
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
ng
i
v
e
nbelow

w
h
e
r
e
x
li
sac
o
n
t
i
n
u
o
u
s
v
a
r
i
a
b
l
eandX
2

X
3a
r
ed
i
s
c
r
e
t
ev
a
r
i
a
b
l
e
s
.X
2musth
a
v
eav
a
l
u
efromt
h
es
e
t[
0
.
51
.
52
.
5
3
.
5
Jandx
3musth
a
v
eav
a
l
u
eumt
h
es
e
t[
0
.
2
20
.
7
51
.7
32.242
.
7
8
J
.

Min
i
m
i
z
e
f
(
x11x3
)=(x1- 2)2+(x1-~)2 +(x1-X3)2+(x2-x3)2 (8.1)

8
.
1
.
1 ProblemRelaxatlon

P44MJ

8
.
1 CONCEPTSI
NDISCRETEPROGRAMMING

(
8
.
2
c
)

Thes
y
m
b
o
lEi
d
e
n
t
i
f
i
e
sa
tt
h
ev
a
r
i
a
b
l
eont
h
el
e
f
tc
a
nh
a
v
eoneo
ft
h
ev
a
l
u
e
sont
h
e
nE
q
u
a
t
i
o
n(
8
.
2
a
)

Rs
t
a
n
d
sf
o
rar
e
a
lv
a
l
u
e
.T
h
i
sr
e
p
r
e
s
e
n
t
sas
t
a
n
d
a
r
du
s
e
r
i
g
h
ts
i
d
e
.I
y
m
b
o
lE di
sa
ne
f
f
e
c
t
i
v
ewayo
fe
x
p
r
e
s
s
i
n
gt
h
ei
d
e
at
h
a
tt
h
ed
i
s
c
r
e
t
ev
a
l
u
e
s
o
f s
C
o
n
l
ybes
e
l
e
c
t
e
dfromag
r
o
u
po
fv
a
l
u
e
s
.T
b
i
si
samixedprogramming(MP)
p
r
o
b
l
e
m
.

3
1
4
2
0
x=
Y
I(
2)+Y
2(
2)+Y
3(
2
)+Y
4(
2)+Y
S(2
)

Suchat
r
a
n
s
f
o
I
I
D
a
t
i
o
ni
sn
o
trecommendedi
fr
ei
sal
a
r
g
enumbero
fi
n
t
e
g
e
r
v
a
r
i
a
b
l
e
s
.E
n
g
i
n
e
e
r
i
n
gd
e
s
i
g
np
r
o
b
l
e
m
sc
a
nb
ee
x
p
e
c
t
e
dt
oc
o
n
t
a
i
nb
o
t
hc
o
n
t
i
n
u
o
u
s
v
a
r
i
a
b
l
e
s
.T
h
e
s
ea
r
et
e
r
m
e
dMixedProgramming(MP)p
r
o
b
l
e
m
si
fe
a
n
dd
i
s
c
r
e
m
a
t
h
e
m
a
t
i
c
a
lmodeli
sl
i
n
e
I
nt
h
i
sbookt
h
ec
l
a
s
s
i
c
a
t
i
o
ni
se
x
t
e
n
d
e
dt
on
o
n
l
i
n
e
a
r
p
r
o
b
l
e
m
st
o
o
.
I
nt
h
ei
n
t
e
r
e
s
to
fm
a
n
a
g
e
a
b
i
l
i
t
y
o
n
l
yt
h
r
e
em
e
t
h
o
d
s ep
r
e
s
e
n
t
e
di
ni
sc
h
a
p
t
e
r
.
o
p
i
co
fd
i
s
c
r
e
t
e
T
h
e
s
em
e
t
h
o
d
sa
r
eo
n
l
yr
e
p
r
e
s
e
n
t
a
t
i
v
ea
n
di
nnowaya
d
d
r
e
s
s t
op
n
i
z
a
t
i
o
ns
u
f
c
i
e
n
t
l
y

l
e
ta
l
o
n
ec
o
m
p
l
e
t
e
l
y
.Th
et
h
i
r
doneh
a
ss
e
e
nl
i
m
i
t
e
du
s
e

b
u
t
i
o
n
a
lr
e
s
o
u
r
c
e
sn
o
tbi
n
gah
u
r
d
l
et
o
d
a
y

i
ti
sd
e
f
i
n
i
t
e
l
ya
t
t
r
a
c
v
e
.The
w
i
t
hcomput
h
o
u
g
hem
e
t
h
o
d
sa
r
eamongt
h
o
s
ea
t
s
e
l
e
c
t
i
o
nd
o
e
sn
o
tr
e
p
r
e
s
e
n
ta
n
yp
r
i
o
r
i
t
yt
h
a
v
eb
e
e
na
p
p
l
i
e
do
f
t
e
ni
ne
n
g
i
n
e
e
r
i
n
g
.Th
em
e
t
h
o
d
so
ft
h
en
e
x
tc
h
a
p
t
e
rh
a
v
ee
v
o
l
v
e
d
t
h
r
o
u
g
ht
h
e
i
ri
n
i
t
i
a
la
p
p
l
i
c
a
t
i
o
nt
od
i
s
c
r
e
t
eo
p
t
i
m
i
z
a
t
i
o
n
.Theyc
o
u
l
da
l
s
ob
e
l
o
n
gt
o
i
sc
h
a
p
t
e
rb
u
thaveb
e
e
nk
e
p
ts
e
p
a
t
eb
e
c
a
u
s
ee
ya
r
et
h
ed
r
i
v
i
n
gf
o
r
c
ei
nt
h
e
sc
h
a
p
t
e
ra
r
e(
l
)E a
u
s
t
i
v
e
s
e
a
r
c
hf
o
rg
l
o
b
a
loptimumt
o
d
a
y
.Them
e
t
h
o
d
so
fi
E
n
u
m
e
r
a
t
i
o
n
(
2
)B
r
a
n
c
handBound(
p
a
r
t
i
a
lo
rs
e
l
e
c
t
i
v
ee
n
u
m
e
r
a
t
i
o
n
)
and(
3
)
r
o
g
r
a
m
m
i
n
g
.
DIamicP

CONCEPTSI
NO
I
S
C
R
E
T
EPROGRAMMING 3
2
1

Ei
sw
e
r
eau
n
c
o
n
s
t
r
a
i
n
e
dp
r
o
b
l
e
mi
nc
o
n
t
i
n
u
o
u
sv
a
r
i
a
b
l
e
s
t
h
es
o
l
u
t
i
o
nc
o
u
l
db
e
=0a
n
ds
o
lv
i
n
geee
q
u
a
t
i
o
n
sf
o
rev
a
l
u
s
o
f X
2
'a
nd
o
b
t
a
i
n
e
db
ys
e
t
t
i
n
gVf
X
3

E
q
t
i
o
n
s8
.
2
band8
.
2
ch
t
ob
eo
v
e
r
l
o
o
k
e
d
.A
l
t
e
m
a
t
e
l
yf
o
rac
o
n
t
i
n
u
o
u
s
i
n
s
p
e
c
t
i
o
no
ft
h
eo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
ny
i
e
l
d
s ef
o
l
l
o
w
i
n
gs
o
l
u
t
i
o
n
:
p
r
o
b
l
e
m

x
.

x
;
=
2
;x
;
=
2
;x
;
=
2
; I=O

(
8.
3
)

F
o
rt
h
eo
r
i
g
i
n
a
lp
r
o
b
l
e
m'
d
X
li
sd
e
f
i
n
e
d
b
u
tn
o
td
f
/
d
x
2o
rd
f
/
d
x
3
's
i
n
c
eX
2 dX3
a
r
ed
i
s
c
r
e
t
ev
a
l
u
e
s
.D
e
r
i
v
a
t
i
v
e
sa
r
edefmedbyt
a
k
i
n
gt
h
el
i
m
i
to
ft
h
er
a
t
i
oo
fc
h
a
n
g
e
o
fo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
n
f
t
ot
h
ec
h
a
n
g
ei
nt
h
ev
a
l
u
eo
ft
h
ed
e
s
i
g
nv
a
r
i
a
b
l
e

a
st
h
ec
h
a
n
g
e
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
n
f
i
so
n
l
yd
e
f
i
n
e
d
i
nt
h
ev
a
r
i
a
b
l
ea
p
p
r
o
a
c
h
e
sz
e
r
o
.Thev
a
l
u
eo
f o
2X
3i
nE
q
u
a
t
i
o
n(
8
.
2
)andi
sn
o
Ld
e
f
i
n
e
de
l
s
e
w
h
e
r
e
.
a
ts
e
l
e
c
t
e
dc
o
m
b
i
n
a
t
i
o
no
fX
S
m
a
l
l
l
i
n
f
i
n
i
t
e
s
i
m
a
lc
h
a
n
g
e
si
nt
h
ed
i
s
c
r
e
t
ed
e
s
i
g
nv
a
r
i
a
b
l
e
sa
n
dt
h
e
r
e
f
o
r
ei
ne
o
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
na
r
en
o
td
e
f
i
n
e
di
nExample8
.
1
.D
e
r
i
v
a
t
i
v
e
sw
i
t
hr
e
s
p
e
c
tt
o
d
i
s
c
r
e
t
ev
a
r
i
a
b
l
e
sdon
o
te
X
I
st
.T
h
i
sc
o
n
c
l
u
s
i
o
ni
so
fm01's
i
g
n
i
f
i
c
a
n
c
ea
si
tmakest
h
i
s
b
o
o
k
'
sp
r
e
v
i
o
u
sbodyo
fworko
fi
I
m
i
t
e
du
s
ei
nep
u
r
s
u
i
to
ft
b
es
o
l
u
t
i
o
nt
ot
h
eDP.
t

e
n
e
c
e
s
s
ya
nds
u
f
f
i
c
i
e
n
tc
o
n
d
i
t
i
o
n
s

whichd
r
o
v
et
h
ea
l
g
o
r
i
t
h
m
s

w
e
r
e
Tor
e
c
o
l
l
e
c
b
a
s
e
dont
h
eg
r
a
d
i
e
n
t
sandt
h
e
i
rd
e
r
i
v
a
t
i
v
e
s
.Twoa
r
e
a
si
nt
h
ep
r
e
v
i
o
u
sc
h
a
p
t
e
r
se
s
c
a
p
e
w
i
t
hal
i
m
i
t
e
di
m
p
a
c
to
fi
ss
t
a
t
e
m
e
n
t
.beyar
LPa
ndz
e
r
o
o
r
d
e
rm
e
t
h
o
d
sf
o
r
n
u
m
e
r
i
c
a
ls
o
l
u
t
i
o
nt
ou
n
c
o
n
s
t
r
a
i
n
e
dn
o
n
l
i
n
e
a
ro
p
t
i
m
i
z
a
t
i
o
n
.B
o
t
ho
fthemp
l
a
ya
s
i
g
n
i
f
i
c
a
n
tr
o
l
ei
nd
i
s
c
r
e
t
eo
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
e
m
s
.I
ti
st
h
e
r
e
f
o
r
ep
o
s
s
i
b
l
et
oc
o
n
c
l
u
d
e
a
ts
o
l
u
t
i
o
nt
oExample8
.
1 e
s
t
a
b
l
i
s
h
e
di
nE
q
u
a
t
i
o
n(
8
.
3
)i
si
n
c
o
r
r
e
c
t(
es
i
d
e
c
o
n
s
t
r
a
i
n
t
sa
r
ei
nv
i
o
l
a
t
i
o
n
)
.
I
nDPt
h
es
o
l
u
t
i
o
ni
nE
q
u
a
t
i
o
n(
8.
3
)r
e
p
r
e
s
e
n
t
st
h
es
o
l
u
t
i
o
nt
oar
e
l
a
x
a
t
i
o
np
r
o
b
l
e
m
.
P
r
o
b
l
e
mr
e
l
a
x
a
t
i
o
nc
a
nt
a
k
es
e
v
e
r
a
l
f
o
r
m
s
.M
o
s
t
l
yi
ti
sa
p
p
l
i
e
dt
oer
e
l
a
x
o
nore
w
e
a
k
e
n
i
n
goft
h
ec
o
n
sa
i
n
o
rt
h
eo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
n
s
.T
h
e
r
e noe
x
p
l
i
c
i
t
n
i
si
n
s
t
a
n
c
e

t
h
er
e
l
a
x
a
t
i
o
nr
e
f
e
r
st
ot
h
er
e
m
o
v
a
lo
ft
h
e
c
o
n
s
t
r
a
i
n
t
si
nExample8
.
1
.I
r
e
s
t
r
i
c
t
i
o
no
fd
i
s
c
r
e
l
e
n
e
s
so
f
t
h
ev
a
r
i
a
b
l
e
s
.T
h
i
si
si
d
e
n
t
i
f
i
e
da
sc
o
n
t
i
n
u
o
u
sr
e
l
a
x
a
t
i
o
n
.
Ther
e
l
e
dp
r
o
b
l
e
mo
rr
e
l
a
x
a
t
i
o
nh
a
ss
e
v
e
r
a
la
d
v
a
n
t
a
g
e
s[
3
]
.

Thes
i
d
ec
o
n
s
t
r
a
i
n
t
sont
h
ed
e
s
i
g
nv
a
r
i
a
b
l
e
sc
a
nb
es
e
tupa
s

.I
fac
o
n
s
t
r
a
i
n
tl
r
l
a
x
a
t
i
o
ni
si
n
f
e
a
s
i
b
l
e

s
oi
st
h
ep
r
o
b
l
e
m
l
m
o
d
e
li
tr
e
l
a
x
e
s
.

x1E R
X2E.
5

x
u

1
.5 2
.
53
.
5
JE

(
8
.
2
a
)

Constraintrelaxationexpandsthesetoffeasiblesolutions.Therelaxedoptimal

(
8
.
2
b
)

r
i
g
i
n
a
l
v
a
l
u
e must i
m
p
r
o
v
eo
re
q
u
a
lt
h
eb
e
s
tf
e
a
s
i
b
l
es
o
l
u
t
i
o
n1
0 eo
p
r
o
b
l
e
m
l
m
o
d
e
l
.

'
l

322 D
I
S
C
R
E
T
EO
P
T
I
M
I
Z
A
T
I
O
N

8
.
1 CONCEPTSI
ND
I
S
C
R
E
T
EPROGRAMMING 323
3

'1+'111t'

-e'iri-

-'''17iboa

o
awU-

q
u
a
t
i
o
n(
8.
3
)a
r
eas
o
l
u
t
i
o
nt
o c
o
n
t
i
n
u
o
l
I
sr
e
l
a
x
a
t
i
o
no
f
Thev
a
l
u
e
se
s
t
a
b
l
i
s
h
e
di
nE
Example8
.1
.I
twase
a
s
yt
oo
b
t
a
i
ni
ss
o
l
u
t
i
o
nw
h
i
l
eat
e
c
h
n
i
q
u
ef
o
rt
h
es
o
l
u
t
i
o
nt
o
e d
i
s
c
r
e
t
e un
i
n
e
d op
z
a
t
i
o
ni
ss
t
i
l
l unknown e KuhnTucker
c
o
n
d
i
t
i
o
n
sa
r
enol
o
n
g
e
ru
s
e
f
ul
.
l
es
o
l
u
t
i
o
ni
nE
q
u
a
t
i
o
n(
8.
3
)i
sn
o
ta
c
c
e
p
t
a
b
l
e
x
i
ja
r
en
o
te
l
e
m
e
n
t
so
ft
h
ep
e
r
m
i
s
s
i
b
l
es
e
1(
8
.
2
)
.Thev
a
l
u
eo
f
I=0willbebetler
andx
b
e
s
td
i
s
c
r
e
t
es
o
l
u
t
i
o
n
l
o
w
e
rbound0"t
h
es
o
l
u
t
i
o
n1
0 o
r
i
g
i
n
a
lp
r
o
b
l
e
m
.

ll1lr

o
p
t
i
m
a
ls
o
l
u
t
i
o
n
sa
ta
r
ee
a
s
i
l
yr
o
u
n
d
e
dt
o
Manyr
e
l
a
x
a
t
i
o
nmodelsp
r
o
d
u
c
i
sa
p
p
e
a
r
st
o
goodf
e
a
s
i
b
l
es
o
l
u
t
i
o
n
so
ft
h
eo
r
i
g
i
n
a
lproblemlmodel.l
d
i
s
c
r
e
t
eo
p
t
i
m
i
z
onf
o
re
n
g
i
n
e
e
r
i
n
gd
e
s
i
g
np
r
o
b
l
e
m
s
.

--tJo--T

---

Theo
p
t
i
m
a
lv
a
l
u
eo
fanyr
e
l
a
x
e
dmodelp
r
o
v
i
d
e
sal
o
w
e
rboundone
o
p
t
i
m
a
l
s
o
l
u
t
i
o
ni
fi
li
sam
i
n
i
m
i
z
a
t
i
o
np
r
o
b
l
e
m
.S
i
m
i
l
a
r
l
yi
le
s
t
a
b
l
i
s
h
e
sa
nu
p
p
e
rbound
f
o
rm
a
x
i
m
i
z
a
t
i
o
np
r
o
b
l
e
m
s
.

!ptlmalSolutlon
8
.
1
.
2 DlscreteC

0
.
5

As
t
a
n
d
a
r
da
p
p
r
o
a
c
h1
0s
o
l
v
i
n
gd
i
s
c
r
e
1
eo
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
e
mp
a
r
t
i
c
u
l
a
r
l
yi
n
e
n
g
i
n
e
e
r
i
n
g
i
s1
0u
s
et
h
ec
o
n
t
i
n
u
o
u
sr
e
l
a
x
a
t
i
o
no
ft
h
em
a
t
h
e
m
a
t
i
c
a
lmodeland
genemteac
o
n
t
i
n
u
o
u
ss
o
l
u
t
i
o
n
.ForExample8
.
1t
h
i
si
sg
i
v
e
ni
nE
q
u
a
t
i
o
n(
8
.
3
)
.More
t
h
a
nl
i
k
e
l
yed
i
s
c
r
e
t
ev
a
r
i
a
b
l
e
sw
i
l
ln
o
tb
e
l
o
n
gt
ot
h
ep d
e
f
i
n
e
dd
i
s
c
r
e
t
es
e
t
s
.T
h
e
s
e
e
a
r
e
s
td
i
s
c
r
e
t
ev
a
l
u
e
s
.Forp
r
o
b
l
e
m
s
v
a
r
i
a
b
l
e
sa
r
et
h
e
nc
h
a
n
g
e
d(
o
rr
o
u
n
d
e
d
)t
oen
e
a
s
i
b
i
l
i
t
yi
sc
h
e
c
k
e
d
.Theb
e
s
tf
e
a
s
i
b
l
es
o
l
u
t
i
o
ni
sl
h
e
nc
h
o
s
e
n
.
w
i
t
hc
o
n
s
t
r
a
i
n
t
sf
U
n
l
i
k
ec
o
n
t
i
n
u
o
u
sp
r
o
b
l
e
m
st
h
e
r
ea
n
on
e
c
e
s
s
a
r
yands
u
f
f
i
c
i
e
n
tc
o
n
d
i
t
i
o
n
st
o
s
a
t
i
s
f
y
.
a
s
eo
fExample 8

1
. E
q
u
a
t
i
o
n
s(
8
.
2
)a
n
d(
8
.
3
)p
r
o
v
i
d
et
h
er
e
l
e
v
a
n
t
l
nec
i
n
f
o
r
m
a
t
i
o
nt
oe
s
t
a
b
l
i
s
ht
h
es
o
l
u
t
i
o
n
.Forc
o
n
v
e
n
i
e
n
c
e

t
J
l
e
ya
r
er
e
p
r
o
d
u
c
e
dh
e
r
e
:
X
J R
1
.
5 2
.
5 3
.
5
]eX2d

(
8
.
2
b
)

0
.
7
5 1
.
73 2
.
2
4 2
.
7
8
]EX
3d

(
8
.2
c
)

X2e [
0
.
5

X3 [
0
.
2
2

(
8
.2
a
)

;
x=2; X
;=2; ;
x=2; I=0

(
8.
3
)

8
.
1i
n
d
i
c
a
t
e
sf
o
u
r
lec
o
n
t
i
n
u
o
u
sv
a
r
i
a
b
l
eX
Ji
sc
h
o
s
e
n1
0h
a
v
eav
a
l
u
eo
f2
.
0
.F
i
g
u
s
e
t
so
fd
i
s
c
ev
a
l
u
e
sf
o
rX2andX3a
r
o
u
n
dt
h
ec
o
n
t
i
n
u
o
u
ss
o
l
u
t
i
o
n
.Thel
e
a
s
tv
a
l
u
eo
f
t
h
eo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
na
tt
h
e
s
ep
o
i
n
t
sw
i
l
lbec
o
n
s
i
d
e
r
e
des
o
l
u
t
i
o
n
.E
v
a
l
u
a
t
i
n
gt
h
e
v
a
l
u
e
so
ft
h
eo
b
j
e
c
t
i
v
en
c
t
i
o
na
tt
h
o
s
ep
o
i
n
t
s
:

;
x=2; X
;=1.5; X
;=1.73; 1=0.
37
5
8
;
x=2; ;
x=1.5; X
;=2.24; 1=0.8552

(
8.
4
a
)
(
8.
4b
)

_L__

0
.
5

1
.
5

D
l
s
c
r
e
t
ev
o
r
l
o
b
l
e~

2
.
5

3
.
5

F
l
g
u
r
e8
.
1 D
l
s
c
r
e
t
en
e
l
g
h
b
o
r
h
o
o
d0
1c
o
n
t
i
n
u
o
u
ss
o
l
u
t
i
o
n
.

X~ =2;

X
;=2

x
i=1
.7
3
; ] =0
.
9
1
5
8

(
8.
4c
)

x~ =2
;

;
x=2.5; xi=2.24; ] =0.3752

(
8
.
4
d
)

tv
a
l
u
eo
feo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
ni
sa
v
a
i
l
a
b
l
ei
nE
q
u
a
t
i
o
n
Fromt
h
i
se
x
e
r
c
i
s
ee 8
(
8
.
4
d
)andwouldber
e
g
a
r
d
e
da
st
h
ea
d
j
u
s
t
e
doptimums
o
l
u
t
i
o
nt
ot
h
eo
r
i
g
i
n
a
lMP
p
r
o
b
l
e
m
. l
i
si
sa
l
s
ot
h
ecommonp
r
a
c
t
i
c
ei
ne
n
g
i
n
e
e
r
i
n
gd
e
s
i
g
no
p
t
i
m
i
z
a
t
i
o
n
.
Th
eb
e
s
ts
o
l
u
t
i
o
n
however

i
so
b
t
a
i
n
e
da
t

x
;
=1.7433; X;=1.5; x
;
=1.73; ]=0.1782

(
8
.
5
)

Thed
i
f
f
e
r
e
n
c
ebetweenE
q
u
a
t
i
o
n
s(
8
.
5
)and(
8.
4a
)
o
n
l
yi
nev
a
r
i
a
b
l
eX
I
s
u
g
g
e
s
t
s
ta
d
d
i
t
i
o
n
a
lc
o
n
t
i
n
u
o
u
so
p
t
i
m
i
z
a
t
i
o
nn
e
e
d
st
ob
ep
e
r
f
o
r
m
e
df
o
re
a
c
hs
e
to
fE
d
i
s
c
r
e
t
ev
a
l
u
e
ss
e
l
e
c
t
e
dc
o
n
f
i
r
m
i
n
gt
h
e r
l
i
e
ro
b
s
e
r
v
a
t
i
o
n
s by F
le
t
c
h
e
r
. u
s
o
p
t
i
m
i
z
a
t
i
o
ns
h
o
u
l
dbee
a
s
i
e
ra
st
h
eo
r
d
e
ro
ft
h
em
a
t
h
e
m
a
t
i
c
a
lmodelwi

lbe duced
byt
h
enumbero
fd
i
s
c
r
e
t
ev
a
r
i
a
b
l
e
ss
i
n
c
et
h
e
yh
a
v
ebeensignedn
u
m
e
r
i
c
a
lv
a
l
u
e
s
.
T
h
i
ss
i
m
p
l
eexample commendsatev
e
r
yl
e
a
s
tat
h
r
e
e
s
t
e
pprocedw
f
o
rp
r
o
b
l
e
m
s
i
n
c

uded
i
s
c
r
e
t
ev
a
r
i
a
b
l
e
s
:

S
t
e
p1
:Aconl
In
uousr
e
l
a
x
a
t
i
o
nt
h
a
ti
d
e
n
t
i
f
i
e
ss
e
v
e
r
a
ls
e
t
so
fd
i
s
c
r
e
t
ev
a
r
i
a
b
l
e
sf
o
r
o
n
.
f
u
r
t
h
e
re
x
p
l
o
r

324

OISCRETEOPTIMIZATION

S
t
e
p2
:F
o
re
a
c
hs
u
c
hs
e
to
fd
i
s
c
r
e
t
ev
a
r
i
a
b
l
ec
o
m
b
i
n
a
t
i
o
nac
o
n
t
i
n
u
o
u
s
o
p
t
i
m
i
z
a
t
i
o
ni
sp
e
r
f
o
r
m
e
d1
0e
s
t
a
b
l
i
s
hanewoptimumv
a
1ueo
ft
h
ec
o
n
t
i
n
u
o
u
s
v
a
r
i
a
b
l
e
sa
n
dt
h
ec
o
r
r
e
s
p
o
n
d
i
n
go
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
n
.I
fa
l
lo
fl
h
ev
a
r
i
a
b
l
e
sa
r
e
d
i
s
c
r
e
t
e

t
h
e
no
n
l
yt
h
ef
u
n
c
t
i
o
nandc
o
n
st
r
a
In
t
sn
e
e
dt
ob
ee
va
1u
a
t
e
da
te
a
c
ho
f
t
h
es
e
lo
fv
a
r
i
a
b
l
e
s
.

82 OISCRETEOPTIMIZATIONTECHNIQUES

StandardFormat-DlscreteO
p
t
l
m
/
z
a
t
l
o
n
:The a
u
t
h
o
ri
sn
o
t aWl
l
I
o
fa
s
t
a
n
d
a
r
df
o
r
m
a
tf
o
r d
i
s
c
r
e
t
eo
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
e
m
.I
ts
u
b
s
u
m
c
s f
o
r
m
a
to
fl
h
e
copondingc
o
n
t
i
n
u
o
u
sr
e
l
a
x
a
t
i
o
np
r
o
b
l
c
m
.I
nt
h
i
sbookI
h
cf
o
l
1
owingf
o
r
m
a
tf
ol
'
t
h
emixedo
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
e
mi
su
s
e
d
.
M
i
n
i
m
i
z
e

S
t
e
p3
:As
i
m
p
l
ecomp
i
s
o
no
ft
h
ea
b
o
v
es
o
l
u
t
i
o
n
s
/
va
1u
e
si
nS
t
e
p2t
oi
d
e
n
t
i
f
ye
optimums
o
l
u
t
i
o
nt
oed
i
s
c
r
e
t
ep
r
o
b
l
e
m
.
T
h
i
su
n
c
o
n
s
t
r
a
i
n
e
do
p
t
i
m
i
z
a
t
i
o
nexamplei
nt
h
r
e
ev
a
r
i
a
b
l
e
sh
a
sd
e
m
o
n
s
t
r
a
t
e
d
t
h
a
td
i
s
c
r
e
t
eo
p
t
i
m
i
z
a
t
i
o
nr
e
q
u
i
r
e
s al
o
to
f work compared t
oc
o
n
t
i
n
u
o
u
s
o
p
t
i
m
i
z
a
t
i
o
n
.T
h
i
sworke
x
p
a
n
d
ss
i
g
n
i
f
i
c
a
n
t
l
yi
ft
h
enumbero
fv
a
r
i
a
b
l
e
si
n
c
r
e
a
s
e
s
o
ri
ft
h
em
a
t
h
e
m
a
t
i
c
a
lmodeli
se
n
h
a
n
c
e
dt
h
r
o
u
g
ht
h
ei
n
c
l
u
s
i
o
no
fc
o
n
s
t
r
a
i
n
t
s
.
A
n
o
t
h
e
re
s
s
e
n
t
i
a
ll
f
r
ei
nt
h
eabovee
x
p
l
o
r
a
t
i
o
ni
s nonewm
a
t
h
e
m
a
t
i
ca
1
c
o
n
d
i
t
i
o
n
sweren
e
c
e
s
s
a
r
yf
o
rs
t
a
b
l
i
s
h
m
e
n
to
ft
h
ed
i
s
c
r
e
t
es
o
l
u
t
i
o
nbeyonda
s
i
m
p
l
ec
o
m
p
a
r
i
s
o
no
ft
h
eo
b
j
e
c
l
i
v
ef
u
n
c
t
i
o
n
.Then
a
t
u
r
eo
fd
i
s
c
r
e
t
ev
a
r
i
a
b
l
e
sa
n
d
d
i
s
c
r
e
l
e f
u
n
c
t
i
o
n
s p
r
e
c
l
u
d
c
s a
n
y s
o
p
h
i
s
t
i
c
a
t
c
d m
a
t
h
e
m
a
t
i
c
a
l c
o
n
d
i
t
i
o
n
s
e
s
t
a
b
l
i
s
h
e
dbyd
e
r
i
v
a
t
i
v
e
so
ft
h
ef
u
n
c
t
i
o
n
si
n
v
o
l
v
e
di
nt
b
emodel
.T
r
a
p
p
i
n
ga
n
d
b
r
a
n
c
h
i
n
gb
a
s
e
donc
o
m
p
a
r
i
s
o
no
fv
a
l
u
e
sa
r
et
h
em
a
i
n
s
t
a
yo
fd
i
s
c
r
e
t
ea
l
g
o
r
i
t
h
m
s
.
T
h
e
r
et
e
c
h
n
i
q
u
e
sa
r
ec
l
a
s
s
i
f
i
e
da
sh
e
u
r
i
s
t
i
cm
e
t
h
o
d
s
.T
h
i
se
n
c
o
u
r
a
g
e
su
n
i
q
u
ea
n
d
p
e
r
s
o
n
a
li
m
p
l
e
m
e
n
t
a
t
i
o
n
so
ft
h
es
e
a
r
c
hf
o
rd
i
s
c
r
e
t
eo
p
t
i
m
i
z
a
t
i
o
nt
h
a
tc
a
nh
t
a
i
l
o
r
e
df
o
rac
l
a
s
so
fp
r
o
b
l
e
m
s
.

8
.
2 DISCRETEOP
T
lMIZATIONTECHNIQUES
Therea
r
et
h
r
e
ed
i
s
c
r
e
t
eo
p
t
i
m
i
z
a
t
i
o
nt
e
c
h
n
i
q
u
e
si
nt
h
i
ss
e
c
t
i
o
n
.Thef
i
r
s
tonei
s
E
x
h
a
u
s
t
i
v
eE
n
u
m
e
r
a
t
i
o
n
.T
h
i
si
n
v
o
l
v
e
si
d
e
n
t
i
f
y
i
n
gt
b
es
o
l
u
t
i
o
nt
o e
m
a
t
h
e
m
a
t
i
c
a
lmodelf
o
ra
l
lp
o
s
s
i
b
l
ec
o
m
b
i
n
a
t
i
o
n
so
ft
h
ed
i
s
c
r
e
t
ev
a
r
i
a
b
l
e
s
.T
h
i
s
i
ss
u
g
g
e
s
t
i
v
eo
ft
h
ez
e
r
o
o
r
d
e
rn
u
m
e
r
i
c
a
lt
e
c
h
n
i
q
u
e
so
fC
h
a
p
t
e
r6
. Those
methodsl
a
c
k
c
ds
o
p
h
i
s
t
i
c
a
t
i
o
na
st
h
e
yi
n
v
o
l
v
e
do
n
l
ye
v
a
l
u
a
t
i
o
no
ft
h
ef
u
n
c
t
i
o
n
s
a
taphenomenalnumbero
fp
o
i
n
t
s
.Theywerea
b
l
et
ot
a
k
ea
d
v
a
n
t
a
g
eo
ft
h
p
l
e
n
t
i
f
u
l computer r
e
s
o
u
r
c
e
sa
v
a
i
l
a
b
l
et
o
d
a
y
.I
m
a
g
i
n
eo
p
e
r
a
t
i
n
gi
nt
h
e
p
e
e
r
t
o
p
e
e
rcomputingenvironmn
ta
f
f
o
r
d
e
dby
MP3"o
r
G
n
u
t
e
l
l
a
"and
s
o
l
v
i
n
gt
h
eproblemu
s
i
n
ga
l
lo
ft
h
ePCsi
nt
h
ew
o
r
l
d
.Thes
e
c
o
n
dmethodi
st
h
e
BranchandBoundm
e
t
h
o
d
.T
h
i
si
sb
a
s
e
donp
a
r
t
i
a
le
n
u
m
e
r
a
t
i
o
nwhereo
n
l
yp
a
r
t
o
ft
h
ec
o
m
b
i
n
a
t
i
o
n
sa
r
ee
x
p
l
o
r
e
d
.Ther
e
m
a
i
n
i
n
ga
r
ep
r
u
n
e
dfromc
o
n
s
i
d
e
r
a
t
i
o
n
b
e
c
a
u
s
et
h
e
ywi
1
1n
o
td
e
t
e
r
m
i
n
et
h
es
o
l
u
t
i
o
n
.T
h
i
si
sc
u
r
r
e
n
t
l
yt
h
emostp
o
p
u
l
a
r
methodf
o
rd
i
s
c
r
e
t
eo
p
t
i
m
i
z
a
t
i
o
nf
o
re
n
g
i
n
e
e
r
i
n
gd
e
s
i
g
n
.Thel
a
s
tmethodi
s
Dynamicprogrammingane
l
e
g
a
n
ta
p
p
r
o
a
c
ht
oo
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
e
m
sb
u
t
whichd
i
dn
o
tg
a
i
nf
a
v
o
rb
e
c
a
u
s
ei
ti
n
v
o
l
v
e
ds
i
g
n
i
f
i
c
a
n
t
l
yl
a
r
g
e
ramountso
f
c
o
m
p
u
t
a
t
i
o
nt
h
a
nc
o
m
p
e
t
i
t
i
v
emethodse
v
e
nf
o
rp
r
o
b
l
e
m
so
fr
e
a
s
o
n
a
b
l
es
i
z
e
.I
t
i
sr
e
s
t
r
i
c
t
e
dt
op
r
o
b
l
e
m
st
h
a
lr
e
q
u
i
r
e as
e
q
u
e
n
t
i
a
ls
e
l
e
c
t
i
o
no
ft
h
ed
e
s
i
g
n
v
a
r
i
a
b
l
e
s
.Todays
u
c
hr
e
s
o
u
r
c
I
im
i
t
a
t
i
o
n
sa
r
ed
i
s
a
p
p
e
a
r
i
n
gi
nt
h
ew
o
r
l
do
f
p
o
w
e
r
f
u
lPCsa
n
dh
e
n
c
el
h
emethodd
e
s
e
r
v
c
sl
obcr
e
v
i
s
i
t
e
d
.

325

f
(
X
Y)]
n
<
;[
Y
]n
d

S
u
b
j
c
c
tt
o
: h(X
Y)=
[
0
]
; [
h
J
I

(
8
.
6
)

(
8
.
7
)

g(X
Y) [
0
]
;[
g
]
m

(
8
.
8
)

J
;SXjS.
i
;
'
; ;
=12..
.n
c

(
8
.
9
)

Y
iE Yd
;
[Yd
]
p
;
i 12
...

nd

(
8
.
1
0
)

;
;
:
;
11

Xr
e
p
r
e
s
e
n
t
sl
h
es
e
to
fncc
o
n
t
i
n
u
o
u
sv
a
r
i
a
b
l
e
s
.Y r
e
p
s
e
n
l
st
h
es
e
to
fndd
i
s
c
r
c
t
e
v
a
r
i
a
b
l
e
s
.fi
seo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
n
.hi
sl
h
es
e
to
f1e
q
u
a
l
i
l
yc
o
n
s
t
r
a
In
t
s
.gi
sl
h
es
e
l
n
e
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
t
s
.E
x
p
r
e
s
s
i
o
n(
8
.
9
)r
e
p
:
s
e
n
t
ses
i
d
ec
o
n
s i
n
t
sone
a
c
h
o
fm i
c
o
n
t
i
n
u
o
u
sv
a
r
i
a
b
l
e
.E
x
p
r
e
s
s
i
o
n(
8
.
1
0
)e
x
p
r
e
s
s
e
ses
i
d
ec
o
n
s
t
r
a
i
n
t
sont
h
ed
i
s
c
r
e
t
e
v
a
r
i
a
b
l
e
s
.Eachd
i
s
c
r
e
l
ev
a
r
i
a
b
l
eY
im
u
s
tb
e
l
o
n
gt
oap
r
e
e
s
t
a
b
l
i
s
h
e
ds
e
to
fP
id
i
s
c
r
e
l
e
'I
fnd=0ni
li
sac
o
n
t
i
n
u
o
u
so
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
e
m
.I
f11c=0cni
ti
s
v
a
l
u
e
sYt/I
m
.I
fb
o
t
ha
r
ep s
e
n
t

t
h
e
ni
ti
saMixedP
r
o
b
l
e
m
.
ad
i
s
c
r
e
t
eo
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l

ContlnuousR
e
l
a
x
a
t
l
o
n
:Thec
o
n
t
i
n
u
o
u
sr
e
l
a
x
o
no
fl
h
em
i
x
e
do
p
t
i
m
i
z
a
t
i
o
n
p
r
o
b
l
c
m(
8
.
6
(
8
.
1
0
)i
si
d
e
n
t
i
c
a
11
0p
r
o
b
l
c
m(
8 (
8
.
1
0
)w
i
t
hYa
sac
o
n
t
i
n
u
o
u
sv
a
r
i
a
b
l
e
t.l
Cd
i
s
c
r
e
n
s
n
l(
8
.
1
0
)i
sr
e
p
l
a
c
b
yac
o
n
t
i
n
u
o
u
ss
i
d
ec
o
n
s
i
n
tf
o
re
d
i
s
c
r
e
t
ev
a
r
i
a
b
l
e
so
ft
h
ef
o
r
mo
f(
8
.
9
)
. i
sn
o
te
x
p
l
i
c
i
t
l
yd
e
v
e
l
o
p
e
ds
i
n
c
e(
8
.
6
)(
8
.
9
)
a
n
dam
o
d
i
f
i
e
d(
8.
10
)i
sf
a
i
r
l
yr
e
p
r
e
s
e
n
t
a
t
i
v
eo
fec
o
n
t
i
n
u
o
u
sr
e
l
a
x
e
dm
a
t
h
e
m
a
t
i
c
a
l
.
l
mode
ReducedModel:The ducedmodeli
si
m
p
o
r
t
a
n
ti
ns
u
b
s
e
q
u
e
n
td
i
s
c
u
s
s
i
o
n
s
.I
ti
s
em
odels
o
l
v
e
da
f
t
e
rap
a
r
to
ft
h
es
e
to
fd
i
s
c
r
e
t
ev
a
r
i
a
b
l
e
sa
r
es
e
ta
tsomea

lo
w
a
b
l
e
v
a
1u
e
s
.T
h
i
sr
e
m
o
v
e
sl
h
o
s
ev
a
r
i
a
b
l
e
sf
r
o
mep
r
o
b
l
e
ma
si
rv
a
l
u
sh
a
v
e
d
i
s
c
r
e
t

b
e
e
ns
e
t
.T
hem
a
t
h
e
m
a
t
i
c
a
lmodeli
st
h
e
nd
e
f
i
n
e
dbyt
h
er
e
m
a
i
n
i
n
gd
i
s
c
r
e
t
ev
a
r
i
a
b
l
e
s
)a
we
l eo
r
i
g
i
n
a
lc
o
n
l
i
n
u
o
u
sv
a
r
i
a
b
l
e
s
.I
fZ(nz
r
e er
e
m
a
i
n
i
n
gd
i
s
c
r
e
t
e
e
e
d1
0b
es
o
l
v
e
d
e
n
v
a
r
i
a
b
l
e
s n

M
i
n
i
m
i
z
e l(X
Z
) [
X
]
"
;
[
Z
]n

(
8
.
1
1
)

Su
e
c
t1
0
:
(X
Z
) [
0
]
; [
1
1
]
/

(
8.
12
)

g(X
z
)~ [
0
]
; [
i
]
m

(
8
.
1
3
)

J
;SXjSx
j
'
; i
=12...nc

(
8
.
1
4
)

Z
iez
.
t

(
8
.
1
5
)

[
z
.
t
)
:

326 D
T
l
ON
I
S
C
R
E
T
EO
P
T
I
M
l
Z
A

8
.
2O
I
S
C
R
E
T
EO
P
T
I
M
I
Z
A
T
I
O
NT
E
C
H
N
I
Q
U
E
S

T
o
b
l
e8
.
1
0 OpmalV
o
l
u
eofCoououousV
o
r
i
a
b
l
e
x
.f
o
rD
i
s
c
r
e
t
eCombloauoo

fXloodx3
x
X
21
)

0
.
2
2
0
0

0
.
7
5
0
0

1
.7
300

2
.
2

2
.
7
8

0
.
5
1
.
5
2
.
5
3
.
5

0
.
9
0
6
7
1
.
2
O
1
.5
7
3
3
1
.
9
0
6
7

1
.
0
8
3
3
1
.
4
1
6
7
1
.7
5
0
0

1
.
4100
1
.
7
4
3
3

1
.5
8
0
0
1
.9
1
3
3

1
.7
6

2
.
0
7
6
7
2.
41
0
0

2
.
2
4
6
7
2
.
5
8
0
0

2
.
0
8
3
3

327

zaodX
3
T
o
b
l
e8
.
1
b OpumolValueo
fO
b
j
e
c
u
v
eFuocuoof
o
rD
l
s
c
r
e
l
eC
o
m
b
i
o
o
t
i
o
no
fX

2
.
0
9
3
3
2
.
4
2
6
7
2
.
7
6
0
0

nt
h
eabove
4
i
sn
o
tanews
e
.
tI
tc
o
pondst
oar
e
d
u
c
e
dYd
.In
Note ti
g
e
n
e
r
a
l
l
yi
ti
s emptys
e
.
tA
I
Io
ft
h
ed
i
s
c
r
e
t
ev
a
r
i
a
b
l
e
s
E
x
h
a
u
s
t
i
v
eE
n
u
m
e
r
a
t
i
o
n
s
e
tt
osomea
l
l
o
w
a
b
l
ev
a
l
u
e
.
8
.
2
.
1 ExhaustiveEnumeration
T
h
i
si
sesimpl
s
to
fed
i
s
c
r
e
t
eo
p
t
i
m
i
z
a
t
i
o
nt
e
c
h
n
i
q
u
S
.I
te
v
a
l
u
a
t
e
sanoptimum
e
s
ts
o
l
u
t
i
o
ni
so
b
t
a
i
n
e
d
s
o
l
u
t
i
o
nf
o
ra
l
lc
o
m
b
i
n
a
t
i
o
n
soft
h
ed
i
s
c
r
e
t
ev
a
r
i
a
b
l
e
s
.Theb
n
v
e
s
t
i
g
a
t
i
o
nf
o
rt
h
eminimum
bys
c
a
n
n
i
n
gt
h
el
i
s
to
ff
e
a
s
i
b
l
es
o
l
u
t
i
o
n
si
neabovei
v
a
l
u
e
.Thet
o
t
a
lnumbero
fe
v
a
l
u
a
t
i
o
n
si
s

x/
X
3

0
.
2
2
0
0

0
.
7
5
0
0

1
.7
300

2
.
2
4
0
0

2
.
7
8
0
0

5
0.
1
.5
5
2.
3
.
5

1
.
9
1
0
7
3
.
3
2
4
0
8
.
0
7
0
7
1
6
.
1
5
0
7

1
.3
5
4
2
1
.3
5
4
2

2
.
7
9
1
5
0
.
1
7
8
2
0
.
8
9
8
2
4
.
9
5
1
5

4
.
8
0
6
0
0
.
8
3
2
7
0
.
1
9
2
7
2
.
8
8
6
0

7
.
8
8
4
0
2.
4
7
0
7
0.
39
0
7
1
.
6440

4
.
6
8
7
5
1
1
.3
5
4
2

model(
3
)i
sr
e
d
u
c
e
dbyenumberofd
i
s
c
r
e
t
ev
a
r
i
a
b
l
e
s(
2
)
.Modelr
e
d
u
c
t
i
o
ni
s
i
n
v
o
l
v
e
di
ne
n
u
m
e
r
a
t
i
o
nt
e
c
h
n
i
q
u
e
s
.

A
l
g
o
r
i
t
h
m
:E
.
'
a
u
s
t
l
v
eEnumeration(
A
8
.
1
)
S
t
e
p1
.!
=i
n
f

X =[
0
0
.
.
0
]
)
1lowablecombinationof(
Y
.
.Y
2

Y
"
d
)=
(
Y
Fore
v
e
r
ya
b
S
o
l
v
eO
p
t
i
m
i
z
a
t
i
o
nProblem(
8
.
1
1
)
(
8
.
1
5
)(
S
o
l
u
t
i
o
nX
)

Ifh(XYb) [
0
]and
0
]and
Ifg(X.Yb}[
I
f
j(XYb)<!
Thenf
f
t
.
XY
b}

I
I
p
/

/
1
.

XX
.

(
8.
16
)

YYb
EndI
f

I
fe
i
t
h
r
n
do
rP
i(
o
rb
o
t
h
)i
sl
a
r
g
en i
sr
e
p
r
e
s
e
n
t
sal
o
tofw
o
r
k
.l
ta
l
s
or
e
p
r
e
s
e
n
t
s
c
a
l
c
u
l
a
t
i
o
n
sw
i
t
h numbero
fd
i
s
c
r
e
t
ev
a
r
i
a
b
l
e
s
.I
na
ane
x
p
o
n
e
n
t
i
a
lgrowthi
nt
h
z
a
t
i
o
np
r
o
b
l
e
mi
swouldi
n
v
o
l
v
en
cc
o
n
t
i
n
u
o
u
soptimums
o
l
u
t
i
o
no
f
mixedop
.I
ft
h
em
a
t
h
e
m
a
t
i
c
a
lmodelandi
t
smputerc
a
l
c
u
l
a
t
i
o
n
sa
r
e sy
t
h
er
e
d
u
c
e
dmodel

tt
h
i
si
sn
o
ta s
e
r
i
o
u
sb
u
r
d
e
n
.I
f

mathematicalmodel q
u
i
r
e
s
t
o implemen
o
r ex
n
p
l
e n
i
t
e e
l
e
m
e
n
t o
r f
i
n
i
t
e d
i
f
f
e
r
e
n
c
e
e
x
t
e
n
s
i
v
e c
a
l
c
u
l
a
t
i
o
n
s f
l
i
m
i
t
i
n
gt
h
ee
l
e
m
e
n
t
so
ft
h
e
c
a
l
c
u
l
a
t
i
o
n
s
t
h
e
nsomeconcemmaya
r
i
s
e
.I
ns
u
c
hc
a
s
e
s
t
ev
a
r
i
a
b
l
e
st
os
m
a
l
l
e
rc
1u
s
t
e
r
smayr
e
p
s
e
n
tau
s
e
f
u
la
p
p
r
o
a
c
h
.
d
i
s
c
r

Exmn
p
l
e8
.
2 TheA
e
r
o
d
e
s
i
g
nClubh
a
si
n
i
t
i
a
t
e
df
u
n
d
r
a
i
s
i
n
ga
c
t
i
v
i
t
i

b
ys
e
l
l
i
n
g

Ex
ample8.1:beexhaustiveenumer
t
i
o
no
fE
x
ample8
.
1i
sa
v
a
i
l
a
b
l
ei
nEx
8_1.m.1
T
a
b
l
e8
.
1ashows
optimumvalueoft
h
ec
o
n
t
i
n
u
o
u
sv
a
r
i
a
b
l
exIf
o
re
v
e
r
ya
1
1
0
w
a
b
l
e

t
at
oi
d
e
n
t
i
f
yam
a
t
h
e
m
a
t
i
c
a
lmodelt
h
a
tw
i
l
lr
u
ce e
i
rc
o
s
twhichd
i
r
e
c
t
l
y
t
r
s
l
a
t
oh
i
g
h
e
rp
r
o
f
i
t
s
.Onlytwot
y
p
e
so
fp
i
z
z
awi
1
Isell-pepperoni(
x
.
) d

i
s
c
r
e
t
ev
a
r
i
a
b
l
e
sX
2andx)
T
a
b
l
e8
.
1bshows c
o
r
r
e
s
p
o
n
d
i
n
g
c
o
m
b
i
n
a
t
i
o
nof d
si
sanu
n
c
o
n
s
t
r
a
i
n
e
d
v
a
l
u
eoft
h
eo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
nf
o
reachoft
h
ec
a
s
e
s
.S
i
n
c
ei
i
s
c
u
s
s
i
o
no
ff
e
a
s
i
b
i
l
i
t
yi
si
n
v
o
l
v
e
d
.Therea
r
e20v
a
l
u
e
sandt
h
eb
e
s
t
problemnod
p
l
e8
.
1i
s
s
o
l
u
t
i
o
nf
o
rEx n

x~ =
l
.i
4
3
3
;

X
;=1.5; ;
x=1.73; r=0.1782

f
EndI
EndFor

I
Iu
s
t
r
a
t
e
s u
s
eo
fa
l
g
o
r
i
t
h
mA
8
.
1w
i
t
hf
e
a
s
i
b
i
l
i
t
yr
e
q
u
i
r
e
m
e
n
t
s
.
Example8
.
2i

seyhavec
o
l
l
e
c
t
e
d
p
i
z
z
ad
u
r
i
n
gl
u
n
c
hi
nt
h
es
t
u
d
e
n
ta
c
t
i
v
i
t
yc
e
n
t
e
r
.Overt
h
ey r

1
Id
e
l
i
v
e
ramaximumof26p
i
z
z
.F
a
c
t
o
r
i
n
gi
n
t
o
c
h
e
e
s
e(
x
2
1
.Thel
o
c
a
lp
i
z
z
ac
h
a
i
nwi
a
c
c
o
u
n
tw
a
s
t
a
g
eandt
h
et
i
m
ea
v
a
i
l
a
b
l
et
omaket
h
es
a
l
eeya
r
r
i
v
ea
tt
h
ef
o
l
1owing
m
a
t
h
e
m
a
t
i
c
a
lm
o
d
e
l
:
M
i
n
i
r
n
i
z
e

(
8
.
5
)

each oft
h
e 20 c
a
s
e
si
se
s
s
e
n
t
i
a
l
l
ya
There i
s some good news t
h
o
u
g
h
.F
i
r
s
t
s
;
n
g
l
e
-l
I
a
r
i
a
b
l
eopz
a
t
i
o
np
r
o
b
l
e
m
.Th
enumbero
fd
e
s
i
g
nv
a
r
i
a
b
l
e
si
n
o
r
i
g
i
n
a
l
ed
o
w
l
l
l
o
a
d
e
df
r
o
mI
h
ew
e
b a
i
n
d
i
c
a
l
c
db
yb
o
l
d
f
u
c
e SS i
f
l
y
p
e
.
I
F
i
le
s10b

EndI
f

f
(
x
.
.X
2
)=
(
X
I-1
5
)
2+(
X
2- 1
5
)
2

2
)
: X
I+X
2S26
S
u
b
j
e
c
tt
o
: 8
.
(
X
hX

g
2
(
X
.x
2
1
: x
.+2
.
5
x
2S37

(
8.
17
)
(
8
.
1
8
a
)
(
8
.I
8
b
)

328 D
I
S
C
R
E
T
EO
P
T
I
MlZAnON

8
.
2D
I
S
C
R
E
T
EO
P
T
I
M
I
Z
A
T
I
O
NT
E
C
H
N
I
Q
U
E
S 329
X
I~ 0
;

x2~

O
.andi
n
t
e
g
e
r

(
8
.
1
8
c
)

Theg
r
a
p
h
i
c
a
ls
o
l
u
t
i
o
nf
o
rc
o
n
t
i
n
u
o
u
sr
e
l
a
x
a
l
i
o
ni
sa
v
a
i
l
a
b
l
ei
nF
i
g
u
r
e8
.
2andt
h
e
a
n
a
l
y
t
i
c
a
lS
o
l
u
l
i
o
n(
n
o
ti
n
c
l
u
d
e
d
)i
d
e
n
t
i
f
i
e
st
h
a
t1
2
.
8
6p
e
p
p
e
r
o
n
ip
i
z
z
a
sa
n
d9
.
6
5
c
h
e
e
s
ep
i
z
z
a
sw
i
l
lp
r
o
v
i
d
ef
o
rt
h
eminirnumc
o
s
to
fo
p
e
r
a
t
i
o
n
.
f
i
l
ea
tw
i
l
lp
e
r
f
o
r
mt
h
ee
x
h
a
u
s
t
i
v
ee
n
u
m
e
r
a
t
i
o
no
ft
h
e
Ex8_2.m i
sem
p
r
o
b
l
e
m
.Thed
i
s
c
r
e
t
es
o
l
u
t
i
o
ni
s

X
;=1
2
;

x
i=10; r
=3
4
;

g
2i
sa
c
t
i
v
e

(
8.
19)

T
h
i
sc
o
n
f
i
r
m
s tn
e
i
g
h
b
o
r
i
n
gd
i
s
c
r
e
t
ev
a
l
u
e
sa
r
o
u
n
dad
i
s
c
r
e
l
es
o
l
u
t
i
o
nworksw
e
l
l
.
Programminge
x
h
a
u
s
l
i
v
ee
n
u
m
e
r
a
t
i
o
ni
ss
t
r
a
i
g
h
t
f
o
r
w
a
r
dandp
a
r
t
i
c
u
l
a
r
l
ye
a
s
yw
i
t
h
MATLAB.G
iven p
r
o
c
e
s
s
i
n
gs
p
e
e
d
l
a
r
g
ememorya
v
a
i
l
a
b
i
l
i
t
yon e
d
e
s
k
t
o
p
. s
y
programmingt
h
r
o
u
g
hs
o
f
t
w
a
r
el
i
k
eMA
T
I
.
.
A
B
.e
t
i
v
ee
n
u
m
e
r
a
t
i
o
ni
sp
r
o
b
a
b
l
ya
v
e
r
ygoodi
d
e
al
o
d
a
y
.Whalrecommendsi
te
v
e
nf
u
r
t
h
e
ri
s t
h
es
o
l
u
t
i
o
ni
sag
l
o
b
a
l
t
i
m
u
m
.Fl
Ir
t
h
e
r
m
o
r
e
.u
n
l
i
k
et
h
ee
a
r
l
i
e
rc
h
a
p
t
r
swheret
h
c
o
n
c
e
p
t
so
fd
e
r
i
v
a
t
i
v
e
s
.
1t
ou
n
d
e
r
s
t
a
n
dandi
m
p
l
e
m
e
n
t
m
a
t
r
i
c
e
s
.l
i
n
e
a
ri
n
d
e
p
e
n
d
e
n
c
e
.a
n
ds
oonweree
s
s
e
n
t
ia
t
h
et
e
c
h
n
i
q
u
e
so
fo
p
t
i
m
i
z
a
t
i
o
n
l
a
c
ko
fs
u
c
bknowledgei
snod
i
s
a
d
v
a
n
g
et
os
o
l
v
i
n
g
e
r
ei
so
n
l
yone
ad
i
s
c
r
e
t
eo
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
e
mu
s
i
n
ge
x
b
a
u
s
t
i
v
ee
n
u
m
e
r
a
t
i
o
n
.Tb
Is
k
i
l
l
I
s
1
a
t
i
n
g m
a
t
h
e
m
a
t
i
c
a
lmodeli
n
t
oprogramc
o
d
e
.WithE
e
s
s
e
n
t
ia
sb
o
o
k
.t
h
i
si
sd
e
f
i
n
i
t
e
l
yn
o
ta
ni
s
s
u
e
.
e
x
p
e
r
i
e
n
c
ep
r
o
v
i
d
e
di
ni
4
0
3
5

es
o
l
u
t
i
o
ni
nE
q
u
a
t
i
o
n(
8
.
1
9
)c
a
nbeo
b
t
a
i
n
e
dw
i
t
hc
o
n
t
i
n
u
o
u
sr
e
l onofE
p
r
o
b
l
e
m
.I
nt
h
i
sc
a
s
et
h
ed
i
s
c
r
e
t
es
o
l
u
t
i
o
nh
a
st
ober
e
c
o
v
e
r
e
dbya
d
j
u
s
t
i
n
gt
h
e
c
o
n
t
i
n
u
o
u
ss
o
l
u
t
i
o
nt
on
e
i
g
h
b
o
r
i
n
gd
i
s
c
r
e
t
ev
a
l
u
e
s
.

Ex
haus
t
#
veEnumeratlonf
o
rContlnuousO
p
t
l
m
l
z
a
t
l
o
nProblems:T
h
i
st
i
t
l
e
mustbequa
I
i
f
i
e
df
u
r
t
h
e
r
i
nl
i
e
uo
fg
r
a
d
i
e
n
t
b
a
s
dn
u
m
e
r
i
ca
It
e
c
h
n
i
q
u
e
s
.The
a
v
a
i
l
a
b
i
l
i
t
yo
fs
e
r
i
o
u
sd
e
s
k
t
o
pc
o
m
p
u
t
i
n
gp
o
w
e
r
.t
h
ep
r
o
m
i
s
eo
fg
l
o
b
a
lo
p
t
i
m
i
z
a
t
i
o
n
.
a
n
d e
a
s
eo
fi
m
p
l
e
m
e
n
t
a
t
i
o
ns
u
g
g
e
s
lam
u
l
t
i
s
t
a
g
ee
x
h
a
u
s
t
i
v
ee
n
u
m
e
r
a
t
i
o
ns
t
r
a
t
e
g
y
f
o
rc
o
n
t
i
n
u
o
u
so
p
t
i
m
i
t
i
o
np
r
o
b
l
e
m
s
.E
x
h
a
u
s
t
i
v
ee
n
u
m
e
r
a
t
i
o
na
f
t
e
ra
l
li
sas
e
a
r
c
h
i
m
i
l
a
r1
0t
h
eo
n
e
su
s
e
db
e
f
o
r
e
.Th
es
e
a
r
c
hf
o
rs
o
l
u
t
i
o
ni
sa
k
i
n1
0e
s
t
r
a
t
e
g
y
.s
z
e
r
o
o
r
d
e
rm
e
l
h
o
d
semployedf
o
ru
n
c
o
n
s
t
r
a
i
n
e
do
p
t
i
m
i
z
a
t
i
o
n(
C
h
a
p
t
e
r6
)
.
Theh
i
g
h
e
rs
t
a
g
e
si
nm
u
l
t
i
s
t
a
g
ee
n
u
m
e
r
a
t
i
o
ni
n
v
o
l
v
ee
n
u
m
e
r
a
t
i
n
gond
e
c
r
e
a
s
e
d
i
n
t
e
r
v
a
Isc
e
n
t
e
r
e
da
r
o
u
n
des
o
l
u
t
i
o
nf
r
o
m p
r
e
v
i
o
u
ss
t
a
g
e
.Theva
I
u
e
so
ft
h
e
d
e
s
i
g
nv
a
r
i
a
b
l
e
su
s
e
df
o
re
n
u
m
e
r
a
t
i
o
nw
i
l
li
n
c
o
r
p
o
r
a
t
ep
r
o
g
r
e
s
s
i
v
e
l
yi
m
p
r
o
v
e
d
t
o
l
e
r
a
n
c
e
.Thes
t
r
a
t
e
g
ymustt
a
k
ei
n
t
oa
c
c
o
u
n
tt
h
ep
o
s
s
i
b
i
l
i
t
yt
h
a
tt
h
ei
n
i
t
ia
Ii
n
t
e
r
v
a
l
m
i
g
h
lo
v
e
r
l
o
o
kg
l
o
b
a
Io
p
t
i
m
a
ls
o
l
u
t
i
o
n
st
l
l
a
tdon
o
ta
p
p
e
a
ra
t
t
r
a
c
t
i
v
ei
nap
a
r
t
i
c
u
l
a
r
s
t
a
g
e
.H
e
n
c
e
.moret
h
a
noneminimumi
st
r
a
c
k
e
di
ne
a
c
hs
t
a
g
e
.Eacho
ft
h
e
s
e
minimumc
a
nb
ec
o
n
s
i
d
e
r
e
d e
n
l
l
m
e
r
a
t
i
o
nc
l
u
s
t
e
r
.Ea
c
hs
u
c
hc
1u
s
t
e
ri
sm
u
l
t
i
s
t
a
g
e
d
u
n
t
i
lap
r
e
s
c
r
i
b
e
dt
o
l
e
r
a
n
c
ei
sa
c
h
i
e
v
e
d
.I
ne
n
g
i
n
r
i
n
g
.
m
u
f
a
c
t
u
r
i
n
gt
o
l
e
r
a
n
c
eo
ft
h
e
sl
i
m
i
tp
r
o
b
a
b
l
yn
e
e
dt
obe
o
r
d
e
ro
f0
.
0
0
0
1i
sa
c
h
i
e
v
a
b
l
et
o
d
a
y
.T
o
l
e
r
a
n
c
e
sbelow i
r
e
ai
sm
i
c
r
o
e
l
e
c
t
r
o
n
i
c
s
.Fromaprogrammingp
e
r
s
p
e
c
t
i
v
ei
tj
l
l
s
t
j
u
s
t
i
f
i
e
du
n
l
e
s
s ea
a
d
d
sa
ne
x
t
r
al
o
o
pt
ot
h
ei
m
p
l
e
m
e
n
t
a
1
i
o
n
. l
ef
i
n
a
ls
o
l
u
t
i
o
no
b
t
a
i
n
e
dr
e
p
r
e
s
e
n
t
sa
c
o
n
t
i
n
u
o
u
ss
o
l
u
t
i
o
nw
i
t
h
i
nt
h
ef
i
n
a
ll
o
l
e
r
a
n
c
ei
m
p
l
e
m
e
n
t
e
dbyt
h
et
e
c
h
n
i
q
u
e
.
Sucha
ne
x
e
r
c
i
s
ei
sl
e
f
l
1
0t
h
es
t
u
d
e
n.
tAsw
i
t
ha
l
li
m
p
l
e
m
e
n
t
a
t
i
o
nt
h
em
e
r
i
to
fl
h
i
s
a
p
p
r
o
a
c
hl
i
e
si
nt
h
et
r
a
c
t
a
b
i
l
i
1
yo
ft
h
em
a
t
h
e
m
a
t
i
c
a
lmodela
n
dt
h
en
e
e
df
o
rag
l
o
b
a
l
optimums
o
l
u
t
i
o
n
.S
i
n
c
et
h
emethodi
sl
a
r
g
e
l
yh
e
u
r
i
s
t
i
c

t
h
e
r
ei
snog
u
a
r
a
n
t
e
et
h
a
tt
h
e
s
o
l
u
t
i
o
ni
sa
c
t
u
a
I
l
yag
l
o
b
a
loptimum(
o
fc
o
u
r
s
ea
s
s
u
m
i
n
gt
h
ef
u
n
c
t
i
o
n
si
n
v
o
l
v
e
da
r
e
n
o
tc
o
n
v
e
x
)
.
8
.
2
.
2 BranchandBound

1
0

.
.
.
g~=O

6ie-7

1
5
2
0
2
5
p
e
p
p
e
r
o
n
lp
l
z
z
a

3
0

F
l
g
u
r
e
8
.
2 G p
h
i
c
a
ls
o
l
u
t
l
o
nExa
m
p
l
e8
.
2
.

3
5

4
0

E
x
h
a
u
s
t
i
v
eo
rc
o
m
p
l
e
t
ee
n
u
m
e
r
a
t
i
o
ni
sp
o
s
s
i
b
l
eo
n
l
yf
o
ral
i
m
i
t
e
ds
e
to
fv
a
r
i
a
b
l
e
s
b
e
c
a
u
s
eo
ft
h
ee
x
p
o
n
e
n
t
ia
Ig
r
o
w
t
ho
fe
f
f
o
r
t q
u
i
dt
oexaminea
Il
p
o
s
s
i
b
l
es
o
l
u
t
i
o
n
s
.
a
r
t
i
a
l(
o
ri
n
c
o
m
p
l
e
t
eo
rs
e
l
e
c
t
i
v
e
)e
n
u
m
e
r
a
l
i
o
nt
oa
r
r
i
v
ea
lt
h
es
o
l
u
t
i
o
nt
ot
h
e
U
s
i
n
gp
d
i
s
c
r
e
l
eo
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
e
mwouldp
r
o
v
i
d
et
l
t
n
e
c
e
s
s
a
r
ya
d
v
a
n
t
a
g
ei
nh
a
n
d
l
i
n
g
l
a
r
g
ep
r
o
b
l
e
m
s
.P
a
r
t
i
a
le
n
u
m
e
r
a
t
i
o
nworksbymakingd
e
c
i
s
i
o
n
sa
b
o
u
tg
r
o
u
p
so
f
s
o
l
u
t
i
o
n
sr
a
t
h
e
r e
v
e
r
ys
i
n
g
l
o
n
e
.Th
e8
r
a
n
c
hand80und(88)a
l
g
o
r
i
t
h
mi
sone
o
ft
h
es
u
c
c
e
s
s
I
a
l
g
o
r
it
l
1mst
ou
s
es
e
l
e
c
t
i
ve/
p
a
r
t
i
a
le
n
u
m
e
r
a
t
i
o
n
.l
td
o
e
ss
obyu
s
i
n
g
r
e
l
a
x
a
t
i
o
nm
o
d
e
l
si
n
s
t
e
a
do
ft
h
eo
r
i
g
i
n
a
lc
o
m
p
l
e
1
ed
i
s
c
r
e
l
em
a
1
h
e
m
a
t
i
ca
Imode
.
lThe
8Ba
l
g
o
r
i
t
h
mu
s
e
sat
r
e
es
t
r
u
c
t
u
r
e1
0r
e
p
r
e
s
e
n
tc
q
m
p
l
e
t
i
o
n
so
fp
a
r
t
i
a
ls
o
l
u
t
i
o
n
s
.The
'
e
e
s
u
c
t
u
r
eu
s
e
sn
o
d
e
sande
d
g
e
s
/
l
i
n
k
s1
0r
e
p
r
e
s
e
n
l
l
h
et
r
a
i
lo
fp
a
r
t
i
a
ls
o
l
u
t
i
o
n
s
.T
h
i
s
f
a
t
l
r
o
m
u
l
gep t
i
a
ls
o
l
u
t
i
o
n
.Thet
e
r
m
i
n
o
l
o
g
yi
sb
o
r
r
o
w
e
df
r
o
mo
p
e
r
a
t
i
o
n
i
sc
a
l
l
r
e
s
e
hl
i
t
e
r
a
t
u
r
e[
3
]andi
se
x
p
l
a
i
n
e
dbelowwi r
e
s
p t
oExample8.
1whichi
s
r
e
p
r
o
d
u
c
e
dh
e
r
ef
o
rc
o
n
v
e
n
i
e
n
c
e
:

8
.
2

3
3
0 D
I
S
C
R
E
T
EO
P
T
I
M
I
Z
A
T
I
O
N
M
i
n
i
m
i
z
e f(Xl
X
2

X
3
)=(
X
I-2
)
2+(
X
I-X
2
)
2+(
X
I-X3)2+(
X
2-X
)
)
2(
8
.
1
)
X
IER

(
8
.
2
a
)

X
2
[
0
.
5 1
.5 2.
5 3
.
5
]EX
2d

(
8
.
2
b
)

X
3E[
0
.
2
2 0
.
7
5

D
1S
C
R
E
T
EO
P
T
I
M
I
Z
A
T
I
O
NT
E
C
H
N
I
Q
U
E
S

BBa
l
g
o
r
i
t
h
ms
e
a
r
c
h
e
sa
1
0
n
gaIC
.l
ts
t
a
r
l
Sa
lt
h
er
o
o
tw
h
e
r
ea
1
1t
h
e
Root:Th

2
3
t
:
3
1
?
J
1
::
;
;
:
?
?
J
2
2
:
:
p
r
e
f
e
r
r
e
dt
oj
u
m
p
i
n
ga
c
r
o
s
sn
o
d
e
si
nesamee
r
a
r
c
h
y
.

l
.7
3 2
.
2
4 2
.
7
8
]
X3
d

(
8
.
2
c
)

P
a
r
t
l
a
lS
o
l
u
t
l
o
n
:As
o
l
u
t
i
o
ni
nw
h
i
c
hsomed
i
s
c
r
e
t
ed
e
s
i
g
nv
a
r
i
a
b
l
e
sa
r
ef
i
x
e
da
n
d
e
r
sa
r
e
j

'
e
eo
ru
n
d
e
t
e
n
n
i
n
e
d
.V
a
r
i
a
b
1
e
st
h
a
ta
r
en
o
tf
i
x
e
da
r
er
e
p
r
e
s
e
n
t
e
db
ye
o
s
y
m
b
o
1
:
"Forexample

x=:
u0.5

(
8
.
2
0
)

s
e
n
t
st
h
eo
p
t
i
m
i
z
a
t
i
o
np
r
o
b
1
e
mw
i X2=0
.
5a
n
dXIa
n
dX3t
ob
ed
e
t
e
r
m
i
n
e
d
.The
r
e
p
a
ls
o
l
u
t
i
o
n
sa
r

a
1
s
ot
e
r
m
e
dt
h
enodeso
ft
h
et
r
e
e
. l
ef
r
e
ev
a
r
i
a
b
l
e
sa
r
eu
s
u
a
l
l
y
p
a
s
o
l
v
e
du
s
i
n
gac
o
n
t
i
n
u
o
u
sr
e
l ono
ft
h
emode
.
l

Compler/ons(
0
'Pa
T
t
S
o
l
u
t
l
o
n
)
:E
n
c
hp
a
r
t
i
a
ls
o
l
u
t
i
o
no
rn
o
d
c
a
ng
i
vt
:r
i
st
:l
u
a
d
d
i
t
i
o
n
n
lp
a
r
t
i
a
1s
o
l
u
t
i
o
n
s
/
n
o
d
e
sd
i
r
e
ct
I
yu
n
d
e
ri
t
.T
h
i
si
sc
a
l
l
e
db
r
a
l
l
c
h
i
n
go
r
e
x
p
a
n
s
i
o
T
lo
fl
h
en
o
d
e
.I
nt
h
i
snews
e
to
fn
o
d
e
sa
n
o
t
h
e
ro
n
eo
f f
r
e
cd
i
s
c
r
e
l
c
v
a
r
i
a
b
l
c
si
sf
i
x
e
da
n
der
e
m
a
i
n
d
e
rs
l
i
l
lc
o
n
t
i
n
u
et
ob
ef
r
e
e
.F
o
re
x
a
m
p
l
eu
n
d
e
r
n
o
d
er
e
p
r
e
s
e
n
t
e
db
yt
h
ep
a
r
t
i
a
ls
o
l
u
t
i
o
ni
nEqu
a
t
i
o
n(
8
.
2
0
)f
i
v
ea
d
d
i
t
i
o
n
a
1n
o
d
e
sa
r
e
s
s
i
b
l
e
:

Incumbents
o
f
t
J
M
o
n
tThei
n
c
u
m
b
e
n
ts
o
l
u
t
i
o
ni
st
h
s
b
G
S
tf
e
a
s
i
b
l
es
o
l
u
t
i
o
nMownS
O
11
d
i
s
c
r
e
t
ev
a
r
i
a
b
l
e
sa
s
s
u
m
i
n
gd
i
s
c
r
e
t
ev
a
l
u
e
s
.Whenl
h
BB
s
e
a
r
c
hf
i
n
a
l
1
ys
t
o
p
s

f
a
rw
i
t
ha
E
h
a
l
i
n
C
U
R
E
b
B
E
E
ts
o
l
u
t
i
o
n

i
f
o
m
e
x
i
s
t
s

i
saz
i
o
b
a
lo
p
t
i
m
u
ms
o
i
t
H
i
o
n
ht
h
e
f
o
l
l
o
w
i
n
g
t
h
e
i
n
c
u
m
b
e
n
t
s
o
l
u
t
i
o
ni
sd
e
n
o
t
e
d
b
y
r
1
3
m
i
n
c
u
m
b
a
u
d
E
S
i
g
nv
m
a
b
l
e
s X
.
.
.

Ex
ample8
.
1- U
slngBranchandBound: H
e
r
eeBBs
e
a
r
c
hi
sa
p
p
l
i
e
d

E
x
a
m
p
l
e8.1.BBEX8JJnc
a
l
c
u
l
a
t
e
sa
l
l
o
ft
h
Ed
a
t
an
e
e
d
e
df
o
rg
e
n
E
r
a
I
i
n
g
Th
eBB i
t
s
e
1
fi
sd
e
v
e
l
o
p
e
di
nF
i
gu
s
8.
3a
n
d8.
4
.
S
t
a
r
t
:NodeO
.A
l
lo
f v
a
r
i
a
b
1
e
sa
r
ef
C
.Thes
o
l
u
t
i
o
nt
ot
h
ec
o
n
t
i
n
u
o
u
sr
e
l
a
x
a
t
i
o
n
i
v
ef
u
n
c
t
i
o
n(
s
o
l
v
e
da
sac
o
n
t
i
n
u
o
u
sp
r
o
b
l
e
m
)i
si
d
e
n
t
i
f
i
e
da
tt
h
e
o
f
t
h
eo
b
j t
n
o
d
e
.
s
1
4
:T
h
i
si
st
h
f
i
r
s
tt
i
e
ro
ft
h
eBBt
r
e

T
hed
i
s
c
r
e
t
ev
a
r
i
a
b
l
eX2i
sa
s
s
i
g
n
e
d
Nod
E
a
c
ho
ft
h
g
p
e
n
n
i
s
B
i
b
l
e
d
i
s
c
r
e
t
ev
a
l
u
e
sa
tt
h
en
u
d
c
si
nt
h
i
st
i
e
r
T
h
i
sE
Si
n
d
i
c
a
t
e
d
X=[f

f
J

X=
f
[0

5
.0
.
2
2
]
; X=
f
[0
.
5
0
.
7
5
]
; X=
f
[0

5
.1
.
7
3
]
;

X=
f
[0
.
5
2
.
2
4
1
; X=
f
[0

5
.2
.
7
8
1

T
h
e
s
cc
o
m
p
l
e
t
i
o
n
sb
r
a
n
c
ho
u
ta
n
dh
c
n
c
eat
r
e
e
l
i
k
cs
t
r
u
c
t
u
r
e
.T
h
e
r
ei
sah
i
e
r
a
r
c
h
i
c
a
l
a
c
hb
r
a
n
c
hr
e
p
r
e
s
e
n
t
s1
e
a
d
st
oat
i
e
ri
nw
h
i
c
hsomev
a
r
i
a
b
l
e
s
t
r
u
c
t
u
r
i
n
go
ft
h
en
o
d
e
s
.E
i
sh
e
l
da
ti
t
sp
r
e
s
c
r
i
b
e
dd
i
s
c
r
e
t
ev
a
l
u
e
.
FathomingtheP
a
r
t
i
a
lS
o
l
u
t
l
o
n
:E
x
p
a
n
d
i
n
gt
h
en
o
d
ea
l
lt
h
ewayt
i
l
lt
h
ee
n
di
s
t
e
r
m
e
dfatlwmingan
o
d
e
.D
u
r
i
n
gt
h
ef
a
t
h
o
m
i
n
go
p
e
r
a
t
i
o
ni
ti
sp
o
s
s
i
b
l
e1
0i
d
e
n
l
i
f
y
t
h
o
s
ep
a
r
t
i
a
ls
o
l
u
t
i
o
n
sa
tan
o
d
et
h
a
tdon
o
tn
e
e
dt
ob
ei
n
v
e
s
t
i
g
a
t
e
df
u
r
t
h
e
r
.O
n
l
yt
h
e
b
e
s
ts
o
l
u
t
i
o
na
ten
o
d
ei
sbranchedo
rc
o
m
p
l
e
t
e
d
.Tho
s
ep
a
r
t

i
a
ls
o

t
i

o
n

s
l
l
I
n

d
e

sa
t
w
i
l
ln
o
t
b
ei
n
v
e
s
t
i
g
a
t
e
df
u
r
t
h
e
ra
r
epruned
0
rt
e
n
n
i
n
a
t
e
d
i
nel
I

C
.

ActlveP
a
r
t
l
a
lS
o
l
u
t
i
o
n
:An
o
d
et
h
a
lh
a
sn
o
tb
e
e
nt
e
n
n
i
n
a
t
e
do
rp
r
u
n
e
da
n
dh
a
s

n
o
lb
c
e
nc
o
m
p
l
e
t
e
d
.
Edgesl
L
l
nks:T
h
e
s
'
e
n
o
d
e
t
on
o
d
ec
o
n
n
e
c
t
i
o
n
si
na.T
h
e
yi
d
e
n
t
i
f
yt
h
ev
a
l
u
e
_r 1.._..~.. .
.
n
_L

'
.
.
.
.
.
.
.:........J

3
3
1

X
*=[
1
.
7
4

5
.1
.
7
3
]
p=
0
.
1
7
8
p
*=0
.
1
7
8
l
n
c
u
m
b
e
n
ls
o
l
u
t
l
o
n
Xo
I
"
.=[
1
.
7
4

5
.1
.
7
3
J
F
l
a
u
r
e8
.
3B
r
a
n
c
ha
n
db
o
u
n
dE
x
a
m
p
l
e8
.
1
.

332

D
l
S
C
R
E
T
EOPnM
I
Z
A
T
I
O
N

8
.
2 D
l
S
C
R
E
T
EOPnM
I
Z
A
T
l
O
NT
E
C
H
N
I
Q
U
E
S 333
i
se
x
a
m
p
l
e1
5p
a
r
t
i
a
ls
o
l
u
t
i
o
n
sw
e
r
en
e
s
s
a
r
y
.Eachp
a
r
t
i
a
ls
o
l
u
t
i
o
nc
a
nb
e

u
m
e
r
a
t
i
o
n
sw
e
r
e
c
o
n
s
i
d
e
r
e
da
ne
n
u
m
e
r
a
t
i
o
n
.I
ne
x
h
a
u
s
t
i
v
ee
n
u
m
e
r
a
t
i
o
n20 n
r
e
q
u
i
r
e
df
o
rExample8
.
1
.T
h
e
r
ei
san
e
lg
a
i
no
f5e
n
u
m
e
r
a
l
i
o
n
s
.Th
BBs
e
a
r
c
h
s
t
h
enumbero
ft
o
t
a
le
n
u
m
e
r
a
t
i
o
n
sn
e
e
d
d1
0s
o
l
v
eep
r
o
b
l
e
m
.
r
e
d
u

A
l
g
o
r
l
t
h
m
:BranchandBound 8
.
2
)
:For even a s
m
a
l
ls
e
to
fd
i
s
c
r
e
t
e

X"=[
1
.
5
7

2
.
5
0
.
2
2
]
P=8.07
X"=[
1.
7
5

5
.0
.
7
5
1

zs

1
.
7
3
1
P=4.68
xe=I107
P=0
.
8
9
8

x
'
=
I
2
.
2
6

zaz.Ml
P=0
.
1
9
2

F
1g
u
r
e8
.
4B
r
a
n
c
ha
n
db
u
n
dExa
m
p
l
e8
.
1
f
a
t
h
o
m
i
n
gN
o
d
e3
.

BBt
r
e
es
t
r
u
c
l
u
r
ei
smostl
i
k
e
l
yad
i
f
f
i
c
u
l
te
n
t
e
r
p
r
i
s
e
.
v
a
r
i
a
b
l
e
sa
c
t
u
a
l
l
yd
r
a
w
i
n
gt
h
'

e
es
t
r
u
c
t
u
r
em
e
r
e
l
yi
d
e
n
l
i
f
i
e
st
h
es
e
a
r
c
hp
r
o
c
e
s
s
.Example8
.
1d
i
dn
o
th
a
v
e
The1
t c
h
o
o
s
i
n
gt
h
e
c
o
n
s
t
r
a
i
n
t
ss
of
e
a
s
i
b
i
l
i
t
ywasn
o
le
x
a
m
i
n
e
d
.Withc
o
n
s
t
r
a
i
n
t
sp
r
e
s
e
n
i
a
ls
o
l
u
t
i
o
nt
oc
o
m
p
l
e
t
eb
a
s
e
do
n
l
yonminimumv
a
l
u
eo
ft
h
eo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
ni
s
p t
n
o
taj
u
s
t
i
f
i
a
b
l
es
t
e
g
y
.I
nt
h
i
sc
a
s
et
h
eb
e
s
lf
e
a
s
i
b
l
es
o
l
u
t
i
o
ni
sec
a
n
d
i
d
a
t
ef
o
r
na
c
t
i
v
ef
e
a
s
i
b
l
ep
a
r
t
i
a
ls
o
l
u
t
i
o
n h
a
sab
e
t
t
e
rv
a
l
u
e
c
o
m
p
l
e
t
i
o
n
.A
l
s
oa
c
u
r
r
e
n
ti
n
c
u
m
b
e
n
ts
o
l
u
t
i
o
ni
sa
l
s
o ac
a
n
d
i
d
a
t
ef
o
rc
o
m
p
l
e
t
i
o
n
. Th
ef
o
l
l
o
w
i
n
g
a
l
g
o
r
i
t
h
mc
a
p

r
e
s e
s
s
e
n
c
eo
ft
h
eB
r
a
n
c
ha
n
dBounda
l
g
o
r
i
t
h
m
.

S
t
e
pO
.E
s
t
a
b
l
i
s
ht
h
er
o
o
tnodew
i
t
ha
l
ld
i
s
c
r
e
t
ev
i
a
b
l
e
s.
l
n
i
t
i
a
l
i
z
es
o
l
u
t
i
o
ni
n
d
e
xst
oO
.

nF
i
g
u
r
e8
.
3
.Th
er
e
m
a
i
n
i
n
gv
i
a
b
l
e
s B
ont
h
ee
d
g
e
sb
e
t
w
e
e
nt
h
en
o
d
e
si
f
r
I
c
o
n
t
i
n
u
o
u
s
.Theoptimums
o
l
u
t
i
o
na
tt
h
ev
a
r
i
o
u
sn
o
d
e
si
sd
i
s
p
l
a
y
e
donE
t
r
e
e
.Theminimums
o
l
u
t
i
o
ni
st
h
eonea
tw
i
l
lb
ef
a
t
h
o
m
e
da
tt
h
en
e
x
ts
t
e
p
.
H
e
r
eNodes2a
n
d3-ec
a
n
d
i
d
a
t
e
sf
o
re
x
p
a
n
d
i
n
gt
h
eb
r
a
n
c
h
e
s
.Nodes1a
n
d4
nI
h
ei
n
c
u
m
b
e
n
t
w
i
l
ln
o
tb
ef
a
t
h
o
m
e
d(
u
n
l
c
s
st
h
e
i
roptimumv
a
l
u
ei
sb
e
t
t
e
ra
yw
i
l
lb
et
e
r
m
i
n
a
t
e
do
rp
r
u
n
e
d eo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
nh
a
sa
s
o
l
u
t
i
o
n
) e
h
i
g
h
e
rv
a
l
u
c
.

l
ec
o
n
t
i
n
u
o
u
sr
e
l
a
x
a
t
i
o
ne
n
s
u
r
e
sal
o
w
e
rboundont
h
eoptimum
v
a
l
u
eo
ft
h
eo
b
j
e
c
t
i
v
ea
tt
h
en
o
d
e
.Node2i
st
a
k
e
nupf
o
rf
a
t
h
o
m
i
n
gb
a
s
e
don
n
o
d
eo
r
d
e
r
i
n
g
C
o
m
p
l
e
t
i
o
n
so
fNode2
:T
h
i
se
x
p
a
n
d
sNode2bys
e
t
t
i
n
ged
i
s
c
r
e
t
ev
a
r
i
a
b
l
eX31
0
i
t
sp
e
r
m
i
s
s
i
b
l
es
e
to
fd
i
s
c
r
e
t
ev
a
l
u
e
s
.I
tc
r
e
a
t
e
sb
r
a
n
c
h
e
sNodes5
9
.l
ti
sa
l
s
o
t
h
es
e
c
o
n
dt
i
e
ro
ft
h
et
r
e
e
.Thes
o
l
u
t
i
o
ni
si
n
d
i
c
a
t
e
da
tev
a
r
i
o
u
sn
o
d
e
s
.I
na
l
l
t
ev
a
r
i
a
b
l
e
sa
r
es
e
t1
0i
n
d
i
c
a
l
e
dv
a
l
u
e
sw
h
i
l
eXli
sf
r
e
e
.
o
fthemt
h
etwod
i
s
c
w
h
i
c
hi
sac
o
n
t
i
n
u
o
u
sv
a
r
i
a
b
l
ebyd
e
f
i
n
i
t
i
o
n
.S
i
n
c
ea
l
lo
ft
h
ed
i
s
c
r
e
l
ev
a
r
i
a
b
l
e
s
b
a
v
ed
i
s
c
r
e
t
ev
a
l
u
e
st
h
i
sn
o
d
ei
sa
l
s
of
a
t
h
o
m
e
d
.Node7i
n
d
i
c
a
t
e
st
h
eb
e
s
t
s
o
l
u
t
i
o
n
.S
i
n
c
ea
l
lo
fed
i
s
c
t
ev
a
r
i
a
b
l
e
sa
r
a
s
s
i
g
n
e
dd
i
s
c
l
ev
a
l
u
e
sE
s
o
l
u
t
i
o
na
tl
h
i
sn
o
d
ei
st
h
ec
u
r
r
e
n
ti
n
c
u
m
b
e
n
ts
o
l
u
t
i
o
nTher
e
m
a
i
n
i
n
gn
o
d
e
sw
i
l
l
b
ep
r
u
n
c
d
.Thec
u
r
r
e
n
li
n
c
u
m
b
e
n
ts
o
l
u
t
i
o
n(
f
.=0.1782)i
sh
i
g
h
e
r
s
o
l
u
t
i
o
na
ln
o
d
e3(
f=0
.
1
5
)
.T
h
i
sr
e
q
u
i
r
e
sNod
31
0b
ef
a
t
h
o
m
e
d
.
C
o
m
p
l
e
t
i
o
n
sofNode3
:F
i
g
u
r
e8.
4i
I
lu
s
t
r
a
t
e
st
h
eb
r
a
n
c
h
i
n
ga
tNode3
.Thep
r
o
c
e
s
s
i
ss
i
m
i
l
a
r1
0c
o
m
p
l
e
t
i
o
n
so
fNode2a
b
o
v
e
.Theb
e
s
ts
o
l
u
t
i
o
ni
sa
tNode1
3w
i
t
h
.
1
9
2
7
.Th
i
si
sl
a
r
g
e
r
t
h
a
nt
h
eC
u
r
r
e
n
ti
n
c
u
m
b
e
n
ts
o
l
u
t
i
o
n
.A
l
Io
f
av
a
l
u
eoff=0
en
o
d
e
sc
a
nb
el
e
r
m
i
n
a
l
e
d
.
Nof
u
r
t
h
e
rn
o
d
e
sa
r
ea
v
a
i
l
a
b
l
ef
o
re
x
p
a
n
s
i
o
n
.Thec
u
r
r
e
n
ti
n
c
u
m
b
e
n
ts
o
l
u
t
i
o
ni
s
a
l
s
ot
h
eg
l
o
b
a
lo
p
t
i
m
a
ls
o
l
u
t
i
o
n
.T
h
i
si
st
h
es
o
l
u
t
i
o
nt
oeo
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
e
m
.I
n

.
.

S
e
lc
u
r
r
e
n
ti
n
c
u
m
b
e
n
ts
o
I
u
t
i
o
nt
of
-= "
X
-= [

]
(
m
i
n
i
r
n
i
z
a
t
i
o
n
p
r
o
b
l
e
m
)
.
S
t
c
p1
.I
fnnya
c
t
i
v
ef
e
a
s
i
b
l
ep
a
r
ua
1s
o
l
u
t
i
o
n
sr
e
m
a
i
n( i
sb
e
u
e
rt
h
a
nt
h
ec
u
r
r
e
n
t
s
e
l
e
c
lonea
sX
<
J
)andgot
oS
t
e
p2
.
i
n
c
u
m
b
e
n
ls
o
l
u
t
i
o
n
)

1
.
I
fnonee
x
i
s
t
sa
n
dt
h
e
r
ee
x
i
s
t
sa
ni
n
c
u
m
b
e
n
ts
o
l
u
t
i
o
ni
li
so
p
t
i
ma

I
fnonee
x
i
s
t
sa
n
dt
h
e
i
snoi
n
c
u
m
b
e
n
ts
o
l
u
t
i
o
nt
h
emodeli
si
n
f
e
a
s
i
b
l
e
.
o
m
p
l
e
t
i
o
no
ft
h
ep
a
r
t
i
a
ls
o
l
u
t
i
o
nXusingac
o
n
t
i
n
u
o
u
s
S
t
e
p2
.O
b
t
a
i
n ec
r
e
l
a
x
a
t
i
o
no
ft
h
eo
r
i
g
i
n
a
lmode
l
.
S
t
e
p3
.I
ft
h
e
r
e enof
e
a
s
i
b
l
ec
o
m
p
l
e
t
i
o
n
so
f p
a
r
t
i
a
ls
o
l
u
t
i
o
nX
t
e
r
m
i
n
a
t
e
X
<
I
)

i
n
c
r
e
m
e
n
tss+1

a
n
dr
e
t
u
mt
oS
t
e
p1
.
S
t
e
p4
.I
ft
h
eb
e
s
tf
e
a
s
i
b
l
ec
o
m
p
l
e
t
i
o
no
fp
a
r
t
i
a
ls
o
l
u
t
i
o
nX
<
I
)cannotimprovet
h
e
i
n
c
u
m
b
e
n
ts
o
l
u
t
i
o
n
t
h
e
nt
e
r
r
n
i
n
a
t
eX
i
n
c
r
e
m
e
n
tss+1

a
n
dr
e
t
u
mt
oS
t
e
p

f
eb
e
s
tf
e
a
s
i
b
l
ec
o
m
p
l
e
t
i
o
ni
sb
e
t
t
e
r t
h
eincumben
t u
p
d
a
t
et
h
e
S
t
e
p5
.I
i
n
c
u
m
b
e
n
ts
o
l
u
t
i
o
n

r
.!<I)

x
.
.
x

t
e
r
m
i
n
a
t
eX
<
s
>

i
n
c
r
e
m
e
n
tss+1

a
n
dr
e
t
u
mt
oS
t
e
p1
.

Ex
ample8
.
2TheBranchandBounda
l
g
o
r
i
t
h
mi
sa
p
p
l
i
e
dt
oExample8
.
2
.The
S
hi
sa
p
p
l
i
e
dt
h
r
o
u
g
ht
h
eBB . l
em
a
t
h
e
m
a
t
i
c
a
1modelf
o
rExample8
.
2
w
h
e
r
e
X
l
a
n
dX2a
r
ei
n
t
e
g
e
r
s
i
s
Minimize

!
(
x
" xv=(
X
l- 1
5
)
2+(x2- 1
5
)
2

(
8
.
1
7
)

3
3
4 D
I
S
C
A
E
T
EO
P
T
I
M
I
Z
A
T
I
O
N

8
.
2D
I
S
C
A
E
T
EO
P
T
I
M
I
Z
A
T
I
O
NT
E
C
H
N
I
Q
U
E
S 3
3
5

+x2:
!
>2
6

(
8.
18
a
)

+2.
5
x2:
!
>3
7

(
8
.
1
8
b
)

S
u
b
j
e
c
t1
0
: 8
1
(
X
"X2):

X(

X2):
8
2
(
X
I

X(

X
I~ 0
;

X22
!0

(
8
.
1
8
c
)

Thes
o
l
u
t
i
o
ni
s

X
;
=12; X
;
=10; f=34; g
2i
sa
c
t
i
v
e

(
8
.
1
9
)

e?Litz522eifz::i::zztsz

iz:;22J31f
f
o
l
l
o
w
i
n
gd
e
v
e
l
o
p
m
e
n
t

l 1
2 1
3
]
X1
E[
1
0 I
E [
9 1
0 1
1 ]
2
)

UslngBranchandBound: eBB ei
sshowni
nF
i
g
u
r
e8
.
5
.Thed
a
t
af
o
re
C
c
u
J
a
t
t
h
r
o
u
g
hBBEx2.mw
h
e
r
et
h
ec
o
n
t
i
n
u
o
u
s l
a
t
i
o
ni
ss
o
l
v

a
n
a
l
y
t
i
c
aI
lyu
s
i
n
gt
h
cKTc
o
n
d
i
t
i
o
n
s
.
'
het
r
e
ei
sb
r
n
n
c
h
e
dn
sf
o
l
l
o
w
s
:

NodeO: i
st
h
er
o
o
tn
o
d
e
.A
ll
t
h
ev
a
r
i
a
b
l
e
s ef
r
.ec
o
n
t
i
n
u
o
u
sr
e
l
a
x
a
t
i
o
n
es
o
l
u
t
i
o
ni
si
n
d
i
c
a
t
e
do
n t
r
e
e
.C
u
r
r
e
n
t
l
yt
h
i
si
s
o
ft
h
ep
r
o
b
l
e
mi
ss
o
l
v
e
d
.Th
t
h
eo
n
l
ya
v
a
i
l
a
b
l
ea
c
t
i
v
ep
a
r
t
i
a
ls
o
l
u
t
i
o
na
n
dh
e
n
c
en
e
e
d
st
ob
ec
o
m
p
l
e
t
e
d
.
a
r
t
i
a
ls
o
l
u
t
i
o
na
tNode0i
sc
o
m
p
l
e
t
e
du
s
i
n
gv
a
r
i
a
b
l
eX
If
r
o
mt
h
e
N
o
d
e
s]-4:ep
r
e
d
u
c
e
dd
i
s
c
r
e
t
es
e
ti
d
e
n
t
i
f
i
e
da
b
o
v
e
.T
h
i
sl
e
a
d
st
of
o
u
rb
r
a
n
c
h
e
sa
n
dt
h
e
o
p
t
i
m
a
lv
a
l
u
e
sa
r
i
n
d
i
c
a
t
e
do
nt
h
e.Theb
e
s
tv
a
l
u
e
sa
r
ei
n
d
i
c
a
t
e
da
te
n
o
d
e
s
.F
o
ri
s
p
r
o
b
l
e
m.ya
r
ea
l
s
of
e
a
s
i
b
l
e
.T
h
e
r
e enowf
o
u
ra
c
t
i
v
ep
a
r
t
i
a
l
s
o
l
u
t
i
o
n
s
.
es
o
]
u
t
i
o
na
tNode3i
sa
l
s
oaf
e
a
s
i
b
l
es
o
l
u
t
i
o
no
ft
h
eo
r
i
g
i
n
a
ld
i
s
c
r
e
t
e
Node3
:Th
o
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
e
ms
i
n
c
et
h
ed
e
s
i
g
nv
a
r
i
a
b
l
e
sa
r
ed
i
s
c
r
e
l
ea
n
db
e
l
o
n
gt
oe
t
.N
ode3i
s
f
a
t
J
lO
m
e
d
. i
sp
r
o
v
i
d
e
st
h
ef
i
r
s
ti
n
c
u
m
b
e
n
t
p
e
r
m
i
s
s
i
b
l
ed
i
s
c
r
e
t
es
e
s
o
l
u
t
i
o
n
.
T
h
r
ea
r
et
h
r
e
ea
v
a
i
l
a
b
l
ea
c
t
i
v
ef
e
a
s
i
b
l
ep
a
r
t
i
a
ls
o
l
u
t
i
o
n
s(No
d
e
s1

a
n
d4
)
f
o
rc
o
m
p
l
e
t
i
o
n
.S
i
n
c
et
h
es
o
l
u
t
i
o
n
sa
ten
o
d
e
sa
r
ea
l
lf
e
a
s
i
b
l
e

o
n
l
yt
h
eb
e
s
t
a
c
t
i
v
ef
e
a
s
i
b
l
ep
a
r
t
i
a
ls
o
l
u
t
i
o
ni
sp
i
c
k
e
du
pf
o
rcompletion-Node4
.N
o
d
e
s1
a
n
d2c
a
nb
et
e
r
m
i
n
a
t
e
d
.
C
o
m
p
l
e
t
i
o
nNode4
:Node4h
a
saf
e
a
s
i
b
l
es
o
l
u
t
i
o
nb
e
t
t
e
r ec
u
r
r
e
n
t
i
n
c
u
m
b
e
n
t
s
o
l
u
t
i
o
n
.
l
t
i
sc
o
m
p
l
e
t
e
db
yb
r
a
n
c
h
i
n
g1
0Nodes5
6

8
.O
n
l
yNode
5h
a
saf
e
a
s
i
b
]
es
o
l
u
t
i
o
n
.Tbeb
e
s
ls
o
l
u
t
i
o
na
lNode5i
sn
o
tb
e
t
t
e
rt
h
a
nt
h
ec
u
r
r
e
n
t
6
7
8c
a
nb
et
e
r
m
i
n
n
t
e
d
.
i
n
c
u
m
b
e
n
ts
o
l
u
t
i
o
n
.N
od
e
s5
a
ls
o
l
u
t
i
o
n
sa
n
d eh
a
sb
e
c
nf
a
t
h
o
m
e
d
.Th
e
T
h
e
r
ea
r
en
omorea
c
t
i
v
ep
a
c
u
r
r
e
n
ti
n
c
u
m
b
e
n
ts
o
l
u
t
i
o
ni
st
h
eo
p
t
i
m
a
ls
o
l
u
t
i
o
n
.

X.=[12.8
6

9
.
9
5
1
1
'
*:
:
:
3
3
.
1
3

l
i
sc
o
m
p
l
e
t
e
st
h
e BB s
e
a
r
c
ha
p
p
l
i
e
dt
oE
x
a
m
p
l
e8
.
2
. The d
u
c
t
i
o
ni
n
e
n
u
m
e
r
a
t
i
o
n
c
o
m
p
a
r
e
dt
o e
x
h
a
u
s
t
i
v
ee
n
u
m
e
r
a
t
i
o
n
i
sd
i
f
f
i
c
u
l
t1
0commento
nf
o
r
t
w
or
e
a
s
o
n
s
.

1
.Thed
e
l
i
b
e
r
a
t
ec
h
o
i
c
et
or
e
s
t
r
l
c
tes
e
1o
fd
i
s
c
r
e
t
ev
a
l
u
e
su
s
e
d1
0b
r
a
n
c
ht
h
e
s
e
e
m
e
da
r
b
i
t
r
a
r
y
.W
h
i
l
ei
sw
a
sd
o
n
ef
o
rc
o
n
v
e
n
i
e
n
c
e
as
e
r
i
o
u
sj
u
s
t
i
f
i
c
a
t
i
o
n
nb
ea
d
v
a
n
c
e
db
a
s
e
do
nt
h
ec
o
n
t
i
n
u
o
u
ss
o
l
u
t
i
o
na
tNodeO
.Ther
e
d
u
c
e
d
d
i
s
c
r
e
t
es
e
ti
sc
h
o
s
e
nt
oi
n
c
l
u
d
eal
i
m
i
1
e
dn
e
i
g
h
b
o
r
h
o
o
d
/
w
i
n
d
o
wo
ft
h
e
o
n
.
c
o
n
t
i
n
u
o
u
s 1

XC
o
:
:
:
[
13

9
)
1
'
*=
40

1
n
f
e
a
s
l
b
l
e

1
n
f
s
l
b
l
e

R
g
u
r
e8
.
5B
r
a
n
c
ha
n
db
o
u
n
dI
r
f
o
rExamp
l
e8
.
2
.

I
n
f

s
l
b
l
e

2
.Thee
x
h
a
u
s
t
i
v
ee
n
u
m
e
r
a
t
i
o
n

b
a
s
e
do
nd
i
s
c
t
ev
a
l
u
e
s
wase
a
s
i
e
r
t
oe
v
a
l
u
a
t
ef
o
r
t
i
a
l
ap
u
r
ed
i
s
c
r
e
t
ep
r
o
b
l
e
m
.A
p
p
l
y
i
n
gKTc
o
n
d
i
t
i
o
n
st
oo
b
t
a
i
nt
h
ep
o
s
s
i
b
]
ep
s
o
l
u
t
i
o
n
st
oec
o
n
t
i
n
u
o
u
sr
e
l
a
x
a
t
i
o
n

a
n
dc
h
o
o
s
i
n
gt
h
ea
c
c
e
p
t
a
b
l
es
o
l
u
t
i
o
ni
s
i
k
eSQP1
0s
o
l
v
et
h
e
a
ni
n
v
o
l
v
e
de
x
e
r
c
i
s
e
.U
s
i
n
gn
u
m
e
r
i
c
a
ll
e
c
h
n
i
q
u
e
sI
p
r
o
b
l
e
m
w
h
i
c
hwi
l
1m
o
s
lI
i
k
e
l
yb
ed
o
n
ei
np
r
a
c
t
i
c
e
h
a
si
t
sownp
r
o
b
l
e
m
s
.
D
e
p
e
n
d
e
n
c
eo
ng
o
o
ds
t
a
r
t
i
n
gv
a
l
u
e
sw
i
n
og
u
a
r
a
n
t
e
eo
fc
o
n
v
e
r
g
e
n
c
er
e
q
u
i
r
e
s
u
s
e
ri
n
l
e
r
v
e
n
t
i
o
n1
0i
d
e
n
t
i
f
yep
a
r
t
i
c
u
l
a
rs
o
l
u
t
i
o
n
.I
tw
o
u
l
dr
e
q
u
i
r
al
o
to
f
5
Ss
oa
s1
0a
p
p
l
yeB
r
a
n
c
ha
n
dBounda
l
g
o
r
i
t
h
m
e
f
f
o
r
t1
0a
u
t
o
m
a
t
et
h
ep
r
o
1
0g
e
n
e
r
a
ln
o
n
l
i
n
e
a
rm
i
x
e
do
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
e
m
s
.L
in
e
a
rm
i
x
e
do
p
t
i
m
i
z
a
t
i
o
n
p
r
o
b
l
e
m
sd
on
o
tp
o
s
s
e
s
si
sl
i
m
i
t
a
t
i
o
n
a
tl
e
a
s
ttot
h
es
a
m
ee
x
t
e
nt
.

3
3
6 D
I
S
C
R
E
T
EO
P
T
I
M
I
Z
A
T
I
O
N

se
x
a
m
p
l
e ea
c
t
i
v
ep
a
r
t
i
a
ls
o
l
u
t
i
o
n
sa
tt
h
er
o
o
ta
n
di
n f
i
r
s
tt
i
e
rw
e
r
ea
l
l
I
ni
o
ro
t
h
e
rn
o
d
e
sh
a
v
ea
ni
n
f
e
a
s
i
b
l
e
f
e
a
s
i
b
l
e
.Whatw
o
u
l
db
et
h
es
t
r
a
t
e
g
yi
ft
h
er
o
o
tn
o
d
a
c
t
i
v
ep
a
r
t
i
a
ls
o
l
u
t
i
o
n
?I
n c
a
s
ee88s
e
a
r
c
hp
r
o
c
e
s
sm
u
s
tc
o
m
p
l
e
t
ea
l
la
c
t
i
v
e
p
a
r
t
i
a
ls
o
l
u
t
i
o
n
sa
n
dn
o
to
n
l
yf
e
a
s
i
b
l
ea
c
t
i
v
ep
a
r
t
i
a
ls
o
l
u
t
i
o
n
s
.T
h
a
ti
sam
o
d
i
f
i
c
a
t
i
o
n
a
tc
a
nb
ee
a
s
i
l
ymadet
ot
h
e8
r
a
n
c
ha
n
d80undA
l
g
o
r
i
t
h
ma
b
o
v
e
.
OtherSearchTechniques:let
wos
e
a
r
c
ht
e
c
h
n
i
q
u
e
si
ni
ss
e
c
t
i
o
nn
n
e
l
y

e
x
h
a
u
s
t
i
v
ee
n
u
m
e
r
t
i
o
na
n
db
r
a
n
c
ha
n
db
o
u
n
d
p
r
o
v
i
d
ea
na
d
d
i
t
i
o
n
a
lf
e
a
t
u
r
eo
f
d
i
s
c
o
v
e
r
i
n
gg
l
o
b
a
loptimums
o
l
u
t
i
o
n
s
.
'
he
ya
l
s
ob
e
l
o
n
gt
ot
h
ec
1a
s
so
fm
e
t
h
o
d
su
s
e
d
t
oa
d
d
r
e
s
sg
l
o
b
a
loptimum.be s
ya
v
a
i
l
a
b
i
l
i
t
yo
fl
a
r
g
ea
n
df
a
s
t
e
rc
o
m
p
u
t
i
n
g
r
e
s
o
u
r
c
e
sh
a
sb
r
o
u
g
h
tr
e
n
e
w
e
de
m
p
h
a
s
i
st
og
l
o
b
a
l
l
yo
p
t
i
m
a
ld
e
s
i
g
n
s
.Simu
U
l
t
e
d
a
n
n
e
a
l
i
n
ga
n
dg
e
n
e
t
i
ca
l
g
o
r
i
t
h
m
sa
r
a
mongt
h
el
e
a
d
i
n
gc
a
n
d
i
d
a
t
e
sf
o
rg
l
o
b
a
l
o
p
t
i
m
i
t
i
o
n
. l
e
y ec
o
v
e
r
e
di
nt
h
en
e
x
tc
h
a
p
t
e
ra
n
dc
a
na
l
s
ob
ec
h
a
r
a
c
t
e
r
i
z
e
da
s
i
o
nc
h
n
i
q
u
e
s

l
i
k
e o
n
e
si
ni
ss
e
c
t
i
o
n
.8
0
t
ho
ft
h
e
mc
a
na
n
da
r
eu
s
e
d
e
n
u
m
e
r
en
o
td
i
s
c
u
s
s
e
di
nt
h
i
sc
h
a
p
t
e
r
.Thea
p
p
l
i
c
a
t
i
o
no
f
f
o
rd
i
s
c
r
e
t
eo
p
t
i
m
i
z
a
t
i
o
n
.They '
m
o
s
to
ft
h
e
s
et
e
c
h
n
i
q
u
e
si
ss
t
i
l
lh
e
u
r
i
s
t
i
ca
n
do
f
t
e
nd
e
p
e
n
d
so
ne
x
p
e
r
i
e
n
c
eb
a
s
e
do
n
so
fp
r
o
b
l
e
m
s
.Somea
p
p
l
i
c
a
t
i
o
n
sa
r
ep
r
o
b
l
e
m
n
u
m
e
r
i
c
a
le
x
p
e
r
i
m
e
n
t
samongc
1a
s
s
s
p
e
c
i
f
i
ca
n
dr
e
q
u
i
r
eu
s
e
ri
n
t
e
r
v
e
n
t
i
o
na
n
dl
e
a
r
n
i
n
g
.S
t
a
n
d
a
r
di
m
p
l
e
m
e
n
t
a
t
i
o
n
si
ne
s
e
m
e
t
h
o
d
s
.e
s
p
e
c
i
a
l
l
yf
o
rn
o
n
l
i
n
e
a
rp
r
o
b
l
e
m
s

a
r
en
o
ly
e
tt
h
en
o
r
m
.
8
.
2
.
3 DynamicProgrammlng
D
i
s
c
r
e
t
ed
y
n
a
m
i
cp
r
o
g
r
a
m
m
i
n
g(DDP)i
sa
ne
x
c
i
t
i
n
gt
e
c
h
n
i
q
u
ef
o
rh
a
n
d
l
i
n
gas
p
e
c
i
a
l
c
a
s
so
fp
r
o
b
l
e
m
s
.E
x
a
m
p
l
e
s8
.
1 d8
.
2a
r
en
o
td
i
r
e
c
tm
e
m
b
e
r
so
ft
h
i
sc

a
s
s
. c
h
a
r
d
8
e
l
l
m
a
n[
4
)w
a
sr
e
s
p
o
n
s
i
b
l
ef
o
r t
i
n
d
u
c
i
n
gec
o
n
c
e
p
ta
n
dt
h
ea
l
g
o
r
i
m
.I
ti
sa
n
z
i
n
gp
r
o

d
u
r
eb
a
s
e
do
nB
e
l
l
m
a
n

sp
r
i
n
c
i
p
l
eo
fo
p
t
i
m
a
l
i
t
y
.Th
ep
r
i
n
c
i
p
l
ei
sb
a
s
e
d
o
p
o
nas
e
q
u
e
n
o
fd
e
c
i
s
i
o
n
so
np
a
r
t
i
a
ls
o
l
u
t
i
o
n
s
.s
oa
twhenc
c
r
t
a
i
nv
a
r
i
a
b
l
e
sh
a
v
eb
e
e
n
d
e
t
e
r
m
i
n
e
d
.t
h
er
e
m
a
i
n
i
n
gv
a
r
i
a
b
l
e
se
s
t
a
b
l
i
s
ha
no
p
t
i
m
u
mc
o
m
p
l
e
t
i
o
no
ft
h
ep
r
o
b
l
e
m[
5
)
.
Ano
e
re
x
p
r
e
s
s
i
o
no
ft
h
i
sn
c
i
p
l
ei
so
b
i
n
e
df
r
o
mR
e
f
e
r
e
n
c
e6
.Ano
p
t
i
m
a
ls
e
q
u
e
n

o
fd
e
c
i
s
i
o
n
sh
ep
r
o
p
e
r
t
yt
h
a
ta
tec
r
e
n
tt
i
m
c
.w
h
a
t
e
v
e
rt
h
ci
n
i
t
i
a
ls
t
a
t
ea
n
de
p
r
e
v
i
o
u
sd
e
c
i
s
i
o
n
s
.E maingd
e
c
i
s
i
o
n
sm
u
s
tb
ea
no
p
t
i
m
a
ls
e
tw
i
t
hr
e
g
a
r
dt
oe
s
t
a
t
e
m n
gf
r
o
mt
h
ef
i
r
s
td
s
i
o
n
.DDPi
sa
b
o
u
tas
e
q
u
e
n
c
eo
fd
e
c
i
s
i
o
n
s(
o
f
t
e
ni
nt
i
m
e
)a
n
d
i
ss
o
m
e
t
i
m
e
st
e
r
m
a
s
e
q
u
e
n
t
i
a
ld
s
i
o
np
r
o
b
l
e
m
.P
r
o
b
l
e
m
si
I
k
eE
x
a
m
p
l
e
s8
.
1 d8.
2
d
on
o
td
i
r
e
c
t
l
yf
i
ti
s
c
l s
i
f
i
t
i
o
nb
u
tt
h
e
y nb
s
f
o
r
m
e
dt
of
i
ti
s
r
u
i
r
e
m
e
n
t(
5
)
.
V
e
r
yo
f
t
e
n
.DDPp
r
o
b
l
e
r
n
sc
a
nb
ed
e
s
c
r
i
b
e
da
s o
p
t
i
m
a
lp
a
d
e
t
e
r
m
i
n
a
t
i
o
np
r
o
b
l
e
m
.
S
u
c
hP
hp
r
o
b
l
e
m
sa
r
es
o
l
v
e
du
s
i
n
gd
i
r
e
c
t
e
dg
r
a
p
h
s(
d
i
g
r
a
p
h
s
)
.l
e
s
ep
r
o
b
l
e
r
n
smay
1yh
l
a
v
en
oc
o
n
n
e
c
t
i
o
nt
op
h
y
s
i
J
p
a
t
h
so
rd
i
s
t
a
n
c
e
s
.L
i
k
et
h
e88treey
c
n
a
b
l
e
a
c
t
u
a
b
e
t
t
e
ra
p
p
r
e
c
i
a
t
i
o
n
l
a
p
p
l
i
c
a
t
i
o
no
ft
h
ea
l
g
o
r
i
t
h
m
i
cp
r
o

d
u
r
ea
n
da
r
en
o
ts
i
c
t
l
y q
u
i
r
e
d
.
u
sd
e
f
i
t
i
o
no
fs
.
t
e
sa
n
ds
t
a
g
e
s
.a
n
do
n
l
yt
h
ef
o
r
m
e
ri
sd
i
s
c
u
s
s
e
dh
e
.l
e
DDPr
b
l
i
s
ht
h
ep
r
i
n
c
i
p
l
eo
fo
p
t
i
m
a
l
i
t
ya
r
e )
)
e
d
Ulc
t
i
o
l
l
a
le
q
u
a
t
i
o
n
s
.
e
q
u
a
t
i
o
n
s

e
dt
oe
s
E
x
a
m
p
l
e8.
3w
i
l
lb
eu
s
e
dt
od
e
f
i
n
ee
n
e
w
t
e
r
m
s
.

Example8
.
3 Ane
s
t
a
b
l
i
s
h
e
du
n
i
v
e
r
s
i
t
yi
si
n
l
s
t
e
di
nd
e
v
e
l
o
p
i
n
ganewC
o
)
)
g
e
f
o
r
. T
a
b
l
e8
.
2i
l
Iu
s
t
r
a
t
e
s d
a
t
a
I
n
f
o
r
m
a
t
i
o
nT
e
c
h
n
o
l
o
g
yd
u
r
i
n
gt
h
ec
u
r
r
e
n
ly
e

D
I
S
C
R
E
T
EO
P
T
I
M
I
Z
A
T
I
O
NT
E
C
H
N
I
Q
U
E
S 3
3
7

T
a
b
l
e8
.
2F
a
c
u
l
t
yH
1r
i
n
gP
o
l
l
c
yD
a
t
af
o
rE
x
a
m
p
l
e8
.
3
l
t
e
m
i
r
e
dF
a
c
u
l
t
y
R u
R
e
c
r
u
i
e
n
t
C
o
s
t
U
n
i
tF
a
c
u
1
t
yC
o
s
t
U
n
i
t
R
e
n
i
n
gC
o
s
t

F
a
l
l
5
1
0

W
i
n
t
e
r

S
p
r
i
n
g

Summer

1
0

1
0
2
2

1
0

1
0

2
2

se
n
t
e
r
p
r
i
s
ef
o
rt
h
ef
i
r
s
ty
e

n
e
c
e
s
s
a
r
yt
oc
a
l
c
u
l
a
l
et
h
ec
o
s
to
fr
e
c
r
u
i
t
i
n
gf
a
c
u
l
t
yf
o
ri
o
fo
p
e
r
n
.Th
ec
o
l
l
e
g
ew
i
l
lo
p
e
r
a
t
ed
u
r
i
n
gf
o
u
rq
u
a
r
t
e
r
s(
F
a
U
.W
i
n
t
e
r
.Sp
g
. d
o
l
u
m
n
s
.D
e
c
i
s
i
o
n
sa
r
ea
s
s
o
c
i
a
t
e
dw
i
t
ht
h
eb
e
g
i
n
n
i
n
go
ft
h
e
Summer)i
n
d
i
c
a
t
e
db
yec
ef
i
r
s
trowi
n
d
i
c
a
t
e
st
h
enewf
a
c
u
l
t
yr
e
q
u
i
r
e
dt
oi
m
p
l
e
m
e
n
tt
h
ec
o
u
r
s
a
t
q
u
a
r
t
e
r
s
.Th
e
c
r
u
i
t
m
e
n
tc
o
s
tf
o
rnewf
a
c
u
l
t
ye
a
c
hq
u
a
r
t
e
r
.
w
i
l
lb
eo
f
f
e
r
e
d
.Thes
e
c
o
n
drowi
ser
o
m
p
e
t
i
t
i
v
ed
m
andf
o
rq
u
a
l
i
f
i
e
dp
e
o
p
l
ei
nt
h
i
s
w
h
i
c
hi
ss
i
g
n
i
f
i
c
a
n
tb
e
c
a
u
s
eo
fec
. l
ei
r
drowi
seu
n
i
tc
o
s
to
fnewf
a
c
u
l
t
yi
nt
e
r
m
so
ft
h
es
t
a
n
d
a
r
di
n
s
t
i
t
u
t

f
a
c
u
l
t
yc
o
s
.
tC
o
u
r
s
es
p
e
c
i
a
l
io
na
n
da
v
a
i
l
a
b
i
l
i
t
ymakest
h
i
sc
o
s
td
i
f
f
e
r
e
n
ti
n
n
i
tc
o
s
lf
o
rr
e
t
a
i
n
i
n
ge
x
t
r
af
a
c
u
l
t
yh
i
r
e
di
nE
d
i
f
f
e
r
e
n
tq
u
a
r
t
e
r
s
.Thel
a
s
trowi
s u
e
a
r
l
i
e
rq
u
a
r
t
e
r(
t
om
i
n
i
m
i
z
ec
o
s
to
v
e
rt
h
ey
e
a
r
)
.Whats
h
o
u
l
db
et
h
eh
i
r
i
n
gp
o
l
i
c
yf
o
r

u
n
i
v
e
r
s
i
l
yl
om
i
n
i
m
i
z
ec
o
s
l
?
l
i
se
x
a
m
p
l
ei
sc
1e
a
r
l
yi
l
1
u
s
t
r
a
t
i
v
eo
fs
e
q
u
e
n
t
i
a
ld
e
c
i
s
i
o
nm
a
k
i
n
g
.D
e
c
i
s
i
o
n
sa
r
e
r
e
q
u
dt
ob
emadee
v
e
r
yq
u
e
r
.F
o
re
x
a
m
p
l
e

i
ti
sp
o
s
s
i
b
l
e1
0c
r
u
i
ta
l
ln
e
e
d
e
df
a
c
u
l
t
y
i
r
s
tq
u
a
r
t
e
ra
n
dp
a
yBn
i
n
gc
o
s
t
.o
ri
tmayc
o
s
tl
e
s
st
oh
i
r
ed
u
r
i
n
g c
h
d
u
r
i
n
gef
quarter or~tforfirst ree q
u
a
r
t
e
r
sa
n
dh
i
r
ea
g
a
i
ni
ne
f
o
.l
ep
r
o
b
l
e
mw
i
l
lb
e
p
t
i
m
a
U
s
h
o
r
t
e
s
tp
a
t
hm
e
t
h
o
de
v
e
no
u
g
h
t
h
e
a
r
e
n
o
r
o
u
i
n
v
o
l
v
e
d
i
n
s
o
l
v
e
du
s
i
n
g o
.
e
p
r
o
b
l
e
m
.

b
eb
e
s
tw
a
y
st
ou
n
d
e
r
s
t
a
n
de
s
h
o
r
t
e
s
t
p
h
a
l
g
o
r
mi
st
od
r
a
wad
i
r
e
c
t
e
d
g
r
a
p
h(
d
i
g
r
a
p
h
)b
e
t
w
n
t
h
en
o
d
e
s
l
s
t
a
t
e
si
nt
h
ep
r
o
b
l
e
m
.

S tes: S
t
a
t
e
s p
r
e
s
e
n
tp
o
i
n
t
s
l
c
o
n
d
i
t
i
o
n
si
n DP w
h
e
r
ed
e
c
i
s
i
o
n
sh
a
v
et
ob
e
c
o
n
s
i
d
e
r
e
d
.The
s
ea
r
es
o
m
e
t
i
m
e
s
'
e
dt
oa
sp
o
l
i
c
yd
e
c
i
s
i
o
n
so
rj
u
s
tp
o
l
i
c
i

.Thi
s
u
e
n
t
iaJ c
o
n
n
e
c
t
i
o
na
sc
u
r
r
e
n
td
e
c
i
s
i
o
n
sc
a
n
n
o
tb
en
s
i
d
e
dU
D e
h
a
sas
p
r
e
v
i
o
u
sd
e
c
i
s
i
o
n
sh
a
v
eb
e
e
nm
a
d
e
.C
l
e
a
r
l
yi
nExample8.3p
r
e
s
e
n
te
b
e
g
i
n
n
i
n
go
fe
a
c
hq
u
a
r
t
e
r
.l
enumbero
fs
t
a
t
e
si
su
s
u
a
l
l
yo
n
emore t
h
e
n
u
m
b
e
r
c
i
s
i
o
n
st
oi
n
c
o
r
p
o
r
a
t
et
h
ee
f
f
e
c
to
ft
h
el
a
s
tp
o
l
i
c
y
/
d
e
c
i
s
i
o
n
.Nodesi
nt
h
ed
i
g
r
a
p
h
ofd
s
e
n
ts
t
a
t
e
si
nt
h
eDPp
r
o
b
l
e
m
.Thed
i
g
r
a
p
hr
e
p
r
e
s
e
n
t
sac
o
n
n
e
c
t
i
o
nb
e
t
w
e
e
ne
a
c
h
r
e
p
ti
sp
e
r
m
i
s
s
i
b
l
e
.T
h
i
sc
a
nb
emadec
1e
a
r
e
rb
ys
t
a
t
i
n
g ti
fan
o
d
ei
s
p
a
i
ro
fs
t
a
t
e
sa
. t
h
e
nt
h
e
r
ei
sad
i
r
e
c
t
g
o
i
n
gt
ob
ea
f
f
e
c
t
e
db
yad
e
c
i
s
i
o
na
ta
n
yp
r
e
v
i
o
u
sn
o
d
Ii
n
e
l
a
r
c
)b
e
t
w
e
e
nenode
.T
hec
o
s
to
ft
h
ec
o
n
n
e
c
t
i
o
ni
si
n
d
i
c
a
t
e
do
nE
c
o
n
n
e
c
t
i
o
n(
f
.Ana
r
r
o
wc
a
nb
eu
s
e
dt
oi
n
d
i
c
a
t
et
h
ed
i
r
e
c
t
i
o
no
ft
r
a
v
e
r
s
ea
l
o
n
gi
s
c
o
n
n
e
c
t
i
o
ni
t
s
e
l
e
s
eI
in

s
/
a
r
c
se
r
e
f
o
r
er
e
p
r
e
s
e
n
td
e
c
i
s
i
o
f
l
s
.O
ncet
h
ed
i
g
r
a
p
hf
o
r
l
i
n
e
l
a
r
cs
e
g
m
e
nt
.Th
t
h
eDPp
r
o
b
l
e
mi
sa
v
a
i
l
a
b
l
e
.t
h
eoptimums
o
l
u
t
i
o
nc
o
r
r
e
s
p
o
f
l
d
st
ot
h
es
h
o
r
t
e
s
tp
a
t
h
fromI
h
eb
e
g
i
n
n
i
l
l
g10t
h
ee
n
d
i
n
gs
t
a
t
eo
f
t
h
ed
i
g
r
a
p
h[
3
)
.

8
.
2D
I
S
C
R
E
T
EO
P
T
I
M
I
Z
A
T
I
O
NT
E
C
H
N
I
Q
U
E
S 339

338 D
I
S
C
E
T
EO
P
T
I
M
I
Z
A
T
I
O
N

V
a
l
u
e(
1
) 0
V
a
l
u
e(
k
)=min(
V
a
l
u
e(
+C
(
/
k
)
;1S1<k
} k=2...1
1+1

9
9

Forexampl

i
nExample8
.
3

V
a
L
u
e(
4
) m
i
n(Value(
1
)+C
(
l
4
)
;V
a
l
u
e(
2
)+C(
2
4
)
;V
a
l
u
e(
3
)+C(3
4
)
}
[
0+
8
5
]
;[
1
5+6
2
]
;[
4
5+2
6
]
)=7
1
V
a
l
u
e(
4
)=min(

w
h
i
c
he
s
t
a
b
l
i
s
h
dt
h
eminimumv
a
l
u
e
)
Node(
4
) 3(

Th
e
s
ec
a
1
c
u
l
a
t
i
o
n
sa
r
eembeddedi
nEx8_3.m.Th
eo
u
t
p
u
tf
r
o
mt
h
ep
r
o
g
r
a
mI
S
:
7
4
F
l
g
u
r
e8
.
6D
i
g
r
a
p
hf
o
rE
x
a
m
p
l
e6
.
3
.

8
.
6i
s d
i
g
r
a
p
hf
o
rt
h
eh
i
r
i
n
gp
o
l
i
c
y
.T
h
e
r
ea
r
ef
i
v
es
t
a
t
e
sw
i
t
ht
h
ef
i
r
s
t
F
i
g
u
r
f
o
u
rr
e
p
r
e
s
e
n
t
i
n
gequ
w
h
e
r
et
h
ed
e
c
i
s
i
o
n
smustb
em
a
d
e
.Node1 P
s
e
n
t
s
i
g
r
a
p
hi
st
h
C
0
5
to
ft
h
ed
e
c
i
5
i
o
no
r
t
h
eF
a
l
lq
u
a
r
t
e
ra
n
d5
0o
n
.Alongt
h
ea
r
c
soned
p
o
l
i
c
ya
twi
J
Io
r
i
g
i
n
a
t
ea
tt
h
ec
u
r
r
e
n
tnodea
n
dmeeta
11oft
h
er
e
q
u
i
r
e
m
e
n
t
sa
te
l
r
n
o
d
e
.I
no
r
d
e
r1
0u
s
t
h
ei
n
f
o
r
m
a
t
i
o
ni
nt
h
ed
i
g
r
a
p
he
f
f
i
c
l
i
v
e
l
y
and1
0d
e
v
e
l
o
p
t
h
eq
u
a
r
t
e
r
sa
r
ea
s
s
i
g
n
e
di
n
d
i
c
e
sf
r
o
m1t
o4
.Thef
o
l
l
o
w
i
n
g
I
h
em
a
t
h
e
m
a
t
i
c
a
lmodel
a
b
b
r
e
v
i
a
t
i
o
n
sa
r
eu
s
e
d
:r
e
q
u
i
r
e
df
a
c
u
l
l
YRF; c
r
u
i
t
m
e
n
tC
O
S
I
RC;u
n
i
tf
a
c
u
l
l
yc
o
s
t
FC;u
n
i
lr
e
l
a
i
n
i
n
gc
o
s
t-h
o
l
d
i
n
gc
o
s
l

HC.Thep
o
l
i
c
yd
e
c
i
s
i
o
n
so
rc
o
s
tb
e
t
w
e
e
nt
h
e
s
t
a
r
tn
o
d
ei
andt
h
ee
n
dnodeji
sC(i

j
)
.F
ore
x
a
m
p
l
e
.C(l

4
)r
e
p
s
e
n
t
sep
o
l
i
c
y
t
h
eF
a
l
lq r
t
e
r

whichw
i
l
lmeetes
t
a
f
f
i
n
gr
e
q
r
e
m
e
n
t
sf
o
reWinlerand
S
p
r
i
n
gq
u
a
r
t
e
r
sa
1
5
0t
h
u
sp
a
y
i
n
gea
p
p
r
o
p
r
i
a
t
er
e
t
a
i
n
i
n
gpena
1t
i
e
s
:

Example 8.3Dynamic programming

j)
Cost between pairs of nodes/states C(
i
o 15 45 85 99
o
0 30 62 74
o 0 0 26 34
o
0 0 0 12
0 0 0 0 0
odes - Values(i)
optimal path values at var~o
o 15 45 71 79
Best previous node o reach current node -Node(i)
0 1 2 3 3
TheV
a
l
u
e(
4
)i
s7
1a
l
l
dNode(
4
)i
s3
.Theoptimumv
a
l
u
ef
o
rt
h
ep
r
o
b
l
e
mi
sV
a
l
u
e
whichi
s
1
U
a
11
07
9
.
(
5
)

S
h
o
r
t
e
s
tP
a
t
h
:Thes
h
o
r
t
e
s
tp
a
t
hi
so
b
t
a
i
n
e
dbyb
a
c
k
t
r
a
c
k
i
n
gf
r
o
mt
l
l
ee
n
ds
t
a
t

C(
l
4
)=
RC
(l
)+[RF
(
l
)+RF(2)+RF(3)).FC
(
l
)
+([RF(2)+RF(3))HC(l
)+[
R
F
(
3
)
]HC(2))
Ex8_3.mi
st
h
em
f
i
l
ea
tg
e
n
e
r
a
t
e
st
h
ed
a
t
af
o
rt
h
ed
i
g
r
a
p
hf
o
rE
x
a
r
n
p
l
e8
.
3 d
0
l
u
t
i
o
nbys
o
l
v
i
n
gt
h
e
f
u
n
c
l
i
o
n
a
leq
l
i
o
n
sa
n
db
a
c
k
t
r
a
c
k
i
n
gepa

i
d
e
n
t
i
f
i
e
se5
t
oi
d
e
n
t
i
f
yt
h
eo
p
t
i
ma
1s
q
u
e
n
c
e
.Notea
tf
o
rd
i
f
f
e
r
n
tp
r
o
b
l
e
m
son
1
yt
h
ec
o
s
t
J
Id
i
f
f
e
r
.Thef
u
n
c
t
i
o
n
a
le
q
u
a
l
i
o
n
sa
n
db
a
c
k
t
r
a
c
k
i
n
g e
l
e
m
e
n
t
so
ft
h
e
c
a
l
c
u
l
a
t
i
o
nwi
DDPmethoda
n
ds
h
o
u
l
db
et
h
es
a
m
e
.

F
u
n
c
t
i
o
n
a
lEquatlonf
o
rDynamlcProgrammlng:Thef
u
n
c
t
i
o
na
1e
q
u
a
t
i
o
ni
s
1g
o
r
i
t
h
mo
rm
e
t
h
o
d
.I
temploysar
e
c
u
r
s
i
v
ep
r
o
d
u

t
h
ee
s
s
e
n
t
i
a
lp
a
r
to
ft
h
eDDPa
1
0i
d
e
n
t
i
f
yt
h
eb
e
s
ls
o
l
u
t
i
o
na
le
a
c
hs
t
a
t
e
.L
e
lV
a
l
u
e(
i
)b
et
h
b
e
s
tv
a
l
u
eo
ft
h
ed
e
c
i
s
i
o
n
a
ln
o
d
ei
.l
a
k
i
n
gi
n
l
oa
c
c
o
u
n
ta
1
1p
o
s
s
i
b
l
ea
r
cs
/
p
a
t
h
st
h
a
ll
e
a
dt
on
o
d
ei
.Le
tNode(
i
)
b
et
h
enode l
e
dt
ot
h
eb
e
s
tp
o
l
i
c
ya
tec
u
r
r
e
n
ts
t
a
t
e
.Th
er
e
c
u
r
s
i
o
nr
e
l
a
t
i
o
n
sf
o
r
I
Js
t
a
t
e
sc
a
nb
es
e
lupa
sf
o
l
l
o
w
s
:

i
n
gt
h
eNode(
i
)
i
n
f
o
n
n
a
t
i
o
n
w
h
i
c
hi
st
h
erowo
fd
a
t
aa
b
o
v
e
.
AtNodB5thEbestnodEfromwhicht
or
e
a
c
hNode5
i
sNode3
o

2 3 3

AINod
3t
h
eb
e
s
tnodefromwhicht
or
e
a
c
hNode3i
sNode2
o

2 3 3

AtNode2t
h
eb
e
s
tn
o
d
efromwhicht
or
e
a
c
hNode2i
sNode1(
t
h
i
si
ses
t
a
r
t
m
g
n
o
d
e
)
o

2 3 3

T
h
i
ss
o
l
u
t
i
o
ni
si
n
d
i
c
a
t
e
dbyet
h
i
c
ks
o
I
i
dl
i
n
ei
nF
i
g
u
r
e8
.
6
.

340 O
I
S
C
R
E
T
EO
P
T
I
M
I
Z
A
T
I
O
N

8
.
3

n
ea
c
a
I

h
l

i
i

r
i
t

I
g

p
o
l
i

cy
a
.
c

c
o
r
l

d
n
g
t

oi

ss
o
l

ut
i

o
n
i

s
t

o
h

i
i

e s
t
t

her

i
I
1
e

df
:
a

c
u

I
t

y
f
o
r Fa
Jlq
u
t

e
r t
h
eFa
Jlq
u
t

e
r
.Dot
h
esamef
o
rt
h
eW
i
n
t
e
rq
u
a
r
t
e
r
.Fort
h
e
S
p
r
i
n
gq
u
a
r
t
e
r

however

h
i f
o
rb
o
t
ht
h
eS
p
r
i
n
gandSummerq
u
a
r
t
e
r
s
.

A
l
t
e
r
n
a
t
ePaths:Therei
sa
na
l
t
e
r
n
a
t
es
o
l
u
t
i
o
nf
o
rt
h
csameo
p
t
i
m
a
lv
a
l
u
e
.The
o
u
t
p
u
tfromEx8_3.mi
n
c
l
u
d
e
si
si
n
f
o
r
m
a
t
i
o
nt
h
o
u
g
hi
twasn
o
ti
n
c
l
u
d
e
da
b
o
v
e
.I
n
eo
u
t
p
u
ti
nt
h
ecommandwindow
e
a
c
hrowc
o
n
t
a
i
n
sa
d
d
i
t
i
o
n
a
ln
o
d
e
sa
tw
i
l
ly
i
e
l
d
sameo
p
t
i
m
a
lva
Ju
e
. l
e
o
n
l
y
e
n
yi
s1i
nt
h
i
r
drowands
e
c
o
n
dc
o
l
u
r
n
n
.T
h
i
ss
u
g
g
e
s
t
s
t
h
a
t et
h
iI
1
dnode
Node(
3
)c
a
n ohavet
h
ev
a
l
u
eo
fl-thati
s
Node1a
l
s
o
p
r
o
v
i
d
e
st
h
esamev
a
Ju
et
or
e
a
c
hNode3
.Th
i
sw
i
l
lc
a
u
s
eeNode(
i
)i
n
f
o
r
m
a
t
i
o
nt
o
changea
s

Thes
h
o
r
t
e
s
tpa
C
e
nbei
n
t
e
r
p
r
e
t
e
da
sf
o
l
l
o
w
s
:
AtNode5t
h
eb
e
s
tnodefromwhicht
or
e
a
c
hNode5i
sNode3

I
nF
i
g
u
r
e8
.
6t
h
i
si
si
n
d
i
c
a
t
e
dbyt
h
et
h
i
c
kd
a
s
h
e
da
r
c
s
.Theh
i
r
i
n
gp
o
l
i
c
ynowwould
bet
oh
i
r
ef
o
rFaUandW
i
n
t
e
ri
nFal
1
.Thenh
i
r
ef
o
rS
p
r
i
n
gandSummeri
nS
p
r
i
n
g
.
~/gorlthm: DynamlcPro

r
ammlnga

Thef
ol
1
owinga
l
g
o
r
i
t
h
mi
si
d
e
n
t
i
f
i
c
di
n
R
e
f
c
r
e
n
c
c3f
o
rpa
st
h
a
tdon
o
tc
y
c
l
e
.Th
e
o
r
i
g
i
n
a
l8el
1mana
l
g
o
r
i
t
h
mi
n
c
Ju
d
e
sa
ni
t
e
r
a
t
i
v
e f
i
n
e
m
e
n
tt
h
a
ta
l
l
o
w
sf
o
rc
y
c
l
i
c
pa
s
.I
nExample8
.
3e
r
ea
r
enoc
y
c
l
i
cpa
s
.T
h
i
sa
l
g
o
r
i
t
h
mi
si
m
p
l
e
m
e
n
t
e
di
n
Ex8_3.m.

S
t
e
p0
;Numbern
o
d
e
si
ns
e
q
u
e
n
t
i
a
lo
r
d
e
r
.Arcsa
r
d
e
f
i
n
e
df
o
rj>i
.S
e
tt
h
es
t
a
r
t
p
t
i
m
a
lv
a
l
u
ea
s
nodeSo

V
a
l
u
e(
s
)=0
S
t
e
p1
:T
e
r
m
i
n
a
t
ei
fa
l
lV
a
l
u
e(
k
)havebeenc
a
l
c
u
l
a
t
e
d
.I
fpr
e
p
r
e
s
e
n
t
st
h
el
a
s
t
u
n
p
r
o
c
e
s
sdn
o
d
e
.t
h
e
n

V
a
l
l
l
e(
P
)=
m
i
n
{V
a
/
u
e(
i
)+C(
i
.p
)
:i
fa
r
c(
i

p
)e
x
i
s
t
s1Si
<p}

p
) t
h
enumbero
fNodeit
h
a
ta
c
h
i
e
v
e
dt
h
eminimumabove
Node(
R
e
t
u
r
nt
oS
t
e
p1

T
h
i
swasas
i
m
p
l
eexampleo
fDDP.I
ti
si
m
p
o
r
t
a
n
tt
on
o
t
ea
ts
o
l
u
t
i
o
ni
nf
i
n
a
l
l
u
e(
i
nc
Ju
d
e
son
1
yt
h
eb
e
s
ts
o
l
u
t
i
o
na
ta
l
ls
t
a
t
e
s
/
n
o
d
e
s
.L
ikeo
t
h
e
rt
e
c
h
n
i
q
u
e
s
t
h
e
'cs
e
v
e
r
a
ld
i
f
f
e
r
e
n
tc
a
t
e
g
o
r
i
e
so
fp
r
o
b
l
e
m
sa
d
d
r
e
s
s
e
dbyDDP.i
n
c
Ju
d
i
n
gs
p
e
c
i
a
l
a
p
s
a
c
kp
r
o
b
l
e
m
s
.I
ti
sa
l
s
op
o
s
s
i
b
l t
ob
r
i
n
gExamples
d
i
s
c
r
e
t
ep
r
o
b
l
e
m
sl
i
k
et
h
eKn
8
.
1and8
.
2u
n
d
e
rt
h
eframeworko
fDDP.t
h
o
u
g
ht
oi
m
p
a
r
tas
e
q
u
e
n
t
i n
a
t
u
r
l
i
nt
h
e
p
r
o
c
e
s
so
fs
e
l
e
c
t
i
n
gd
e
s
i
g
nv
a
r
i
a
b
l
e
si
nt
h
i
sc
a
s
emayi
n
v
o
l
v
eat
o
t
r
a
n
s
f
o
r
m
a
t
i
o
no
f
ep
r
o
b
l
e
m
.Event
h
e
ni
tmayn
o
tb
ee
f
f
i
c
i
e
n
tt
os
o
l
v
ei
tu
s
i
n
gDDP.I
np
r
a
c
t
i
c
e
DDP
h
a
sa
l
w
a
y
sb
e
c
nr
e
g
a
r
d
e
da
sc
o
m
p
u
t
a
t
i
o
n
a
l
l
ycomplexandp
r
o
h
i
b
i
t
i
v
e
.I
ti
ss
e
l
d
o
m
u
s
e
df
o
rp
r
o
b
l
e
m
swi al
a
r
g
enumbero
fs
t
a
t
e
so
ri
ft
h
ee
s
t
i
m
a
t
i
o
no
fc
o
s
ti
s
i
n
t
e
n
s
i
v
e
.A
n
o
t
h
e
rc
r
i
t
i
c
i
s
mo
fDDP i
s a
ti
tca
e
so
u
t al
o
to
fr
e
d
u
n
d
a
n
t
c
o
m
p
u
t
a
t
i
o
nf
o
rc
o
n
t
i
n
u
o
u
sp
r
o
b
l
e
m
s
.T
o
d
a
y
.t
h
e
s
ea
r
el
e
s
so
fa
ni
s
s
u
e i
nt
h
e
.Fors
t
o
c
h
a
s
t
i
candd
i
s
c
r
e
t
ep
r
o
b
l
e
m
sDDPmays
t
i
1
1beane
f
f
e
c
t
i
v
et
o
o.
l8el
1man
p
a
st
h
a
sa
l
s
oi
n
d
i
c
a
t
e
dt
h
a
ti
tc
b
eu
s
c
dt
os
o
l
v
eas
e
to
fu
n
s
t
a
b
l
es
y
s
t
e
m
so
fe
q
u
a
t
i
o
n
s
.
DDPa
l
s
ocompu
sd
e
s
i
g
ns
e
n
s
i
t
i
v
i
t
ya
sp
a
r
to
fi
t
ss
o
l
u
t
i
o
n
.Th
i
ss
e
c
t
i
o
nwas
prim
ym
eantt
oi
n
t
r
o
d
u
c
eDDPa
sau
s
e
f
u
lt
o
o
lf
o
rsomec
a
t
e
g
o
r
i
e
so
fp
r
o
b
l
e
m
s
.
R
e
a
d
e
r
sa
r
ce
n
c
o
u
r
a
g
e
dt
oc
o
n
s
u
l
ter
e
f
e
r
e
n
c
e
sf
o
rmorei
n
f
o
r
m
a
t
i
o
nonesu
e
c
.
t

AtNode3t
h
eb
e
s
tnodefromwhicht
or
e
a
c
hNode3i
sNode1

A
O
O
I
T
I
O
N
A
LE
X
A
M
P
L
E
S 3
4
1

8
.
3 ADDITIONALEXAMPLES
I
nt
h
i
ss
e
c
t
i
o
n
.twoa
d
d
i
t
i
o
n
a
le
x
a
m
p
l
e
sa
r
ep
r
e
s
e
n
t
e
d
.Thef
i
r
s
to
n
ei
sa
ne
n
g
i
n
e
e
r
i
n
g
n
p
l
ef
r
o
mm
e
c
h
a
n
i
c
a
ld
e
s
i
g
nu
n
di
l
l
u
s
t
r
a
t
e
s how t
h
ed
i
s
c
t
ev
e
r
s
i
o
no
fa
c
x
f
o
u
r
v
a
r
i
a
b
l
eproblemv
e
r
ye
f
f
e
c
t
i
v
e
l
yr
e
d
u
c
e
st
oas
i
n
g
l
e
v
a
r
i
a
b
l
ep
r
o
b
l
e
ma
t
c
a
n
b
ee
a
s
i
l
ya
d
s
e
dby e
n
u
m
e
m
t
i
o
n
. An u
n
b
e
l
i
e
v
a
b
l
er
e
d
u
c
t
i
o
ni
nc
o
m
p
l
e
x
i
t
y
y
i
e
l
d
i
n
gg
l
o
b
a
lo
p
t
i
m
i
z
a
t
i
o
ni
so
b
t
a
i
n
e
dbye
l
e
m
e
n
t
a
r
yc
a
l
c
u
l
a
t
i
o
n
s
.I
td
e
f
i
n
i
t
e
l
y
i
n
v
i
t
e
ss
c
r
u
t
i
n
yr
e
g
a
r
o
i
n
g E
o
r
ti
ns
o
l
v
i
n
gc
o
n
t
i
n
u
o
u
sn
o
n
l
i
n
e
a
ro
p
t
i
m
i
z
a
t
i
o
n
p
r
o
b
l
e
m
s
.Th
es
e
c
o
n
donei
sea
p
p
l
i
c
a
t
i
o
noft
h
ebmnchandbounds
e
a
r
c
ht
oa0-1
i
n
t
c
g
c
rprogmmmingp
r
o
b
l
e
m
.Mucho
ft
h
ea
p
p
l
i
c
a
t
i
o
no
fd
i
s
c
r
e
t
eo
p
t
i
m
i
z
a
t
i
o
ni
si
n
e '
e
ao
fd
e
c
i
s
i
o
nma
n
g
.andd
e
c
i
s
i
o
n
sl
i
k
ef
o
ro
ra
g
a
i
n
s
t
.y
e
so
rno

a
r
ep
o
s
s
i
b
l
c
l
.
u
s
i
n
ga0-1l
i
n
e
a
ri
n
t
e
g
e
rprogrammingmode
8
.
3
.
1 Example8
.
IBeamDeslgn

u
r
a
1designproblem.I
nf
a
ci
ti
sa
n
o
t
h
e
rv
a
r
i
a
t
i
o
n
l
a
p
p
l
i
c
a
t
i
o
n

h
i
scxamplei
sas
t
r
u
ct
l
i
e
r
.Th
ec
o
n
t
i
n
u
o
u
sproblcmwasd
e
v
e
l
o
p
e
da
n
d
o
ft
h
eI
b
c
a
ms
t
r
u
c
t
u
r
es
e
e
n r
s
o
l
v
e
da
sp
a
r
to
fas
t
u
d
e
n
tp
r
o
j
e
c
.
tI
ni
sexampleebeami
su
s
e
da
sah
o
r
i
z
o
n
t
a
l
E"s
e
r
i
e
sgan
c
m
n
e
.Th
eh
o
i
s

tw
i
t
hac
a
p
a
c
i
t
yo
fW 2000I
b
.
memberi
naf
i
x
e
d
i
sa
tt
h
ec
e
n
t
e
ro
ft
h
eb
e
1
.Th
eo
b
j
e
c
t
i
v
eisdesignaminimummassb
e
nt
h
a
tw
i
l
l
n
o
tf
a
i
1underb
e
n
d
i
n
g
.s
h
c
a
r
ands
p
c
c
i
f
i
e
dd
e
t
l
e
c
t
i
o
n
.Th
el
e
n
g
t
hLo
ft
h
ebeami
s
e
n
.Thcmoduluso
f
s
p
e
c
i
f
i
c
da
s96i
n
.S
t
e
e
lwasc
h
o
s
c
nf
o
rt
h
cm
a
t
e
r
i
a
lo
feb
e
l
a
s
t
i
c
i
t
yo
fs
t
e
e
li
sE=
29E+0
6p
s
.
iTheweightd
e
n
s
i
t
yo
fs
I
i
sP=
0.284Ib/in 3
Th
ey
ic
Jds
n
g
i
n oni
s
'
y=36E+0
3p
si
.Th
ey
i
e
l
ds
t
r
e
n
g
i
ns
h
e
a
ri
sy=
2
1E+0
3p
si
.
esp
i
f
i
e
da
l
l
o
w
a
b
l
ed
e
t
l
t
i
O
Di
s0
.
2
5i
n
.Af
a
c
t
o
ro
fs
a
f
e
t
yo
f1
.
4on
t
h
el
o
a
di
sa
s
s
u
m
e
d
.F
ig
u
r
e8
.
7i
I
Iu
s
t
e
st
h
ec
e
n
cl
o
a
d
i
n
gont
h
ebeama
sw
e
l
la
s
A
11
c
a
l
c
u
l
a
t
i
o
n
sa
r
ei
ni
n
c
h
e
sandp
o
u
n
d
s
.
)
t
h
ed
e
s
i
g
nv
a
r
i
a
b
l
e
s
.(

/
/

.
:
.
.
.
.
:
<
<
'
1
x
/
/

342 D
I
S
C
R
E
T
EOpnMIZAnON
1

111

8
.
3A
D
D
I
T
I
O
N
A
LE
X
A
M
P
L
E
S 343

Z
X
4

v-MY

F
i
g
u
r
e
8
.
7D
e
s
i
g
nv
a
r
i
a
b
l
e
s
:Exa
m
p
l
e8
.
4
.

X~ =5 2
8
;

Them
a
t
h
e
m
a
t
i
c
a
lmodelf
o
rt
h
i
sp
r
o
b
l
e
mc
a
nb
ee
x
p
s
s
e
da
s
:
M
i
n
i
m
i
z
e

f
.
i
X)=pL

.
A

S
u
b
j
e
c
tt
o
: g
l
(
X
)
:W U-12EI~ 0 (
d
e
f
l
t
i
o
n
)
g
2
(
X
)
:Wlxl- 8/y~O

(
n
o
r
m
a
ls
l
r
e
s
s
)

1
X
g
3
(
X
)
:W -2

yS0 (
3
s
h
e
a
rs
5
S
)

Someg
e
o
m
c
t
r
i
cc
o
n
s
t
r
a
i
m
st
or
e
l
a
t
ct
h
ev
a
r
i
o
u
sd
e
s
i
g
nv
a
r
i
a
b
l
e
sa
r

g
4
(
X
)
:xl- 3
X
2:
;0
:
;
;0
g
5
(X
)
:2
x
2-xl:

g
6
(
X
)
:x
3-1
.
5
x
4~ 0
g
7
(
X
)
:0
.
5
X
4-x
3~ 0

Th
es
i
d
ec
o
n
s
t
r
a
i
n
t
soned
e
s
i
g
nv
a
r
i
a
b
l
e
sa
r
e
3
:
:
;
x1S20
2
:
:
;
x
:
;
1
50.
125:
:
;
x
.
7
5 0
.
2
5:
:
;
X
;1
.
2
5
2:
3S0
4:

r
o
s
s
s
e
c
t
i
o
n
a
la
a
(
A
c
)
momento
fi
n
e
r
t
i
a
and
Thef
o
l
l
o
w
i
n
gr
e
l
a
t
i
o
n
sdefmeec
r
e
aa
b
o
u
tt
h
ec
e
n
i
d(
Q
c
)
:
f
i
r
s
tmomento
fa

A
c=2x
X
r4+X1X)-2
4
1=X~ ~ _(
x
-X4
)
(
X1-2
x
3
)
3
2
=- ---=--'--"'------::
1
2
1
2
Qc=0.5.iXI-X~+0.5x3(XI -X~2

Thcs
o
l
u
t
i
o
nt
ot
h
ec
o
n
t
i
n
u
o
u
sr
e
l
a
x
a
t
i
o
nm
o
d
e
li
s

'
t
i
l

X
;=2.
;x;=0.1563;

X:=O

1
"=48.239

T
h
eD
l
s
c
r
e
t
eP
r
o
b
l
e
m
:Thediscreteproblemsharesthesamemathematical
modele
x
c
e
p
tt
h
ev
a
l
u
e
sf
o
rt
h
ed
e
s
i
g
nv
a
r
i
a
b
lsw
i
l
lb
e
l
o
n
gt
oas
e
to
fd
i
s
c
r
e
t
e
v
a
l
u
e
s
.I
ti
sp
o
s
s
i
b
l
et
ou
s
ei
t
h
c
o
m
p
l
e
t
ee
n
u
m
e
r
a
t
i
o
no
rt
h
eBBs
e
a
r
c
ht
o
s
o
l
v
e p
r
o
b
l
e
m
.B
e
f
o
r
ep
r
o
c
e
e
d
i
n
gf
u
r
t
h
e
ri
nt
h
i
sd
i
r
e
c
t
i
o
ni
ti
su
s
e
f
u
lt
o
r
e
c
o
g
n
i
z
e t
h
er
e
a
s
o
nf
o
rd
i
s
c
r
e
t
eo
p
t
i
m
i
z
a
t
i
o
ni
st
oc
h
o
o
s
ea
no
f
f
t
h
e
s
h
e
l
f
I
b
e
a
mwhichw
i
l
lkeept
h
ec
o
s
tandp
r
o
d
u
c
t
i
o
nt
i
m
edown.S
e
v
e
r
a
lm
i
l
l
sp
r
o
v
i
d
e
i
n
f
o
r
m
a
t
i
o
nons
t
a
n
d
a
r
dr
o
l
l
i
n
gs
t
o
c
ke
ym
a
n
u
f
a
c
t
u
r
e
.R
e
f
e
r
e
n
c
e[
7
]i
su
s
e
dt
o
c
o
l
l
e
c
td
a
t
af
o
rt
h
i
se
x
a
m
p
l
e
.l
fas
t
a
n
d
a
r
dS-beami
sn
e
e
d
e
der
e
f
e
r
e
n
c
el
i
s
t
s
s
e
v
e
r
a
lbeamsw
i
t
ha
c

a
lv
a
l
u
e
sf
o
rt
h
ed
e
s
i
g
nv
a
r
i
a
b
l
e
sX
I
X
2

X
3
X
4

C
h
o
o
s
i
n
g
a
nI
b
e
a
mfromt
h
el
i
s
ta
l
s
oi
d
e
n
t
i
f
i
e
sa
l
lo
ft
h
ed
e
s
i
g
nv
a
r
i
a
b
l
e
s
.T
h
i
sbecomesa
s
i
n
g
l
e
v
a
r
i
a
b
l
ep
r
o
b
l
e
mt
h
ev
a
r
i
a
b
l
eb
e
i
n
gt
h
ep
a
r
t
i
c
u
l
a
ri
t
e
mfromt
h
el
i
s
to
f
b
e
a
m
s
.I
nt
h
i
sc
a
s
et
h
e
r
ea
r
enog
e
o
m
e
t
r
i
c
a
lc
o
n
s
t
r
a
i
n
t
sn
e
c
e
s
s
a
r
yandn
e
i
t
h
e
ra
r
e
es
i
d
ec
o
n
s
t
r
a
i
n
t
s
.A c
o
m
p
l
e
t
ee
n
u
m
e
r
a
t
i
o
nc
a
nbep
e
r
f
o
r
m
e
dont
h
ec
a
n
d
i
d
a
l
e
beamsfrom s
t
o
c
kl
i
s
t
andt
h
eb
e
s
tbeamc
a
ne
a
s
i
l
yb
ei
d
e
n
t
i
f
i
e
d
.Comparet
h
i
s
e
n
u
m
e
r
a
t
i
o
nt
os
o
l
v
i
n
gt
h
en
o
n
l
i
n
e
a
rm
a
t
h
e
m
a
t
i
c
a
lmodelw
i
t
hs
c
v
e
nf
u
n
c
t
i
o
n
a
l
c
o
n
s
t
r
a
i
n
t
s
.Whens
u
c
ha
no
p
p
o
r
t
u
n
i
t
ya
r
i
s
e
s
e
x
p
l
o
i
t
i
n
gt
h
ed
i
s
c
r
e
t
co
p
t
i
m
i
z
a
t
i
o
n
p
r
o
b
l
c
mi
sh
i
g
h
l
yrccommendcd.Ont
h
eo
t
h
e
rhandc
o
n
s
i
d
e
rad
i
f
f
e
r
e
n
tp
r
o
b
l
e
m
1
Ibemoldedi
nt
h
em
a
n
u
f
a
c
t
u
r
i
n
gf
a
ci
1
it
y
.Hcret
h
e
whereap
l
a
s
t
i
cI
b
e
a
mwi
c
o
n
t
i
n
u
o
u
sproblemn
e
e
d
st
ob
es
o
l
v
e
df
o
rt
h
eminimumc
o
st
.
Ex8_
4.mwi
I
lp
e
r
f
o
r
mac
o
m
p
l
e
t
ee
n
u
m
e
r
a
t
i
o
no
ft
h
i
se
x
a
m
p
l
e
.I
ti
n
c
l
u
d
e
sa
t
r
u
n
c
a
t
e
ds
t
o
c
kl
i
s
tfromw
h
i
c
ht
h
ef
m
a
ls
o
l
u
t
i
o
ni
so
b
t
a
i
n
e
d
.Th
ep
r
o
g
r
a
m
m
i
n
gi
s
s
t
r
a
i
g
h
t
f
o
r
w
a
r
d
.Th
eo
u
t
p
u
ti
nt
h
eCommandwindowi
s
Solution found
Optimum Beam is :W6X9
Mass of the Beam 71.543 (lb)
DepthWidthWeb ThicknessFlange Thickness (inches)
5.900 3.940 0.170 0.215
Constraints: Deflection Normal Stress Shear Stress
(<= 0)

-2422048221.600

-2468680.570

-86086.261

o
l
u
t
i
o
ni
ss
i
g
n
i
f
i
c
a
n
t
l
yh
i
g
h
e
ra
nt
h
eo
n
eo
b
t
a
i
n
e
di
nt
h
ec
o
n
t
i
n
u
o
u
s
N
o
t
et
h
a
tes
r
e
l
a
x
a
t
i
o
no
n
c
ea
g
a
i
ni
l
l
u
s
t
i
n
g t
h
ec
o
n
t
i
n
u
o
u
sr
e
l
a
x
a
t
i
o
np
r
o
v
i
d
e
sal
o
w
e
r
boundont
h
es
o
l
u
t
i
o
n
.

8
.
3
.
2 ZerneIntegerProgrammlng
ManyLPp
r
o
grams/
m
o
d
e
l
sh
a
v
es
p
e
c
i
a
lf
e
a
t
u
r
e
sw
h
i
c
ha
r
eu
s
u
a
l
l
ye
x
p
l
o
i
t
e
dt
b
r
o
u
g
h
s
p
e
c
i
a
la
l
g
o
r
i
t
h
m
s
.Ani
m
p
o
r
t
a
n
tc
l
a
s
so
fLPp
r
o
b
l
e
m
sr
e
q
u
i
r
eev
a
r
i
a
b
l
e
sb
eb
i
n
a
r
y
t
h
a
ti
syc
a
nh
a
v
eo
n
eo
ftwov
a
l
u
e
s
.Th
e
s
ep
r
o
b
l
e
m
sc
a
nb
es
e
tups
o v
a
l
u
e
s

344

8
.
3A
D
D
I
T
I
O
N
A
LEXAMPLES 345

D
I
S
C
E
T
EO
P
T
I
M
I
Z
A
T
lON

f
o
rev
a
r
i
a
b
l
e
sa
r
ez
e
r
oo
ro
n
e
--d
e
f
i
n
i
n
gaZe
ro-OneI
n
t
e
g
e
rP
r
o
g
r
a
m
m
i
n
g

a
p
i
1
B
u
d
g
e
t
i
n
gE
x
a
m
p
l
e
.
3C
T
a
b
l
e
8

P
r
o
b
l
e
m(
Z
I
P
)
.Theg
e
n
e
r
a
lf
o
r
mo
faZIPi
s

3'a

S
u
b
j
e
c
tt
o
: 3
X
1+2x
x
4+5
x
sS1
0
2+x3+2

&

JU

X
6[Oor1
]
X"

si
sn
o
tam
i
n
i
m
i
z
a
t
i
o
np
r
o
b
l
e
ma
sr
e
q
u
i db
yt
h
es
t
a
n
d
a
r
df
o
r
m
a
.
t
N
o
t
ei
Ao
r
i
t
h
m
s
: .
e
r
ea
r
es
e
v
e
r
a
1a
l
g
o
ms.1h
a
v
eb
e
e
nd
e
v
e
l
o
p
e
df
o
rs
o
l
v
i
n
gZ
I
P
m
o
d
e
l
s
.Th
e
ya
l
lu
s
ei
m
p
l
i
c
i
t
J
s
e
l
e
c
t
i
v
ee
n
u
m
e
r
a
t
i
o
nr
o
u
g
hB
r
a
n
c
ha
n
dBounds
h

t
o
g
e
t
h
e
rw
i
t
hc
o
n
t
i
n
u
o
u
sr
e
l
a
x
a
t
i
o
n
.However
t
h
eb
i
n
a
r
yn
a
t
u
r
eo
fev
a
r
i
a
b
l
e
si
s
e
x
p
l
o
i
t
e
dwhenc
o
m
p
l
e
t
i
n
gt
h
ep
a
r
t
i
a
ls
o
l
u
t
i
o
n
so
r v
e
r
s
i
n
gt
h
et
r
l
.l
eB
a
l
a
s
Ze
r
o
O
n
eA
d
d
i
t
i
v
eA
l
g
o
r
ithmlM
e
t
h
o
di
so
n
eo
f commend
i
q
u
e
sf
o
r s
c
1a
s
so
fp
r
o
b
l
e
m
s[
6
]
.Th
ea
l
g
o
r
i
t
h
mf
r
o
m s
a
m
er
e
f
e
r
e
n
c
ei
si
n
c
1
u
d
e
db
e
l
o
w
.

Ao
r
l
m:B
alasZer
o-O
neA
d
d
i
t
i
v
eAo.r~~h!! (
A
8
.
4
)
Thee
n
u
m
e
r
a
t
i
o
ni
sa
c
k
e
db
yt
w
ov
e
c
t
o
r
sUa
i
t
dX
.B
o
t
h o
fl
e
n
g
t
hn
.l
fap
a
r
t
i
a
l
s
o
l
u
t
i
o
nc
o
n
s
i
s
t
so
f1..
sa
s
s
i
g
n
e
di
no
r
d
e
r
t
h
ev
a
l
u
e
so
fXc
o
r
r
e
s
p
o
n
d
i
n
g
s
s
i
g
n
e
dv
a
r
i
a
b
l
s
a
r
e
O
o
r1
.
'
h
er
e
m
a
i
n
i
n
gv
i
a
b
l
e
'
e
ea
n
dh
a
v
et
h
ev
a
l
l
l
e
t
oea
1
.Th
ecomponenofUa
r
e

i
I
fk=J
a
ndi

s
c

o
m
p

I
l
e
m

n
I
t
h
asn

o
tb

c
n
c

o
n
ere

I
l
I
lk=

:
k~if =1o
r0a
n
di
t
s

oI

p
J
e
m
h
e
e
nc
o
n
s
i
d
d
ifk>s

S
t
e
p1
.I
fb:
2
:0

o
p
t
i
m
a
ls
o
l
o
ni
sX=
[
0
]
.O
t
h
e
r
w
i
s
U=
[
0
]X=

[J
].
f
.=00
X.=
X(
i
n
c
u
m
b
e
n
ts
o
l
u
t
i
o
n
)
S
t
e
p2
.C
a
l
c
l
l
l
a
t
e

M
a
x
i
m
i
z
e 20X1+3X2+30X3+5X4+IOxs+5X6

AuqJAUZJhueJ

nunu-Enu'A

40

xa

'
0

Example C
a
p
i
t
a
lBudgeting: l
eC
o
l
l
e
g
eo
fE
n
g
i
n
e
e
r
i
n
gi
sc
o
n
s
i
d
e
r
i
n
g
d
e
v
e
l
o
p
i
n
gau
n
i
q
u
eP
h
.
D
.p
r
o
g
r
a
r
ni
nM
i
c
r
o
s
y
s
t
e
m
si
nt
h
en
e
x
tt
h
r
l
y
e
a
r
s
.T
h
i
s
l
.S
i
xi
n
d
e
p
e
n
d
e
n
t
e
x
p
a
n
s
i
o
nc
a
l
l
sf
o
rs
i
g
n
i
f
i
c
a
n
ti
n
v
e
s
t
m
e
n
ti
nr
e
s
o
u
r
c
e
sa
n
dp
e
r
s
o
n
n
e
f
o
c
u
sa
r
e
a
sh
a
v
eb
e
e
ni
d
e
n
t
i
f
i
e
df
o
rc
o
n
s
i
d
e
r
a
t
i
o
n
a
1
1o
fw
h
i
c
hc
n
o
tb
ed
e
v
e
l
o
p
e
d
.
T
a
b
l
e8.
3p
I
1
C
S
e
n
t
st
h
ey
e
a
r
l
yc
o
s
ta
s
s
o
c
i
a
t
e
dw
i
t
hc
r
e
a
t
i
n
gt
h
ev
a
r
i
o
u
sg
r
O
l
l
p
So
v
e
rt
h
e
ey
r
s
.l
el
a
s
tl
i
n
ei
net
a
b
l
er
e
p
r
e
s
e
n
t
st
h
ea
v
a
i
l
a
b
l
eb
l
l
d
g
e
t(
i
nm
i
l
l
i
o
n
so
f
d
o
l
l
a
r
s
)p
e
ry
e
a
ra
l
l
o
t
t
e
dt
ot
h
i
sd
e
v
e
l
o
p
m
e
nt
.
l
el
a
s
tc
o
l
u
m
ni
sev
a
l
l
l
e(
i
n
$100

0
0
0
)t
h
eg
r
o
u
pi
se
x
p
e
c
t
e
dt
og
e
n
e
r
a
t
ei
nt
h
en
e
x
t5y
e
a
r
s
.
X
6
)
o
n
ef
o
re
a
c
h
Thec
a
p
i
t
a
lb
u
d
g
e
t
i
n
gp
r
o
b
l
e
ml
l
s
e
ss
i
xb
i
n
a
r
yv
a
r
i
a
b
l
e
s(Xh.
i
t
e
mi
net
a
b
l
e
.TheZ
I
Pp
r
o
b
l
e
mc
a
nb
ef
o
n
u
l
a
t
e
da
s

+3

I
nea
b
o
v
ep
r
o
b
l
e
m
t
h
eunknownv
a
r
i
a
b
l
e
sa
r
en
o
tb
i
n
a
r
yv
a
r
i
a
b
l
e
sb
u
t ei
n
t
e
g
e
r
s
n
o
n
e
e
l
e
s
s
.

'
em
t
e
g
e
r
s

Z+

X"X
2

X
3
X
4

5b

S
l
l
b
j
e
c
t

: X
X
2+I
O
x
3+2
5
x
4=
b
1+5

Mi
n
i
m
i
z
e X
1+x2+x3+X4

++

'
0

Va
Ju
e

1
0

E
x
c
e
p
tf
o
rt
h
eb
i
n
a
r
ys
i
d
c
o
n
s
t
r
a
i
n
t
so
ned
e
s
i
g
nv
a
r
i
a
b
l
e
si
sr
e
p
r
e
s
e
n
t
se
m
a
t
h
e
m
a
t
i
c
a
lm
o
d
e
lf
o
ra
n LPp
r
o
b
l
e
m(
t
h
es
t
a
n
d
a
r
dm
o
d
e
l[1r
e
se
q
u
a
l
i
t
y
c
o
n
s
i
n
t
so
n
l
y
)
.Mosta
l
g
o
r
i
t
h
m
sa
s
s
u
m
e M p
o
s
i
t
i
v
e
.I
fa
n
yc
ji
sn
e
g
a
t
i
v
e

t
h
e
n
ss
l
l
b
s
t
i
t
l
l
t
e
df
o
rx
Ji
ne
p
r
o
b
l
e
m
.
1 i
S
e
v
e
r
a
lv
a
r
i
e
t
i
e
so
fo
p
t
i
m
i
z
a
t
i
on
J
d
e
c
i
s
i
o
np
r
o
b
l
e
m
si
n
v
o
l
v
eb
i
n
a
r
yv
a
r
i
a
b
l
e
s
.They
c
1u
d
et
h
ek
n
a
p
s
a
c
k(
ap
u
r
eILPw
i
t
has
i
n
g
l
em
a
i
nc
o
n
s
t
r
a
i
n
t
)

c
a
p
i
t
a
lb
l
l
d
g
e
t
i
n
g(
a
i
n
m
u
l
t
i
d
i
m
e
n
s
i
o
n
a
lk
n
a
p
s
a
c
kp
r
o
b
l
e
m
)a
s
s
e
m
b
l
yl
i
n
eb
a
l
a
n
c
i
n
gm
a
t
c
h
i
n
gs
e
t
c
o
v
e
r
i
n
ga
n
df
a
c
i
l
i
t
yl
o
c
a
t
i
o
n
a
n
ds
oo
n
.Somee
x
a
m
p
l
e
sa
r
ed
e
t
a
i
l
e
db
e
l
o
w
.

Example aKnapsackProblem:M
i
n
i
m
i
z
e e n
l
l
m
b
e
ro
fp
e
n
n
i
e
s(
X
1
)
n
i
c
k
e
l
s(
x
2
1
d
i
m
e
s(
X
3
)
a
n
dq
u
a
r
t
e
r
s(
X
4
)t
op
r
o
v
i
d
ec
o
r
r
e
c
tc
h
a
n
g
ef
o
rbc
e
n
t
s
:

3 3
nU44

B
u
d
g
e
t

=0or1 j=12...n

Xj

-E

i=1

.
.
.
m

3 TZJZO

S
y
s
t
e
m
s
Mi
c
r
o
s
y
s
t
e
m
s
S
o
f
t
w
a
r
e
M
o
d
e
l
i
n
g
M I
s
P
h
o
t
o
n
i
c
s

Y
e
a
r
3

Y
e
a
r

Y
e
a
r
l
321250

J
X

}
=
1

nT'MMF

S
u
b
j
e
c
tt
o
:

G
r
o
u
p

I
t
e
m

M
i
n
i
m
i
z
e f
(
X
)
::
2

2
:a

y=b
j-

Ji
st
h
es
e
to
fa
s
s
i
g
n
e
dv
a
r
i
a
b
l
e
s

l~

~~

OISCRETEOPTIMATION

8
.
3 AOO
l
T
lONALEXAMPLES

Ti--vMW/

346

y=minY
i
' i
=1

2
.
.

I
=
Lfj

347

S
u
b
j
e
c
t
t
o
: -8x1- l
l
X
2-9
X
3-1
8
x
4S-12

-2x.-2x
x3-14x4S-14
27

9
x
.-6
x
x
4s
-lO
2- 3X3- 6

X
.
.X
2
'X
3
.X
4E [
01
)

I
f
y O d 1
<
1
' then I=

J and X XGot
oS
t
e
p6
0e
r
w
i
s
ec
o
n
t
i
n
u
e
.C
r
e
a
t
eas
u
b
s
e
tTo
f-eev
a
r
i
a
b
l
e
sd
e
f
i
n
e
da
s
S
t
e
p3
T=ij:f+<1

a
i
j<0 f
o
ri
s
u
c
haty/<O)
e
m
p
t
ys
e
t
)t
h
e
nGot
oS
t
e
p6
E
I
s
ec
o
n
t
i
n
u
e
1fT= (
t
:
S
t
e
p4
.I
n
f
e
a
s
i
b
i
l
i
t
yt
e
s
I
fis
u
c
ht
h
a
tY
i<0a
n
d

5UH
Mmrd
li

T
h
e
o
u
t
p
u
tomBalas.m:

Solution exists

-------

The variab1es are:


1

The objective function 52.00


constraints -Va1ue and Right hand side
-26 -12
-16 -14
-15 -10

Lmin(Oaij)<0

Y
i-

jGT

Got
oS
t
e
p6
E
I
s
ec
o
n
t
i
n
u
e
Ia
sb
r
a
n
c
h
i
n
gt
e
s
t
:F
o
re
a
c
hf
r
e
ev
a
r
i
a
b
l
ec
r
e
a
t
et
h
es
e
t
S
t
e
p5
.Ba

Th
i
sc
o
r
r
e
s
p
o
n
d
st
ot
h
es
o
l
u
t
i
o
ni
d
e
n
t
i
f
i
e
di
nR
e
f
e
r
e
n
c
e6
.
CompleteEnumeratlon:Th
eB
a
l
a
sa
Ig
o
r
i
t
h
mp
e
r
f
o
r
m
sp~a~ e
n
u
m
e
r
a
t
i
o
n
.The
c
o
m
p
l
e
t
eo
re
x
h
a
u
s
t
i
v
ee
n
u
m
e
r
a
t
i
o
n(
n
.
)f
o
rnb
i
n
a
r
yv
a
r
i
a
b
l
e
si
sg
i
v
e
nbyt
h
er
e
l
a
t
i
o
n

=(
i
:Y
j-a
i
j<0
)
I
fa
I
Is
e
t
sMja
r
eemptyGot
oS
t
e
p6
.O
t
h
e
r
w
i
s
eCa
Ic
u
l
a
t
ef
o
r e
e

ne=2n

v
a
r
i
a
b
l
e

T
y
p
i
c
a
l
l
y

e
r
I
t
e
E
p
r
i
s
ed
e
c
i
s
i
o
nmakingi
n
v
o
l
v
e
ss
e
v
e
r
a
lh
u
n
d
r
e
dv
a
r
i
a
b
l
e
sa
n
dl
i
n
e
a
r
m
o
d
e
l
s
.Ont
h
e0
e
rh
d

e
n
g
i
n
r
i
n
gp
r
o
b
l
e
m
sc
h
a
r
a
c
t
e
r
i
z
e
dbyan
o
n
l
i
n
e
a
rmodel
mayhavea
b
o
u
t20b
i
n
a
r
yv
a
r
i
a
b
l
e
sm
q
u
i
r
i
n
gne=l
O48

5
7
6
e
n
u
m
B
E
a
I
i
o
n
s
T
h
i
si
s
n
o
b
i
'
d
e
a
lw
i
t
hd
e
s
k
t
o
pr
e
s
o
u
r
c
e
stoday.KEepi
nmind compl
t
ee
n
u
m
e
r
a
t
l
o
n
a
v
o
i
d
sa
n
yr
e
l
a
x
a
t
i
o
ns
o
l
u
t
i
o
n(
p
u
r
e
l
yd
i
s
c
r
e
t
ep
r
o
b
l
e
m
so
n
l
y
)
.

= a
i
j
)
ieM}

w
h
e
r
e =0i
f
t
h
es
e
t i
se
m
p
t
y
.Addt
ot
h
ec
u
r
r
e
n
tp
a
r
t
ia
Is
o
l
u
t
i
o
nt
h
e
Got
oS
t
e
p2
Ub
ys
e
t
t
i
n
gt
h
es
i
g
no
ft
h
er
i
g
h
o
s
tp
o
s
i
t
i
v
ecomponen
.
tAl
I
S
t
e
p6
.Modi
e
l
e
m
e
n
t
st
ot
h
er
i
g
h
ta
r
eO
.Retumt
oS
t
e
p2
I
ft
h
e
r
ea
r
enop
o
s
i
t
i
v
ee
l
e
m
e
n
t
si
nU
ee
n
u
m
e
r
a
t
i
o
ni
sc
o
m
p
l
e
t
e
.Th
e

;
__.dI
o
p
t
i
maIs
o
l
u
t
i
o
ni
si
nt
h
ei
n
c
u
m
b
e
n
tv
e
c
t
o
rX
.Theo
b
j
e
c
t
i
v
ei
si
n
i
s
t
h
e
r
ei
snof
e
a
s
i
b
l
es
o
l
u
t
i
o
n
.
V i
a
b
l
eX
j
a
tm
a
x
i
m
i
z
e
s

Balas.mi
sem
f
i
l
e i
m
p
l
e
m
e
n
t
st
h
eBa
Ia
sa
l
g
o
r
i
t
h
m
.I
twast
r
a
n
s
l
a
t
e
df
r
o
ma
P
a
s
ca
Ip
r
o
g
r
a
ma
v
a
i
l
a
b
l
ei
nR
e
f
e
r
e
n
c
e6
.I
ti
sap
r
o
g
r
a
ma
n
dp
r
o
m
p
t
st
h
eu
s
e
rf
o
rt
h
e
p
r
o
b
l
e
mi
n
f
o
r
m
a
t
i
o
n
n
m
A
b
c
.l
et
e
s
tp
r
o
b
l
e
mg
i
v
e
ni
nR
e
f
e
r
e
n
c
e6i
si
n
c
l
u
d
e
d
b
e
l
o
w
.
Exampk8.5

M
i
n
i
m
i
z
e

f
(
X
)
:l
O
x
.+1
4
x
2+2
l
x
3+4
2
x
4

3i-ijii

.
rJ
f
.
r

Ca
l
I
Simplex.m: i
saf
u
n
c
t
i
o
nm
f
i
l
et
h
a
tc
b
eu
s
e
dt
os
o
l
v
eLPp
r
o
g
r
a
m
s
.I
t
cmbeusdtoobtains
o
l
u
t
i
o
n
st
oLPm
l
a
x
a
t
i
o
nm
o
d
e
l
sd
u
r
i
n
gt
h
eBBS
E
a
r
c
h
.
I
ti
s
I
g
o
r
i
t
h
ma
v
a
i
l
a
b
l
ei
nR
e
f
e
r
e
n
c
e6
.I
ti
st
r
a
n
s
l
a
t
e
df
r
o
maP
a
s
c
a
lp
r
o
g
r
a
m
.
b
a
s
e
dont
h
ea
l
em
a
t
h
e
m
a
t
i
c
a
Imodeli
s
M
i
n
i
m
i
z
e

T
x
c

S
u
b
j
e
c
tt
o
: [
A
]]=[
b
]
x~o

S
l
a
c
kv
a
r
i
a
b
l
e
sa
r
ei
n
c
J
u
d
e
di
nec
o
e
f
f
i
c
i
e
n
tm
a
t
r
i
x[
A
]
.G
r
e
a
t
e
r o
re
q
u
a
lt
o
c
o
n
s
t
r
n
t
sc

b
es
e
t up w
i
t
hn
e
g
a
t
i
v
es
l
a
c
kv
a
r
i
a
b
l
e
sa
st
h
ep
r
o
g
r
a
mw
i
l
l

348 O
I
S
C
R
E
T
EO
P
T
I
M
I
Z
A
T
I
O
N

PROBLEMS 349

a
u
t
o
m
a
t
i
c
a
l
l
ye
x
e
c
u
t
ead
u
a
l
p
h
a
s
es
i
m
p
l
e
xm
e
t
h
o
d
.A
n
o
t
h
e
ra
l
t
e
m
a
t
i
v
ei
st
o e
LP p
r
o
g
r
a
m
sp
r
o
v
i
d
e
d by MATLAB. The f
u
n
c
t
i
o
ns
t
a
l
e
m
e
n
tf
o
ru
s
i
n
gC
a
l
l
Slmplex.mi
s
]
function [xsol f

CallSimplex(A
B
C
M
N)

.B
.C.M
.Nhavet
ob
ed
e
f
i
n
e
d
.Mi
s numbcro
fc
o
n
s
t
r
a
i
n
t
s
Tou
s
et
h
ef
u
n
c
t
i
o
n
.A
st
h
enumbero
fd
e
s
i
g
nv
i
a
b
l
e
s
.
andNi

8
.
1
0 Youa
r
et
h
eownero
fa
na
u
t
o
m
o
b
i
l
ed
e
a
l
e
r
s
h
i
pf
o
rBrandX c
a
r
s
.S
e
tupaDP
f
o
rm
o
n
t
h
l
yi
n
v
e
n
t
o
r
yc
o
n
t
r
o
lb
a
s
e
donc
a
r
se
x
p
e
c
t
e
d1
0b
es
o
l
d
.h
i
g
ho
r
d
e
r
i
n
g
c
o
s
t
s
.u
n
i
tc
o
s
t
sf
o
rt
h
ec
a
r
.andi
n
v
e
n
t
o
r
yc
o
s
t
sf
o
re
x
c
e
s
sc
a
r
s
.S
o
l
v
ei
tu
s
i
n
g
t
h
ecodei
nEx8_3.m.
8
.
1
1S
o
l
v
et
h
cf
o
l
l
o
w
i
n
gu
s
i
n
gB
a
l
a
sa
l
g
o
r
i
t
h
m
u
s
i
n
gc
o
m
p
l
e
t
cc
n
u
m
e
r
a
t
i
o
n

M
i
n
ir
n
i
z
e

f(X)=X
I+X2+X3

S
u
b
j
e
c
t 3 x I+4X2+5
x
38

2
x1+5
x
2+3X3~ 3
REFERENCES

1
.F
l
e
t
c
h
e
r
.R
.
P
r
a
c
l
i
c
a
lM
e
l
h
o
d
so
l
O
p
'
i
m
i
z
a
l
i
o
n
W
i
l
e
y
NewY
o
r
k1
9
8
7
.
2
.P
a
p
a
n
b
r
o
sP
.Y
.a
n
dW
i
l
d
eJ
.D
.P lCi
p
l
e
s0
1Opli~I D
e
s
i
g
n
M
o
d
e
l
i
n
gand
U
l
a
l
i
o
n
.C
a
m
b
r
i
d
g
eU
n
i
v
e
r
s
i
t
yPress

NewY
o
r
k
1
9
8
8
.
Con
3
.R
a
r
d
i
n
R
.L
.
O
p
l
i
m
i
l
a
l
i
o
ni
nO
p
e
r
a
l
i
o
n
sR
e
s
e
a
r
c

J
lP
r
cn
l
i
c
e
H
a
l
l
E
n
g
l
e
w
o
o
dC
I
f
s
NJ
1
9
9
8
.
4
.B
e
l
l
m
a
n
R
.
D
y
n
a
m
i
cP
r
o
g
r
a
n
u
n
g

P
r
i
n
c
e
t
o
nU
n
i
v
e
r
si
L
yP
r
e
s
s
P
r
in
c
e
t
o
n
NJ1
9
5
7
.
5
.W
h
i
t
e
D
.J
.DynamicP
r
o
g
r
a
m
m
i
n
g
H
o
l
d
e
n
D
a
y
S
a
nF
r
a
n
c
i
s
c
o1
9
6
9
.
6
.S
y
s
l
oM. M.Deo
N
.a
n
dK
o
w
a
l
i
k
J
.S
.D
l
s
c
r
e
l
eO
p
l
i
m
i
z
a
t
i
o
nA
l
g
o
r
l
l
h
m
s

n
t
i
c
e
H
a
l
l
E
n
g
l
e
w
o
o
dC
l
i
f
f
s
N
J
.1
9
8
3
.
7
.R
y
e
r
s
o
nS
l
o
c
kS
l
e
e
lL
i
s
l
.J
o
s
e
p
hT
.R
y
e
r
s
o
n
I
n
c
.
1
9
9
5
.

PROBLEMS

.
1i
n
t
oad
i
s
c
r
e
t
eo
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
e
mands
o
l
v
eby
8
.
1 T
r
a
n
s
l
a
t
eExample 1
e
x
h
a
u
s
t
i
v
ee
n
u
m
e
r
a
t
i
o
n
.
8
.
2 T
r
a
n
s
l
a
t
eExample 1
.
2 1
0ad
i
s
c
r
e
t
eo
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
e
mands
o
l
v
eby
e
x
h
a
u
s
t
i
v
ee
n
u
m
e
r
a
t
i
o
n
.
8.
3 T
r
a
n
s
l
a
t
eExample 1
.3i
n
t
oad
i
s
c
r
e
t
eo
p
t
i
r
n
i
z
a
t
i
o
np
r
o
b
l
e
mands
o
l
v
eby
e
x
h
a
u
s
t
i
v
ee
n
u
m
e
r
a
t
i
o
n
.
8
.
4 Im
p
l
e
m
e
n
t a m
u
l
t
i
s
t
a
g
e e
x
h
a
u
s
t
i
v
ee
n
u
m
e
r
a
t
i
o
n f
o
rt
h
c c
o
n
t
i
n
u
o
u
s
t
i
o
no
fExample1
.
1
.
o
p
t
i
m
i
8.
5 T
r
a
n
s
l
a
E
xample1
.
1i
n
t
oad
i
s
c
r
e
t
eo
p
t
i
r
n
i
z
a
t
i
o
nproblemands
o
l
v
ebyb
r
a
n
c
h
db
oundm
e
t
h
o
d
.
i
o
nproblemands
o
l
v
ebyb
r
a
n
c
h
8
.
6 T
r
a
n
s
l
a
t
eExample1
.
2i
n
t
oad
i
s
c
r
e
t
eo
p
t
i
m
i t
andboundm
e
t
h
o
d
.
8
.
7 T
r
a
n
s
l
a
t
eExample1
.
3i
n
t
oad
i
s
c
r
e
t
eo
p
t
i
m
i
z
a
t
i
o
nproblemands
o
l
v
ebyb
r
a
n
c
h
andboundm
e
t
h
o
d
.
8
.
8 I
d
e
n
t
i
f
ya
p
p
r
o
p
r
i
a
t
es
t
kavailabilityfortheflagpoleprobleminChapter4and
s
o
l
v
et
h
ec
o
r
r
e
s
p
o
n
d
i
n
gd
i
s
c
r
e
t
eo
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
e
m
.
8
.
9 S
o
l
v
eExample1
.
3u
s
i
n
g s
t
o
c
ki
n
f
o
r
m
a
t
i
o
na
v
a
i
l
a
b
l
ei
nEx
8_
4.m.

X.X2.X3[01
]

8
.
1
2P
r
o
g
r
1a
ne
x
h
a
u
s
t
i
v
ee
n
u
m
e
r
a
t
i
o
nt
e
c
h
n
i
q
u
cf
o
rz
e
r
ooneI
Pu
s
i
n
gMATLAB.
S
o
l
v
eP
r
o
blem8
.
1
0
.
8
.
1
3I
si
tp
o
s
s
i
b
l
et
os
o
l
v
et
h
ec
a
p
i
t
a
lb
u
d
g
c
t
i
n
ge
x
n
p
l
e
?I
fsowhati
st
h
es
o
l
u
t
i
o
n
u
s
i
n
gc
o
m
p
l
e
t
ee
n
u
m
e
r
a
t
i
o
n
?
8
.
1
4S
o
l
v
ec
a
p
i
t
a
lb
u
d
g
e
t
i
n
gexampleu
s
i
n
gB
a
l
a
sp
r
o
g
r
a
m
.
8
.
1
5C
r
e
a
t
ea
nexamplewhich i
n
v
o
l
v
e
smakingb
i
n
a
r
yc
h
o
i
c
c
sr
e
g
a
r
d
i
n
gy
o
u
r
p
a
r
t
i
c
i
p
a
t
i
o
ni
nc
o
u
r
s
ea
n
de
x
t
r
a
c
u
r
r
i
c
u
l
a
ra
c
t
i
v
i
t
i
c
s1
0maximizey
o
u
rg
r
a
d
e
s
.
S
o
l
v
ei
.
t

9
.
1 PAOBLEMD
E
F
I
N
I
T
I
O
N 3
5
1

~uess) Whi
l
ei
tm
a
k
e
ss
e
n
s
et
ol
o
o
kf
o
rt
h
eg
l
o
b
a
1optimum i

5n
om
e
t
h
o
dt
o

L
E
n
6

f
v
f J
c
h
a
r

t
l
t
c
i
n
u
o
1

u
1
s
c

.
'

0
1
m

v
.

F
e
x
;
E
r
o

b
o
l
e
m
s g
u
a
r
a
n
a

El
I
s

l
u

t
i
i
s d
i

f
f

i
i
c

a
E
t

t
h

a
t
S
l
ad
r
e
p
r
e
s
e
n
t

s
o

n
lyami

n
i
s

u
r
P
i
o
np
b
l
e

m
s
n
o
t
n

c
e
s
s

l
yc
o

n
v
T
h
e
r
e
f
o
r
e

i
nt

h
el

t
t

r
b
o
t
ho
f
t

he
s
em
e
t
h
o
d
sa
r
ee
x
p
c
t
e
d
~oproduce s
o
l
u
t
i
o
n
sc
1
0
s
e~ot
h global optimum
s
o
l
u
t
i
o
n

r
a
t
h
e
rt
h
eg
l
o
b
a
1s
o
l
u
t
i
o
ne
m
s
e
l
v
e
s
'
I
1
1
i
si
sn
o
t
t
os
u
g
g
e
s
t
t
h
a
ti
n
v
e
s
t
i
g
a
t
i
o
n
l
u
u
O
n
i
sawa
o
f ei
ft
h
ep
o
s
s
i
b
i
l
i
t
yo
fo
n
e n
o
t
b
l
i
s
h
e
d
o
f
t
h
e
g
l
o
b o

;2
2
2
2
2
1
;
t
i
:
:
;
f
r
;
;
L
P
2
2
2
J
Z
;
2
1

GLOBALOPTIMIZATION

T
h
i
sc
h
a
p
t
e
ri
n
t
r
o
d
u
c
e
st
w
oo
ft
h
em
o
s
tp
o
p
u
l
a
rt
e
c
h
n
i
q
u
e
sf
o
rf
i
n
d
i
n
gt
h
eb
e
s
tv
a
l
u
e
e
c
t
i
v
ef
u
n
c
t
i
o
n

a
l
s
or
e
f
e
r
r
e
dt
oa
st
h
eg
I
o
b
a
lo
p
t
i
m
u
ms
o
l
u
t
i
o
n
.Theya
r
e
f
o
rt
h
e0
o
r
i
t
h
m(
G
A
)
.Then
a
m
e
sa
r
en
o
t
S
i
m
u
l
a
t
e
dA
n
n
e
a
l
i
n
g(
S
A
)a
n
dt
h
eG
e
n
e
t
i
cAlg
a
t
u
r
a
l
p
r
o
c
e
s
s
e
s
.SAi
sb
a
s
e
d
a
c
c
i
d
e
n
t
a
1s
i
n
c
et
h
ei
d
e
a
sb
e
h
i
n
dt
h
e
ma
r
ed
e
r
i
v
e
domn
i
nd
e
f
i
n
e
dc
r
y
s
t
a
1
l
i
n
es
t
r
u
c
t
u
r
e
sb
a
s
e
do
nminimum
o
nc
o
o
l
i
n
go
fm
e
t
a
l
st
oob
e
n
e
si
n
p
o
t
e
n
t
ia
1e
n
e
r
g
y
.GAi
sb
a
s
e
dont
h
ec
o
m
b
i
n
a
t
i
o
na
n
dr
e
c
o
m
b
i
n
a
t
i
o
no
feg
t
h
eb
i
o
l
o
g
i
c
a
1s
y
s
t
e
ml
e
a
d
i
n
gt
oi
m
p
r
o
v
e
dDNAs
e
q
u
e
n
c
e
so
rs
p
e
c
i
e
ss
e
l
e
c
t
i
o
n
.I
n
h
e
ya
r
m
ore l
i
k
e ac
o
n
t
i
n
u
a
t
i
o
no
fC
h
a
p
t
e
r8 (
D
i
s
c
r
e
t
e
a
c
t
u
a
la
p
p
l
i
c
a
t
i
o
nt
e
c
a
u
s
et
h
e
s
em
e
t
h
o
d
sc
a
nb
eu
s
e
df
o
rc
o
n
t
i
n
u
o
u
sa
n
dd
i
s
c
r
e
t
e
O
p
t
i
m
i
z
a
t
i
o
n
)b
e
ys
h
a
r
ea
ni
m
p
o
r
t
a
n
tc
h
a
r
a
c
t
e
r
i
s
t
i
cw
i
t
h
c
a
t
e
g
o
r
i
e
so
ft
h
eo
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
e
m
s
.Th
o
t
h
e
ro
p
t
i
m
i
z
a
t
i
o
nt
e
c
h
n
i
q
u
e
si
n e
ya
r
ep
r
i
m
i
l
ys
e
a
r
c
ht
e
c
h
n
i
q
u
e
s
t
h
e
y
ys
e
a
r
c
h
i
n
gt
h
ed
e
s
i
g
ns
p
a
c
ef
o
rt
h
es
o
l
u
t
i
o
n
.U
n
l
i
k
ee
i
d
e
n
t
i
f
yeoptimumb
g
r
a
d
i
e
n
t
b
a
s
e
dm
e
t
h
o
d
s

t
h
e
s
et
e
c
h
n
i
q
u
e
sa
r
el
a
r
g
e
l
yh
e
u
r
i
s
t
i
c

g
e
n
e
r
a
1
1
yi
n
v
o
l
v
el
a
r
g
e
o
u
n
t
so
fcomput
o
n

a
n
di
n
v
o
l
v
esomes
t
a
t
i
s
t
i
c
a
lr
e
a
5
0
n
i
n
g
.G
l
o
ba
1o
p
t
i
m
i
z
a
t
i
o
n
i
sb
e
i
n
gp
u
r
s
u
e
ds
e
r
i
o
u
s
l
yt
o
d
a
ya
t
c
t
i
n
gas
i
g
n
i
f
i
c
a
n
ta
m
o
u
n
to
fnewr
e
s
e
a
r
c
ht
oe
a
r
e
ao
fa
p
p
l
i
e
do
p
t
i
m
i
z
a
t
i
o
na
f
t
e
ral
o
n
gh
i
a
t
u
s
.As
i
g
n
i
f
i
c
a
n
td
r
a
w
b
a
c
kt
omanyo
f
t
h
e
s
e
t
e
c
h
n
i
q
u
e
s
i
se
yr
e
q
u
i
r
ee
m
p
i
r
i
c
a
lt
u
n
i
n
gb
a
s
e
do
nt
h
ec
l
a
s
so
fp
r
o
b
l
e
m
s
b
e
i
n
gi
n
v
e
s
t
i
g
a
t
e
d
. Th
e
r
ei
sa
l
s
o no s
y way t
od
e
t
e
c
m
i
n
et
e
c
h
n
i
q
u
e
l
m
e
t
e
r
st
oi
m
p
l
e
m
e
n
t a
u
t
o
m
a
t
i
co
p
t
i
m
i
z
a
t
i
o
nt
e
c
h
n
i
q
u
e
.
p
r
o
b
l
e
m
s
e
n
s
i
t
i
v
ep
a
r
Theg
l
o
b
a
lo
p
t
i
m
u
mc
a
n
n
o
tr
e
a
l
lyb
ec
h
a
r
a
c
t
e
r
i
z
e
d

b
e
i
n
gd
i
f
f
e
r
e
n
tf
r
o
mt
h
el
o
c
a
l
a
t
t
e
ri
sr
e
f
e
r
e
n
c
e
dwheno
n
l
yap
a
r
to
ft
h
ed
e
s
i
g
ns
p
a
c
e(
a
sop
s
e
dt
o
o
p
t
i
m
u
m
.Thel
ec
o
m
p
l
e
s
p
a
c
e
)i
sv
i
e
w
e
di
nt
h
en
e
i
g
h
b
o
r
h
o
o
do
fap
a
r
t
i
c
u
l
a
rd
e
s
i
g
np
o
i
n
.
tThe
e
s
t
a
b
l
i
s
h
e
da
n
du
s
e
di
ne
a
r
l
i
e
rc
h
a
p
t
e
r
s(
f
o
r
s
t
a
n
d
a
r
dKuhnT
u
c
k
e
rc
o
n
d
i
t
i
o
n
s
w
i
l
li
d
e
n
t
i
f
yel
o
c
a
1optimumo
n
l
y
.T
h
i
si
sa
1
s
omore
c
o
n
t
i
n
u
o
u
sp
r
o
b
l
e
m
s
)
e
m
p
h
a
t
i
ci
nn
u
m
e
r
i
ca
1i
n
v
e
s
t
i
g
a
t
i
o
n
so
fc
o
n
t
i
n
u
o
u
sp
r
o
b
l
e
m
sw
h
e
r
et
h
ea
l
g
o
r
i
t
h
m
sa
r
e
a
l
u
e
sc
l
o
s
et
ow
h
e
r
et
h
:
ya
r
es
t
a
r
t
e
d(
i
n
i
t
i
a
l
e
x
p
e
c
t
e
dt
omovet
o
w
a
r
dt
h
eop mumv

350

0
p
t
i
I
I
l
i
z
a
G
o
nt
e
c
h
n
i
q
u
e
sa
r
es
i
m
p
l
et
oi
m
p
l
e
m
e
n
tw
h
i
l
e mquutng e
x
h
a
u
s
t
i
V
E
c
a
l
c
u
l
a
t
i
o
n
s
.
T
h
i
sc
a
np
o
t
e
n
u
d
l
ys
o
a
kupa
l
lo
fei
d
l
et
i
m
o
nt
h
ep
e
r
s
o
n
a
1d
e
s
k
t
o
p
c
o
m
p
u
t
i
n
gr
e
s
o
u
r
c
e
l
e
n
gi
t m
e
d
i
a
t
ea
a
c
t
i
v
e
n
e
s
s
.
s
p
o
u
s
e
d
23et
e
c
h
n
i
q
u
e
sa
n
dt
h
e
i
ra
s
c
u
s
s
I
o
na
x
eV
E
r
yb
a
s
I
Cf
o
l
l
o
w
i
n
gEemee
i
n a
p
t
e
r8
.e d
e
ri
se
n
c
o
u
r
a
g
e
dt
o i
tj
o
u
r
n
a
lp
u
b
l
i
c
a
t
i
o
n
s m
o
n
i
t
o
r
m
a

U
1
I

E
z
z
Eo
d
f h

n
i

q
u

s
.l
n

sc
h
a
p
t

t
s
o
m

ei
s
s

s
s
r

e
g

gg

1
0


u
o
n
'
e
p
n

t
f

0
1

1
1

9
.
1 PROBLEMDEFINITION

'
h
f
o
l
l
o
w
i
n
g
d
c
v
c
l
o
p
m
e
n
t
o
r
g
l
o
b
a
lo
p
t
i
m
i
z
a
d
o
ni
sb
a
s
donu
n
c
o
n
s
t
r
a
i
n
e
d
c
o
n
t
i
n
u
o
u
s
z
a
t
i
o
n
.S
o
l
v
i
n
gc
o
n
s
i
n
E
d
p
m
b
l
e
m
s
d
G
e
m
o
t
a
p
p
e
a
r
t
o
h
a
v
e
e
v
o
l
v
e
dtoacommon
o
p
t
e
c
di
sn
o
ti
l
l
u
s
.
t
e
di
nt
h
i
sb
o
o
k
.N
o
t
ea
tc
o
n
s
i
n
e
dp
r
o
b
l
e
m
sc
a
nb
e
i
n
e
dp
r
o
b
l
e
mu
s
i
n
gap
e
n
a
l
t
yf
o
r
m
u
l
a
t
i
o
n(
I

l
a
p
7
)
l
e
c
o
n
v
e
r
t
e
dt
oa
nu
n
c
o
n
s
p
r
o
b
l
e
mc
a
nb
ed
e
f
i
n
e
da
s

Mi
n
i
m
i
z
e

f
[
X
)
;[
X
)
n

S
u
b
j
e
c
t
t
o
: x! Xj S. X~

(
9
.
1
)
(
9
.
2
)

(
o
n
l
yo
n
e um)eo
p
o
nc
o
u
l
d
l
yb
e
t
a
b
l
i
s
h
e
db
ya
n
yo
ft
h
ea
ifO
p
r
i
a
t
t
e
c
h
n
i
q
u
e
su
s
e
di
nt
h
eb
ksincetheloca
1minimum
i
sa
l
s
o
eg
l
o
b
a
lm
i
n
i
m
u
m
.To
v
e
s
t
i
g
a
t
h
eg
l
o
b
a
lmin
ni
ti
st
h
e
r
e
f
o
r
ee
x
p
e
c
t
e
d
u
ldbem
y
l
o
n
i
n
i
r
r
1
5

s
e
v
e
r
a
lo
f nhavingd
r
e
n
tv
a1r

o
b
j
e
c
t
i
v
e
. l
e
a
s
tamongt
h
e
mi
st
h
e
neg
l
o
b
a
ls
o
l
u
t
i
o
n
.L
o
c
a
1minimums

i
nese
o
fc
o
n
t
i
n
u
o
u
sp
r
o
b
l
e
m
s
a
r
eo
s
e s
a
f
yeKTc
o
n
d
i
t
i
o
n
s
dw
i
l
lben
u
m
e
r
i
c
a
l
l
y
d
e
t
e
r
m
i
j
f
es
t
a
r
t
i
n
g sw
i
ne
n
e
i
bo
o
odof p
m
inim
u
m
.T
h
i
s
d
e
f
i
n
i
t
i
o
nmayb
ea
p
p
l
i
e
dt
od
i
s
c
r
e

eop
z
a
t
i
o
np
r
o
b
l
e
m
st
h
a
tu
s
c
o
n
t
i
n
u
o
u
sr
e
l
a
x
a
t
i
o
n
ob
i
nes
o
M
U
o
n
F
o
rd
i
s
c
m
t
ep
r
o
b
l
e
m
su
s
i
n
gc
o
m
p
l
e
e
n
u
m
e
r
a
t
i
o
n
t
h
eg
l
o
b
a
l
o
o
t
i
m
i
m
u
o
nS
h
o
u
l
dh
d
i
s
c
o
v
e
r
E
da
u
t
o
m
a
t
i
c
a
l
l
y
.
F
I
T
1
e
E
E
f
G
E
e
t
h
ed
i
s
c
u
s
s
i
o
na
p
p
l
i
e
st
o
d
i
s
c
r
e
t
e
p
l
e
m
st
h
a
t
u
o
n
l
yi
n
c
o
m
p
l
e
t
ee
n
u
m
e
r
a
t
i
o
n
.

sp
r
o
b
l
e
mw
e
r
e

9
.
1
.
1 GlobalMinimum
Theg
l
o
b
a
lr
n
i
n
i
m
u
mf
o
rt
h
efunctio~f(X) e
x
i
s
t
si
fef
u
n
c
t
i
o
ni
sc
o
n
t
i
n
u
o
u
sona
n
o
n
e
m
p
t
yf
e
a
s
i
b
l
es
e
tSa
ti
sc
l
o
s
e
da
n
db
o
u
n
d
e
d[
1
]

~t!!.":~=<

9
.
1

352 G
L
O
B
A
Lo
p
n
M
I
Z
A
T
I
O
N
l
es
e
tSd
e
f
i
n
e
sar
e
g
i
o
no
fed
e
s
i
g
ns
p
a
c
e i
sf
e
a
s
i
b
l
e
.An
yp
o
i
n
ti
nSi
sa
t
h
ep
o
i
n
tX
c
a
n
d
i
d
a
t
ed
e
s
i
g
np
o
i
n
t X. I
nt
h
en
d
i
m
e
n
s
i
o
n
a
lC
a
r
t
e
s
i
a
ns
p
a
p
o
n
d
st
oav
e
c
t
o
ro
fl
c
n
g
t
hI
I
X
I
I
.Theb
o
u
n
d
a
r
i
e
so
ft
h
er
e
g
i
o
na
r
ep
a
r
to
ft
h
es
t
c
o
S(
c
1
o
s
e

.
l
el
e
n
g
t
ho
ft
h
ev
t
o
rX s
h
o
u
l
db
eaf
i
n
i
t
enumber(
b
o
u
n
d
e
d
)
. 1
d
e
f
i
n
i
t
i
o
nf
o
rt
h
eg
l
o
b
a
lminimumX.c
a
nb
es
t
a
t
e
d

P
R
O
B
L
E
MD
E
F
I
N
I
T
I
O
N 353

N
o
n
c
o
n
v
e
x
S
e
.

X2J
f
(
X
.
)f
(
X
)
; XeS

(
9
.
3
)

IX1~f

Thec
o
n
d
i
t
i
o
n
sf
o
rt
h
ee
x
i
s
t
e
n
c
eo
ft
h
eg
l
o
b
a
lminimuma
sg
i
v
e
na
b
o
v
ea
r
ea
t
t
r
i
b
u
t
e
d
t
oW
e
i
e
r
s
t
r
a
s
s
.Whileeg
l
o
b
a
lminimumi
sg
u
a
r
t
di
ft
h
ec
o
n
d
i
t
i
o
n
sa
r
es
a
t
i
s
f
i
c
d
e
ydon
o
tn
e
g
a
t
eee
x
i
s
t
e
n
c
eo
ft
h
eg
l
o
b
a
ls
o
l
u
t
i
o
ni
fc
o
n
d
i
t
i
o
n
sa
r
en
o
ts
a
t
i
s
f
i
e
d
.
y
p
eo
fo
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
e
m
st
h
ec
o
n
d
i
t
i
o
n
sa
r

C
o
n
s
i
d
e
r
i
n
gt
h
eb
r
o
a
dr
a
n
g
e dt
f
a
i
r
l
yl
i
m
i
t
e
di
n u
s
e
l
n
e
s
s
.Th
ec
o
n
d
i
t
i
o
n
sf
o
r
i
alm
umareesame
a
sc
o
n
s
t
r
a
i
n
t(
9.
3
)e
x
c
e
p
tt
h
es
e
tSi
sl
i
m
i
t
e
dt
oas
m
a
l
lr
e
g
i
o
na
r
o
u
n
dX
.
.

iJ

ConvexSet:lIsu
s
u
a
l
l
yr
e
l
a
t
e
st
o d
e
s
i
g
nv
e
c
t
o
r(
o
rd
e
s
i
g
np
o
i
n
t
s
)o
fe
p
r
o
b
l
e
m
.C
o
n
s
i
d
e
rtwod
e
s
i
g
np
o
i
n
t
sX.andX2Asb

t
h
es
e
tSd
e
f
i
n
e
sad
e
s
i
g
n
e
f
o
r
e
X
.a
n
dX2b
e
l
o
n
gt
oS
.Th
el
i
n
ej
o
i
n
i
n
gX.a
n
dX2
r
e
g
i
o
na
n
dbyi
m
p
l
i
c
a
t
i
o
nbo
c
o
n
t
a
i
n
sawholes
e
r
i
e
so
fnewd
e
s
i
g
np
o
i
n
t
s
.I
fal
1o
ft
h
e
s
ep
o
i
n
t
sa
l
s
ob
c
l
o
n
gt
oS
e
nSi
sac
o
n
v
e
xs
et
.Ac
I
o
s
e
ds
e
tSl
r
u
i
s
t
h
a
tt
h
ep
o
i
n
t
sX1 dX2c
a
na
l
s
ob
eon
t
h
eb
o
u
n
d
a
r
y
.
f
I however.t
h
el
i
n
ej
o
i
n
i
n
gX.andX2c
o
n
t
a
i
n
ssomep
o
i
n
t
st
h
a
tdon
o
t
nSi
sn
o
taconvexs
e
.
tForat
w
o
d
i
m
e
n
s
i
o
n
a
ld
e
s
i
g
np
r
o
b
l
e
m
.convex
b
e
l
o
n
g
t
o
S
.e
dn
onconvexs
e
t
sa
r
ci
l
l
u
s
t
r
a
t
e
di
nF
i
g
u
r
e9
.1
.

g
u
r
e9
.
1

34kuoh
y

Convexlty: I
fano
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
e
mc
a
nbeshownbec
o
n
v
e
x

t
h
e
nel
o
c
a
l
minimumw
i
l
la
l
s
ob
eag
l
o
b
a
lminimum.Ac
o
n
v
e
xo
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
e
mr
e
q
u
i
r
thc

t
h
e
na
1of i
l
n
e
q
u
a
l
i
t
y
o
b
j
c
c
t
i
v
ef
u
n
c
t
i
o
nbec
o
n
v
e
x
.I
fc
o
n
s
t
r
a
i
n
t
sa
r
ep
r
e
s
e
n
t
c
o
n
s
t
r
a
i
n
t
smustb
econvex
w
h
i
l
et
h
e u
a
l
i
t
yc
o
n
s
t
r
a
i
n
t
smustb
el
i
n
e
a
r
.

X
l
Con
v
e
xa
n
dn
o
n
c
o
n
v
e
xs
e
!
.

p
r
o
b
l
e
m
s
.Fromap
r
a
c
t
i
c
a
lv
i
e
w
p
o
i
n
t
.t
h
ee
i
g
e
n
v
a
l
u
e
so
f f
u
n
c
t
i
o
nmustb
eg
r
e
a
t
e
r
a
no
re
q
u
a
lt
oz
e
r
oa
ta
1
lp
o
i
n
t
si
nS
.
l
ti
si
m
p
o
r
t
a
n
tt
or
e
a
l
i
z
e ag
l
o
b
a
loptimummays
t
i
1
le
x
i
s
te
v
e
ni
ft
h
ec
o
n
v
e
x
i
l
y
t
.l
nr
e
a
lo
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
e
m
s
.i
tmayb
ed
i
f
f
i
c
u
l
t10e
s
t
u
b
l
i
s
h
c
o
n
d
i
t
i
o
n
sa
r
en
o
tme
c
o
n
v
e
x
i
t
y
.I
np
r
a
c
t
i
c
e
.c
o
n
v
e
x
i
t
yc
h
e
c
k
sa
r
er
a
r
e
l
yp
e
r
f
o
r
m
e
d
.Ana
s
s
u
m
p
t
i
o
ni
smade
ag
l
o
b
a
loptimume
x
i
s
t
sf
o
rt
h
ep
r
o
b
l
e
m
. i
sas
u
b
j
e
c
t
i
v
ee
l
e
m
e
n
tf
o
r
c
e
d
bye
x
p
e
r
i
e
n
c
e
.i
n
s
i
g
h
t
.a
n
di
n
t
u
i
t
i
o
n
.T
h
i
si
sa
l
s
oo
f
t
e
nj
u
s
t
i
f
i
e
dbyep
o
s
s
i
b
i
l
i
t
y
ab
e
t
t
e
rs
o
l
u
t
i
o
nt
h
a
nec
u
r
r
e
n
tonec
a
nb
ef
o
u
n
d
.

ConvexFunct/ons: A convexf
u
n
c
t
i
o
ni
sd
e
f
i
n
e
do
n
l
yf
o
rac
o
n
v
e
xs
et
.Once
a
g
a
i
nl
e
tSb
eaconvexs
e
tfromwhichtwod
e
s
i
g
np
o
i
n
t
sX.andX2a
r
cs
e
l
c
c
t
e
d
.
Letfbet
h
ef
u
n
c
t
i
o
nwhosec
o
n
v
e
x
i
t
yi
st
obec
s
t
a
b
l
i
s
h
e
d
.C
o
n
s
i
d
e
ra
n
yp
o
i
n
tX"
onal
i
n
ej
o
i
n
i
n
gX. dX2.Thef
u
n
c
t
i
o
n
f
i
sc
o
n
v
e
xi
f
j{X
Ji
sl
e
s
s va
1
ueof
ec
o
r
r
e
s
p
o
n
d
i
n
gp
o
i
n
ton l
i
n
e
j
o
i
n
i
n
g
f
t
X
.
)andftXu.l
nnd
i
m
e
n
s
i
o
n
s i
si
s
n
o
te
a
s
yt
os
e
b
u
tao
n
e
d
i
m
e
n
s
i
o
n
a
lr
e
p
r
e
s
e
n
t
a
t
i
o
ni
sp
r

e
n
t
e
di
nF
i
g
u
r
e9
.
2
.l
e
c
o
n
d
i
t
i
o
nf
o
rc
o
n
v
e
x
i
t
yi
nef
i
g
u
r
ei
s

f
(
x
)Sr

4
)
(
9.
X]

Am
a
t
h
e
m
a
t
i
c
a
lc
o
n
d
i
t
i
o
no
fc
o
n
v
e
x
i
t
yi
st
h
a
tt
h
H
e
s
s
i
a
nma
x(
m
a
t
r
i
xo
fs
e
c
o
n
d
d
e
r
i
v
a
t
i
v
c
s
)o
ft
h
ef
u
n
c
t
i
o
n
f
m
u
s
lb
cp
o
s
i
t
i
v
es
e
m
i
d
c
f
i
n
i
t
ca
ta
1
lp
o
i
n
t
si
nt
h
es
e
tS
.
Notet
h
a
lt
h
i
sc
o
r
r
e
s
p
o
n
d
st
oeKTc
o
n
d
i
t
i
o
n
sf
o
ru
n
c
o
n
s
t
r
a
i
n
e
dm
i
n
i
m
i
z
u
t
i
o
n

Xa

X
2

F
l
g
u
r
e9
.
2C
o
n
v
e
xf
u
n
c
t
l
n
.

1P
R
O
B
L
E
MO
E
F
I
NI
O
N 3
5
5

3
5
4 G
L
O
B
A
Lo
p
n
M
I
Z
A
n
O
N

9
.
1
.
2 Natureo
ftheS
o
l
u
t
l
o
n

Se
v
e
r
a
lm
l
n
i
m
aa
n
dmaxlmaw
i
t
hg
l
o
b
a
ls
o
l
u
t
i
o
n
-c
o
n
t
o
u
rp
l
o
t

(
9.
5
)

F
i
g
u
r
e9.
4i
I
l
u
s
t
r
a
t
e
sec
o
n
t
o
u
rp
l
o
tTh
ec
o
n
t
o
u
rp
l
o
tc
a
nbeb
e
t
r
a
p
p
c
i
a
t
e
di
n
c
o
l
o
r
u
s
i
n
g c
o
l
o
r
b
a
rf
o
ri
n
t
e
r
p
r
e
t
a
t
i
o
n
.R
u
n
n
i
n
gt
h
em-f
i
I
ew
i
l
lc
r
e
a
t
eb
o
t
hf
i
g
u
r
e
s
.
Fromr
e
l
a
t
i
o
n(
9
.
5
)i
ti
sc
I
e
a
rt
h
er
i
p
p
l
e
swi
I
ls
u
b
s
i
d
i
z
ea
sX
Ia
n
dx2movemuchf
a
r
t
h
c
r
awayf
r
o
mw
h
e
r
et
h
eg
l
o
b
a
lminimumo
c
c
u
r
sb
e
a
u
s
o
ft
h
i
n
f
l
u
c
n
c
eo
ft
h
e
d
e
n
o
m
i
n
a
t
o
r
.Then
u
m
e
r
a
t
o
ri
sc
o
n
s
t
r
a
i
n
c
dt
oI
icb
e
t
w
e
e
n1a
n
d+
1m
u
l
t
i
p
l
i
e
db
yt
h
c

1522F5Jl

s
i
n(
0
.
1+
(X.-4)l+(X2-4)
f
(
x
.
X
z
)=-20
.--~
0.
1+

1-4)z+(x2-4)
Z

14

l
o
b
a
loptimums
o
l
u
t
i
o
nwi
I
lb
eo
b
t
a
i
n
e
dn
u
m
e
r
i

a
11
y
.Ag
o
o
d
I
ti
sa
s
s
u
m
e
d eg
u
n
d
e
r
s
t
a
n
d
i
n
go
ft
h
en
a
t
u
r
eo
ft
h
es
o
l
u
t
i
o
ni
sp
o
s
s
i
b
l
ew
i
t
hr
e
f
e
r
e
n
c ac
o
n
t
i
n
u
o

d
e
s
i
g
ne
x
a
m
p
l
e
.E
x

n
s
i
o
nt
od
i r
e p
r
o
b
l
e
m
si
sp
o
s
s
i
b
l
eb
yc
o
n
s
i
d
e
r
i
n
gt
h
a
ta
n
y
c
o
n
t
i
n
u
o
u
ss
o
l
u
t
i
o
nc
o'
e
s
p
o
n
d
st
oas
p
e
c
i
f
i
cc
o
m
b
i
n
a
t
i
o
no
ft
h
es
e
to
fd
i
s
c
t
ev
a
l
u
e
s
f
o
rt
h
ed
e
s
i
g
nv
a
r
i
a
b
l
e
s
.F
i
g
u
r
e9.
3i
sa
ni
1
I
u
s
no
faf
i
g
u
r
ew
i
t
hs
c
v
e
r
a
lminima
a
n
dmaximaw
i
t
has
o
ngg
l
o
b
a
lminimuma
tx =4X2=4
. l
ef
i
g
e
a
v
a
i
l
a
b
l
e
r
o
u
g
hf
l
g
9
_
3
.
m
1i
st
h
ep
l
o
to
fef
u
n
c
t
i
o
n
(a
I
s
oE
x
a
m
p
l
e9.
1
)

S
e
v
e
r
a
lm
l
n
l
m
n
dm
a
x
l
m
aw
l
t
hg
l
o
b
a
ls
o
l
u
l
l
o
n
~

0 2 4

xv
a
l
u
e
s

F
l
g
u
r
e9
.
4C
o
n
l
o
u
rp
l
o
to
fE
x
a
m
p
l
e9
.
1
.

50

qH

1
5
-20

.
t
:
:

1
0
xz

1
0 1
0
F
i
g
u
r
e9
.
3S
u
r
f
a
c
ep
l
o
to
fExa
m
p
l
9
.
1
.

I
p
i
l
e
s1
0bed
o
w
n
l
o
a
d
e
dmt
h
ew
e
bs
i
!
ea
r
i
n
d
i
c
n
l
l
l
db
yb
o
l
d
f
a
c
es
n
n
ss
e
r
i
ft
y
p
c
.

sc
a
n
s
c
a
l
i
n
gf
a
c
t
o
ro
f2
0
.I
nt
h
el
i
m
i
tt
h
er
i
p
p
l
e
ss
h
o
u
l
ds
t
a
r
tt
obecomeap
l
a
n
e
.Thi
a
r
i
a
b
l
e
s
.
b
ec
o
n
f
i
r
m
c
db
ye
x
e
c
u
t
i
n
gf
l
g
9
_
3
.
mw
i
t
hd
i
f
f
e
r
e
n
tr
a
n
g
e
sf
o
r v
l
o
b
a
Is
o
l
u
t
i
o
nb
yo
b
s
e
r
v
a
t
i
o
n
.S
u
r
r
o
u
n
d
i
n
g
FromF
i
g
u
r
e9
.
3i
ti
se
a
s
yt
os
p
o
t g
eg
l
o
ba
Iminimuma
r
er
i
n
g
so
f1
0
1e
x
t
r
m
a
.N
u
m
e
r
i
c
a
l
l
yi
ti
sac
h
a
I
l
e
n
g
c1
0
1
lo
ft
h
ebumps a
n
dd
i
p
s
.A s
i
m
p
l
eway t
oc
h
a
r
a
c
t
c
.
en
u
m
e
r
i
c
a
l
n
e
g
o
t
i
a
t
ea
t
e
c
h
n
i
Q
u
e
sf
o
ram
i
n
i
m
i
z
a
t
i
o
np
r
o
b
l
e
mi
st
h
a
te
ya
t
t
e
m
p
tt
of
i
n
dad
i
t
c
h(
o
rv
a
l
l
e
y
)
a
Il
i
n
t
o
.I
ft
h
i
si
s a
sa
nu
n
c
o
n
s
t
r
a
i
n
c
dminimump
r
o
b
l
e
m

t
h
eK
1
c
o
n
d
i
t
i
o
n
s
t
of
p
c
d yoneofel
o
c
a
Iminimumr
i
n
g
s
d
e
p
e
n
d
i
n
g
w
i
l
l
l
e
a
dt
h
es
o
l
u
t
i
o
nt
obe p
t
a
r
t
i
n
gp
o
i
n
tf
o
rt
h
ci
t
e
r
a
t
i
o
n
s
.l
ft
h
ei
t
e
r
a
t
i
o
n
s'Cs
t
a
r
t
e
dc
Io
s
e
o
nt
h
el
o
c
a
t
i
o
no
fcs
4x2=
4e
ni
ti
smoreIik
e
l
ycg
l
o
b
a
loptimumwi
l
1b
ed
i
s
c
o
v
e
r
e
d
.
t
oX( =
-5
X
2=
-5
u
s
i
n
gs
t
a
n
d
a
r
dn
u
m
e
r
i
c
a
lt
h
n
i
q
u
e
sw
i
l
lp
r
o
b
a
b
l
yn
o
tl
e
a
d
S
t
a
r
i
n
ga
t
x
l=
t
o g
l
o
b
a
Im
inimum.
T
h
i
si
sas
i
m
p
l
ei
l
l
u
s
t
r
a
t
i
o
nb
u
tnotGthema
r
e
M
a
n
i
t
B
l
o
c
a
lminimumsa
n
d
g
np
r
o
b
l
e
m
sw
i
m
o
r
c 2
m ums daEingleglobalmmm.In i
V i
a
b
l
e
s

i
ti
sn
o
tr
e
a
d
i
l
ya
p
p
a
r
e
n
thowmanyl
o
ca
Iminimumse
x
i
s
t a
n
di
fe
g
l
o
b
a
I
minimumi
sr
c
a
l
l
yt
h
a
td
i
s
t
i
n
c
te
v
e
ni
fi
ti
sp
o
s
t
u
l
a
t
t
oc
x
i
st
.Wh
a
ta
b
o
u
tt
h
e
c
r
ei
snowayt
h
cg
l
o
ba
Iminimumc
a
ne
s
c
a
p
et
h
c
c
o
m
p
l
c
t
ee
n
u
m
e
r
a
t
i
o
nt
e
c
h
n
i
q
u
e
?Th

3
5
6

GLOBALOP
T
1MIZATION

9
.
2 NUMEAICALTECHNIQUESANDADDITIONALEXAMPLES

s
i
f
t
i
n
gd
u
et
oi
st
e
c
h
n
i
q
u
e
.T
h
e
r
ei
sa
1
s
on
ot
e
c
h
n
i
q
u
ea
ti
ss
i
m
p
l
e
r c
o
m
p
l
e
t
e
ep
r
o
b
l
e
mi
sI
het
i
m
et
of
i
n
i
s
hs
e
a
r
c
h
i
n
gI
hei
n
f
i
n
i
t
ep
o
i
n
t
so
fI
h
e
e
n
u
m
e
r
a
t
i
o
n
.l
h
es
o
l
u
t
i
o
n
.
d
e
s
i
g
ns
p
a
c
et
od
i
s
c
o
v
e
rI
Th
es
o
l
u
t
i
o
nt
og
l
o
b
a
lo
p
t
i
m
u
mm
u
s
tI
h
e
r
e
f
o
r
el
i
eb
e
t
w
e
e
ns
t
a
n
d
a
r
dn
u
m
e
r
i
c
a
l
t
e
c
h
n
i
q
u
e
sa
n
dc
o
m
p
l
e
t
ee
n
u
m
e
r
a
t
i
o
n
.Thef
o
r
m
e
rw
i
l
ln
o
tf
i
n
dI
h
es
o
l
u
t
i
o
nu
n
l
e
s
si
t
a
t
t
e
rc
a
r
m
o
tf
i
n
dI
h
es
o
l
u
t
i
o
ni
nt
i
m
et
obe u
s
e
f
u
l
.G
l
o
b
a
l
i
sf
o
r
t
u
n
a
t
e
.The1
o
p
z
a
t
i
o
nc
h
n
i
q
u
e
st
e
n
dt
of
a
v
o
rt
h
el
a
t
t
e
ra
p
p
r
o
a
c
hb
e
c
a
u
s
eo
fes
i
m
p
l
i
c
i
t
yo
f
f
u
l
l
h
es
o
l
u
t
i
o
nm
u
s
tb
ed
i
s
c
o
v
e
r
e
di
nar
e
a
s
o
n
a
b
l
et
i
m
e
.
i
m
p
1
e
m
e
n
t
a
t
i
o
n
.Tob
e e

9
.
1
.
3 ElementsofaNumerlcalTechnlque
l
ep
r
o
b
1
e
m
sf
a
c
e
db
yan
u
m
e
r
i
c
a
lt
e
c
h
n
i
q
u
es
e
a
r
c
h
i
n
gf
o
rI
h
eg
l
o
b
a
lminimumh
a
v
e
b
nt
o
u
c
h
e
do
ni
nep
r
e
v
i
o
u
ss
e
c
t
i
o
n
.I
nm
o
s
ts
t
a
n
d
a
r
dn
u
m
e
r
i
c
a
lt
e
c
h
n
i
q
u
e
sE

l
o
c
a
1r
n
i
n
i
m
u
ms
e
r
v
e
sa
sas
t
r
o
n
gma
e
to
rab
e
a
c
o
nf
o
rI
h
es
o
l
u
t
i
o
n
.I
nm
o
s
tc
a
s
e
s

d
e
s
i
g
nc
h
a
n
g
e
sa
r
en
o
ta
c
c
e
p
t
e
di
fI
h
e
r
ei
sn
oi
m
p
r
o
v
e
m
e
n
ti
nt
h
es
o
l
u
t
i
o
n
.I
fs
t
a
n
d
a
r
d
n
u
m
c
r
i
c
a
lt
e
c
h
n
i
q
u
e
sh
a
v
et
obea
d
a
p
t
e
dt
ol
o
o
kf
o
rg
l
o
b
a
lr
n
i
n
i
m
u
m
I
h
e
nI
h
e
ym
u
s
t
e
n
c
o
u
r
a
g
es
o
l
u
t
i
o
n
sa
t
i
nes
h
o
r
tr
u
nmayi
n
c
r
e
a
s
eI
heo
b
j
e
c
t
i
v
ev
a
l
u
es
oe
y
c
a
nf
i
n
da
n
o
l
h
e
rv
a
l
l
e
ya
tw
h
o
s
eb
o
t
t
o
mi
seg
l
o
b
a
lm
i
n
i
m
u
m
.The
ym
u
s
ta
c
h
i
e
v
e
I
hi
sw
i
l
h
o
u
ts
a
c
r
i
f
i
c
i
n
g s
i
g
n
i
f
i
c
t
p
o
t
e
n
c
yo
fI
hel
o
c
a
lr
n
i
n
i
m
u
m
.b
e
c
a
u
s
eal
o
c
a
l
minimums
o
l
u
t
i
o
ni
sb
e
t
t
e
rt
h
a
nn
os
o
l
u
t
i
o
na
ta
1
1
.F
i
g
u
r
e9
.
5i
l
l
u
s
t
r
a
t
e
sI
h
e
s
ei
d
e
a
s
u
s
i
n
gas
i
n
g
l
ev
a
r
i
a
b
l
e
.
I
nF
ig
u
9
.
5
ei
t
e
r
a
t
i
o
n
ss
t
a
a
tp
o
i
n
tA.I
nas
t
a
n
d
a
r
dn
n
e
r
i
1
t
e
c
h
n
i
q
u
e
.a
S hd
i
r
t
i
o
ni
sf
o
u
n
dw
h
i
c
hl
e
a
d
st
op
o
i
n
tB
.AtB
.i
fn
v
e
n
t
i
o
n
a
ln
u
m
e
r
i
c
a
l
t
e
c
h
n
i
q
u
e
sw
e
r
eu
s
e
d

as
e
a
r
c
hd
i
r
e
c
t
i
o
nl
e
a
d
i
n
gt
op
o
i
n
tCw
i
l
lbea
c
c
e
p
t
e
d

w
h
i
l
ea

i
fp
o
i
n
tDw
a
s
s
e
a
r
c
hd
i
r
e
c
t
i
o
nl
e
a
d
i
n
gt
op
o
i
n
tDw
i
l
lb
es
t
r
o
n
g
l
yd
i
s
c
o
u
r
a
g
e
d
.B
u
t
somehowr
e
a
c
h
a
b
l
ee
ni
tc
o
u
1
dmoveo
nt
oEa
n
de
nFa
n
dc
o
n
t
i
n
u
et
oI
heg
1
o
b
a
1

3
5
7

m
i
n
i
m
u
m
.Them
a
i
ni
d
e
ai
nf
i
n
d
i
n
gt
h
eg
l
o
b
a
lminimumi
st
oe
n
c
o
u
r
a
g
es
o
l
u
t
i
o
n
sa
t
w
i
l
le
s
c
a
p
et
h
ea
l c
t
i
o
no
fI
hel
o
c
a
lm
i
n
i
m
u
m
.T
h
i
sm
e
a
n
sa
lp
o
i
n
tEi
ti
sq
u
i
t
e
p
o
s
s
i
b
l
ees
e
a
r
c
hd
i
r
e
c
t
i
o
nmay1
e
a
dt
op
o
i
n
tG
.1i
se
nl
r
u
i
r
e
da
ta1
a
r
g
e
n
u
m
b
e
ro
fi
t
e
r
a
t
i
o
n
sb
ep
e
r
m
i
t
t
e
ds
oa
tI
h
ev
a
1
1
e
yh
o
1
d
i
n
gI
heg
1
o
b
a
1minimumwi
1
l
b
er
e
v
i
s
i
t
e
d
.Fromi
sb
r
i
e
fi
1
l
u
s
t
r
a
t
i
o
nI
hec
h
a
r
a
c
t
e
r
i
s
t
i
c
sf
o
rat
e
c
h
n
i
q
u
e e
dt
of
i
n
d
ag
l
o
b
a
lminimumc
a
nb
es
u
m
m
a
r
i
z
e
da
s
:
S
h
o
r
t
t
e
r
mi
n
c
r
e
a
s
e
si
nt
h
eo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
nm
u
s
tb
ee
n
c
o
u
r
a
g
e
d
.
r
1
yu
s
e
f
u
lb
c
a
u
s
e
T
r
a
d
i
t
i
o
n
a
lg
r
a
d
i
e
n
t
b
a
s
e
dt
e
c
h
n
i
q
u
e
sw
i
l
ln
o
tb
ep
a
r
t
i
c
u
1
a
e
ya
r
eb
a
s
e
do
ns
e
a
r
c
hd
i
r
c
t
i
o
n
sa
td
eI
heo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
na
tI
he
c
u
r
r
e
n
ti
t
c
r
a
t
i
o
n
.The
ywi
1
1n
o
tp
c
r
m
i
ti
n
c
r
o
fI
h
eo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
nv
a
l
u
e
.
I
fd
e
s
c
e
n
t
b
a
s
e
ds
e
a
r
c
hd
i
r
e
c
t
i
o
n
sa
r
enotp
a
r
to
fI
h
et
e
c
h
n
i
q
u
e a
n
y
e
c
r
e
e
seo
b
j
e
c
t
i
v
es
h
o
u
I
dbea
c
c
e
p
t
e
dt
op
r
o
m
o
t
ea
t
t
r
a
c
t
i
o
n
d
i
r
e
c
t
i
o
n d
b
yI
h
e1
0
c
a
lr
n
i
n
i
m
u
m
.
As l
o
c
a
t
i
o
no
fI
h
eg
l
o
b
a
lminimumi
sn
o
tknown
s
e
a
r
c
hm
u
s
tbep
e
r
m
i
t
t
e
d
o
v
e
ra1
a
r
g
cr
e
g
i
o
nw
i
l
hn
ob
i
a
s

a
n
df
o
ra1
a
r
g
en
u
m
b
e
ro
fi
t
e
r
a
t
i
o
n
s
.
S
i
n
c
ea
c
c
e
p
t
a
b
l
ed
i
r
e
c
t
i
o
n
sc
a
nd
e
c
r
e
a
s
eo
ri
n
c
r
e
a
s
e o
b
j v
ef
u
n
c
t
i
o
n
ah
c
u
r
i
s
t
i
cd
e
t
e
r
m
i
n
a
t
i
o
no
f S hd
i
r
e
c
t
i
o
n(
o
rd
e
s
i
g
nc
h
a
n
g
e
)m
u
s
t
v
a
l
u
e
bep
t
o
f
t
h
ec
h
n
i
q
u
e
.
h
ez
e
r
o
o
r
d
e
rt
e
c
h
n
i
q
u
ef
o
ru
n
c
o
n
s
t
r
a
i
n
e
d
T
h
e
s
ef
e
a
t
u
r
e
sr
e
s
e
m
b
1
eI
z
a
t
i
o
n dI
h
e
r
c
f
o
r
ew
i
l
lr
c
q
u
i
r
eal
a
r
g
en
u
m
b
e
ro
fi
t
e
r
a
t
i
o
n
st
ob
e
o
p
e
f
f
e
c
t
i
v
e
.l
el
a
r
g
en
u
m
b
e
ro
fi
t
e
r
a
t
i
o
n
swi
1
lf
o
r
c
eI
het
e
c
h
n
i
q
u
e1
0b
es
i
m
p
1
e
.
Thes
o
l
u
t
i
o
nc
a
no
n
l
yb
ec
o
n
c
l
u
d
e
db
a
s
e
do
nI
h
en
u
m
b
e
ro
fi
t
e
r
a
t
i
o
n
sa
s
n
e
c
e
s
s
a
r
yc
o
n
d
i
t
i
o
n
s(
a
v
a
i
l
a
b
1
ef
o
rc
o
n
t
i
n
u
o
u
sp
r
o
b
l
e
m
so
n
l
y
)c
a
n
n
o
t
d
i
s
t
i
n
g
u
i
s
hb
e
t
w
e
e
nl
o
c
a
1a
n
dg
l
o
b
a
lo
p
t
i
m
u
m
.

I
h
eb
e
s
tt
e
c
h
n
i
q
u
ei
sc
o
m
p
l
e
t
eo
re
x
b
a
u
s
t
i
v
ee
n
u
m
e
r
a
t
i
o
n
.I
ti
s
W
i
l
h
o
u
tad
o
u
b
t
I
h
e
r
e
f
o
r
en
e
c
e
s
s
a
r
yt
oa
s
s
u
m
et
h
a
tc
o
m
p
l
e
t
ee
n
u
m
e
r
a
t
i
o
ni
sI
lOt
a
l
lo
p
t
i
o
l
li
nf
u
r
t
h
e
r
d
i
s
c
u
s
s
i
o
n
.
Thep
r
o
b
1
e
md
e
s
c
r
i
p
t
i
o
n
i
l
l
u
s
n

a
n
d
d
i
s
c
s
i
o
nh
a
v
ed
e
a
l
tw
i
l
hu
n
c
o
n
s
a
i
n
e
d
a
i
n
t
sc
a
nb
ei
n
c
l
u
d
e
db
ya
d
d
r
e
s
s
i
n
gI
h
ea
b
o
v
ec
h
a
r
a
c
t
e
r
i
s
t
i
c
st
o
r
n
i
n
i
m
i
z
a
t
i
o
n
.C
o
n
s
a
p
p
l
yt
of
e
a
s
i
b
l
es
o
l
u
t
i
o
n
so
n
l
y
.I
ti
s1
e
f
tt
oI
her
e
a
d
e
rt
oc
o
n
t
m
p
l
a
t
eea
b
i
l
i
t
yt
o
s
t
i
c
a
1
l
yg
e
n
e
r
a
t
eal
a
r
g
en
u
m
b
e
ro
ff
e
a
s
i
b
l
ed
e
s
i
g
nc
h
a
n
g
e
sf
o
rc
o
n
s
i
d
e
r
a
t
i
o
n
.
h
e

f
(
x
)

9
.
2 NUMERICALTECHNIQUESANDADDITIONALEXAMPLES
Thi
ss
e
c
t
i
o
ni
n
t
r
o
d
u
c
e
st
w
op
o
p
u
l
a
rm
e
l
h
o
d
sf
o
rg
l
o
b
a
lo
p
t
i
r
n
i
z
a
t
i
o
n
.B
o
l
ha
r
ed
e
r
i
v
e
d

x
F
1g
u
r
e9.6 S
t
r
a
t
e
g
yf
o
rg
l
o
b
a
lo
p
t
i
m
u
ms
e
a
r
c
h
.

f
r
o
mn
a
t
u
r
a
Ip
h
y
s
i
c
a
lp
r
o
c
e
s
s
a
th
e
l
pi
d
e
n
t
i
f
yo
p
t
i
m
u
ms
t
a
t
e
s
.l
ef
i
r
s
to
n
e

S
i
m
u
l
a
t
e
dAnn
e
a
l
i
n
g(
S
A
)
.r
e
f
e
r
s ap
h
y
s
i
c
a
lp
r
o
c
e
s
sa
te
s
t
a
b
l
i
s
h
e
sd
i
f
f
e
r
c
n
t
c
r
y
s
t
a
1
l
i
n
es
t
r
u
c
t
u
r
e
s
e
a
c
ho
fw
h
i
c
hc
a
nb
ea
s
s
o
c
i
a
t
e
dwi
d
i
f
f
e
r
e
n
tminimum
p
o
t
e
n
t
i
a
le
n
e
r
g
y
.T
h
isi
sa
t
t
e
m
p
t
e
db
yr
e
h
e
a
t
i
n
ga
n
dc
o
o
l
i
n
go
f m
a
t
e
r
i
a
l
s
/
m
e
t
a
1
s
w
h
i
l
ec
o
n
t
r
o
l
l
i
n
g r
a
l
eo
fc
o
o
l
i
n
g
.l
es
e
c
o
n
d

I
h
eG
e
n
e
t
i
cA
l
g
o
r
i
l
h
m(
G
A
)
w
o
r
k
s
t
h
r
o
u
g
hm
u
t
a
l
i
o
no
fc
a
n
d
i
d
a
t
es
o
l
u
t
i
o
n
su
n
t
i
l
l
h
eb
e
s
ts
o
l
u
t
i
o
nc
a
nb
ee
s
l
a
b
l
i
s
h
e
da

358

GLOBALOPTIMlZATION

9.
2 NUMERICALTECHNIQUESANOAOOtnONALEXAMPLES

p
r
o
c
e
s
s a
tc
a
nb
er
e
l
a
t
e
dt
on
a
t
u
r
1s
o
l
u
t
i
o
no
rb
i
o
l
o
g
i
c
a
le
v
o
l
u
t
i
o
n
. l
e
s
e
I
Iu
s
t
r
a
t
e
dw
i
t
hr
e
s
p
e
c
tt
ou
n
c
o
n
s
t
r
a
i
n
e
dp
r
o
b
l
e
m
si
nt
h
i
ss
e
c
t
i
o
n
.
t
e
c
h
n
i
q
u
e
sa
r
ei

Thea
l
g
o
r
i
mi
sh
e
u
r
i
s
t
i
c
.
Thea
l
g
o
r
i
t
h
mi
ss
u
g
g
e
s
t
i
v
co
fab
i
a
s
e
drandomw
a
l
k
.
D
e
s
c
e
n
to
ft
h
ef
u
n
c
t
i
o
ni
sd
i
r
e
c
t
l
ya
c
c
o
m
m
o
d
a
l
e
d
.

9
.
2
.
1 SlmulatedAnneaJ
lng(SA)
Thcf
i
r
s
ta
p
p
l
i
c
a
t
i
o
nu
s
i
n
gSAf
o
ro
p
t
i
m
i
z
a
t
i
o
nwasp
r
e
s
e
n
t
c
dbyK
i
r
k
p
a
t
r
i
c
ke
ta
l
.
[21.Theexamplesc
o
v
e
r
e
dp
r
o
b
l
e
m
s frome
l
e
c
t
r
o
n
i
cs
y
s
t
e
m
s
i
np
u
t
i
c
u
l
a
r

componentp
l
a
c
e
m
e
n
t e
nbyt
h
eneedf
o
rminimumw
i
r
i
n
gc
o
n
n
e
c
t
i
o
n
s
.The
d
e
m
o
n
s
t
r
a
t
c
dby
a
l
g
o
r
i
t
h
mi
t
s
e
l
fwasb
a
s
e
dont
h
eu
s
eo
fs
t
a
t
i
s
t
i
c
a
lm
e
c
h
a
n
i
c
s
M
e
t
r
o
p
o
l
i
se
ta
l
.[
3
]
t
oe
s
t
a
b
l
i
s
ht
h
e
r
m
a
lc
q
u
i
l
i
b
r
i
u
mi
nacol
1c
c
t
i
o
no
fa
t
o
m
s
.The
p
a
r
t
i
c
u
l
a
r
l
y
t
e
r
mi
t
s
e
l
fi
smorer
e
p
r
c
s
e
n
t
a
t
i
v
eo
ft
h
ep
r
o
c
e
s
so
fc
o
o
l
i
n
gm
a
t
c
r
i
a
l
s
m
e
t
a
l
sa
f
t
e
rr
a
i
s
i
n
gt
h
e
i
rt
e
m
p
e
r
a
t
u
r
et
oa
c
h
i
e
v
ead
e
f
i
n
i
t
cc
r
y
s
t
a
l
l
i
n
es
t
a
t
c
.As
t
h
ee
x
t
e
n
s
i
o
nt
oo
p
t
i
m
i
z
a
t
i
o
ni
sm
a
i
n
l
yh
c
u
r
i
s
t
i
c
.Th
cs
i
m
p
l
e
m
e
n
t
i
o
n
e
de
a
r
l
i
e
r
i
d
e
ab
e
h
i
n
dSAcanb
E
i
l
l
u
s
t
r
a
t
e
dt
h
r
o
u
g
ht
h
eI
n
o
d
i
a
e
db
a
s
i
ca
l
g
o
r
i
t
h
mo
fSAdue
t
oBochaveskyc
ta
l
.(
4
)
.

A
l
g
o
r
l
t
h
m
:BaslcSlmulatedAnneallng(
A
9
.
1
)

S
t
c
p1
.Chooscs
t
a
r
t
i
n
gd
e
s
i
g
n"
0
.Calculatef
oJ
(
"
o
)
(Necds
t
o
p
p
i
n
gc
r
i
t
c
r
i
a
)
S
t
e
p2
.Choos
a
randomp
o
i
n
tont
h
cs
u
r
f
a
c
eo
fau
n
i
tn
dI
lI
I
I
:R
s
i
o
n
a
lh
y
p
c
r
s
p
h
e
r
e
(
i
ti
sas
p
h
e
r
ei
nn
d
i
m
e
n
s
i
o
n
s
)t
oe
s
t
a
b
l
i
s
has
e
a
r
c
hd
i
r
e
c
t
i
o
nS
S
t
e
p3
.U
s
i
n
gas
t
e
p
s
i
z
e

c
a
l
c
u
l
a
t
e

f
t=f(Xo+);

359

Af=

S
t
e
p4
.I
fAfS0
t
h
e
np=1
E
l
s
ep=e
f
J
t
.
f
S
t
e
p5
. Arandomnumberr(
0Sr<1
)i
sg
e
n
e
r
a
t
e
d
.

I
frSp
t
h
e
nt
h
es
t
e
pi
sa
c
c
e
p
t
e
dandt
h
ed
e
s
i
g
nv
e
c
t
o
ri
sup
t
e
d
s
e

noc
h

g
ei
smadet
oe
d
e
s
i
g
n
Got
oStep2

D
i
r
l
t
i
o
n
sa
ti
n
c
r
e
a
s
eev
a
l
u
eo
ft
h
ef
u
n
c
t
i
o
na
r
es
o
m
e
t
i
m

p
e
r
m
i
t
t
e
d
.T
h
i
s
o
t
e
n
t
i
a
lt
od
i
s
c
o
v
e
rt
h
g
l
o
b
a
l
i
si
m
p
o
r
t
a
n
tt
oe
s
c
a
p
el
o
c
a
lo
p
t
i
m
u
m
.T
h
i
si
st
h
ep
optimumt
h
r
o
u
g
hea
l
g
o
r
i
t
h
m
.

Th
i
sb
a
s
i
ca
l
g
o
r
i
t
h
mc
a
nd
e
f
m
i
t
e
l
yb
ei
m
p
r
o
v
e
d
.D
u
r
i
n
gt
h
ei
n
i
t
i
a
li
t
e
r
a
t
i
o
n
si
ti
s
n
e
c
e
s
s
a
r
yt
oe
n
c
o
u
r
a
g
es
o
l
u
t
i
o
n
st
oe
s
c
a
p
e l
o
c
a
lo
p
t
i
m
u
m
.Af
t
e
rar
e
a
s
o
n
a
b
l
yl
a
r
g
e
h
eo
p
p
o
r
t
u
n
i
t
yt
oc
s
c
a
p
cl
o
c
a
l optimum (window o
f
number o
fi
t
c
r
a
t
i
o
n
st
o
p
p
o
r
t
u
n
i
t
y
)mustb
ec
o
n
t
i
n
u
o
u
s
l
yr
c
d
u
c
e
ds
ot
h
a
tt
h
cs
o
l
u
t
i
o
l
ld
o
c
sn
o
twandcra
l
l
o
v
e
rt
h
ep
l
a
c
ea
n
dc
a
n
c
c
lt
h
ei
m
p
r
o
v
e
m
e
n
t
st
ot
h
es
o
l
u
t
i
o
na
th
a
v
ea
l
r
e
a
d
yt
a
k
e
n
b
ys
c
h
e
d
u
l
i
n
gc
h
a
n
g
e
s ep
a
m
e
t
e
r~
p
1
a
c
e
.T
h
i
sc
a
nb
eimp1emen
i
t
e
r
a
t
i
o
n
si
n
c
r
e
a
s
e
.I
nSAt
e
n
n
i
n
o
l
o
g
yi
si
sr
e
f
e
r
r
e
dt
o ea
n
n
e
a
l
i
n
gs
c
h
e
d
u
l
e
.
l
e
s
cm
o
d
i
f
i
c
a
t
i
o
n
sdon
o
ti
na
n
ywayd
e
c
r
e
a
s
et
h
ep
r
o
b
l
em!
u
s
e
rd
e
p
e
n
d
e
n
c
eo
ft
h
e
a
l
g
o
r
i
t
h
m
.T
h
i
sp
r
o
b
l
e
md
e
p
e
n
d
e
n
c
i
su
s
u
a
l
l
yu
n
a
t
t
r
a
c
t
i
v
ef
o
ra
u
t
o
m
a
t
i
co
rs
t
a
n
d
a
r
d
i
m
p
l
e
m
e
n
t
a
t
i
o
no
ft
h
eSA m
c
t
h
o
d
.R
c
s
e
a
r
c
h
e
r
sa
r
ee
n
g
a
g
c
di
nc
o
n
s
t
r
u
c
t
i
o
no
f
tdon
o
td
e
p
e
n
donu
s
c
r
s
p
e
c
i
f
i
e
da
l
g
o
r
i
t
h
mc
o
n
t
r
o
1p
a
r
a
m
e
t
e
[
5
J
.
a
l
g
o
r
i
t
h
m
sa
TheSAa
l
g
o
r
i
t
h
mf
o
rc
o
n
t
i
n
u
o
u
sv
a
r
i
a
b
l
e
si
m
p
l
c
m
e
n
t
e
di
nt
h
ebooki
n
c
l
u
d
e
smany
o
f b
a
s
i
ci
d
e
a
sf
r
o
mt
h
eo
r
i
g
i
n
a
la
l
g
o
r
i
t
h
m
.T
h
e
r
ea
r
emanya
d
d
i
t
i
o
n
a
lc
h
a
n
g
e
s
h
e
.F
i
r
s
t n
e
g
a
t
i
v
es
i
g
ni
nt
h
ep
r
o
b
i
l
i
t
y rmi
sf
o
l
d
e
di
n
t
o~ i
t
s
c
lf
.A
no
s
i
l
ep
r
o
c
e
d
u
r
ei
si
n
c
l
u
d
e
df
o
ro
b
t
a
i
n
i
n
gt
h
ev
a
l
u
eof~. I
ti
sb
a
s
e
dont
h
ei
n
i
t
i
a
lv
a
l
u
c
0andf
o
.~ i
sc
a
1
c
u
l
a

yicldav
a
l
u
eo
fp=0
.
7f
o
raA
fu
a
lt
o50%o
f
o
f"
i
n
i
ea
l
g
o
r
i
t
h
mf
a
v
o
r
san
e
g
a
t
i
v
ev
a
l
u
eo
f~ byd i
.S
i
n
c
ct
h
ep
r
o
c
e
s
sf
o
r
e
s
t
a
b
l
i
s
h
i
n
ga
ni
n
i
t
i
a
l~ i
1
I
u
s t
e
da
b
o
v
ec
a
nd
e
t
c
r
m
i
n
capo v
ev
a
l
u
e

t
h
ea
l
g
o
r
i
t
h
m
ti
sl
a
t
c
ri
l
l
u
s
t
r
a
t
e
dt
h
r
o
u
g
hn
u
m
e
r
i
c
a
le
x
p
e
r
i
m
e
n
t
st
h
a
ta
c
o
r
r
e
c
t
st
h
i
so
c
c
u
r
r
e
n
c
c
.I
s
m
a
l
1m
a
g
n
i
t
u
d
eo
ft
h
cp
a
r
a
r
n
c
t
e
r~ a
l
l
o
w
st
h
esolutionwandera
l
lo
v
e
rt
h
ep
1
a
c
c
w
h
i
l
ea1
a
r
g
ev
a
l
u
ep
r
o
v
i
d
e
sas
'O
ng1
0c
a
lata
c
t
i
o
n
.I
ns
t
e
a
do
ft
h
eSAs
t
e
pb
e
i
n
g
i
n
c
o
r
p
o
r
a
t
e
dwhen o
b
j
e
c
t
i
v
ei
n
c
r
e
a
s
e
s
i
ti
st
r
i
g
g
e
r
e
dwhen o
b
j
e
c
t
i
v
ei
s1
i
k
e
l
y
l
o
p
e
.I
n
ea
1
g
o
r
i
t
h
mb
e
l
o
w
t
h
ec
o0
1p
a
r
a
m
e
t
e
r~ i
sk
c
p
t
t
oi
n
c
r
e
a
s
cb
a
s
c
don s
c
o
n
s
t
a
n
ta
ti
t
si
n
i
t
i
a
lv
a
l
u
e
.Thec
a
l
c
u
l
a
t
i
o
no
fisb
o
r
r
o
w
e
dfromt
h
et
r
a
d
i
t
i
o
n
a
l
o
n
e
d
i
m
e
n
s
i
o
n
a
ls
t
e
p
s
i
z
ec
a
l
c
u
l
a
t
i
o
nu
s
i
n
gt
h
eGoldenS
e
c
t
i
o
nmethod(
C
h
a
p
t
e
r5
)i
f
t
h
eo
b
j
e
c
t
i
v
ei
sl
i
k
e
1
yt
od
e
c
r
e
e
.I
ft
h
es
l
o
p
eo
ft
h
co
b
j
e
c
t
i
v
ei
sp
o
s
i
t
i
v
eo
rz
e
r
oa
f
i
x
e
disc
h
o
s
e
n
.T
h
i
sv
a
l
u
ei
samaximumo
foneo
ft
h
ep
r
e
v
i
o
u
s

I
fimplemented a
l
g
o
r
i
t
h
mh
a
st
oe
x
e
c
u
t
e
df
o
rar
e
a
s
o
n
a
b
l
yl
a
r
g
enumbero
f
i
t
e
r
a
t
i
o
n
sf
o
rt
h
es
o
l
u
t
i
o
nt
o t
oag
l
o
b
a
lminimum.Th
es
i
m
p
l
i
c
i
t
yo
ft
h
ea
l
g
o
r
i
t
h
m

r
e
f

11

0b
s
c
r
v
a
6

onw
i
l
l
i nd

ica

t
e at d~ i
mn
t
i
sv
e
r
ya
p
p
e
a
l
i
n
g
.A c
a
p

a
r

thep1
eter~can b
e
r
l

e
l
a
t
t

o
t

heBoltzma
n
n
pr

o
b
a

.
b

i
l
i
t
yd
i

s
t

r
i

bu
t

in
.I
ns
t
a
t
i
s
t
i
c
a
l
s
s
e
da
sk
lT
;Ti
sc
o
n
s
i
d
e
r
e
dca
n
n
e
a
l
i
n
gt
e
m
p
e
r
a
t
u
a
ndke
m
e
c
h
a
n
i
c
si
ti
sexp
Boltzmannc
o
n
s
t
a
nt
.I
td
i
c
t
a
t
e
st
h
ec
o
n
d
i
t
i
o
n
a
lp
r
o
b
a
b
a
taw
o
r
s
es
o
l
u
t
i
o
nc
a
n
p
t
e
d
.Fort
h
ea
l
g
o
r
i
t
h
mt
ob
ee
f
f
e
c
t
i
v
e
i
ti
s commendedt
h
a
tpb
ei
nt
h
er
a
n
g
e
beac
0.
5SpS0
.
9
.I
nf
o
r
m
a
lSAt
e
r
m
i
n
o
l
o
g
y

~r
e
p
r
s
e
n
t
seann
i
n
gt
e
m
p
e
r
a
t
u
r
e
.Th
e
f
o
I
l
owingi
d
e
a
si
nt
h
ea
l
g
o
r
i
t
h
ma
r
ca
p
p
a
r
e
n:
t

A
l
g
o
r
l
t
h
m
:SimulatedAnneallng9
.

S
t
e
p1
.Choose"
0
.
J("o) p=
0
.
7~ =
l
n
(
p
)
/
0
.
5xJ
(
"
o
)
< l

s=1
I
fs>O

s=lxs

360 G
L
O
B
A
LO
P
T
I
M
I
Z
A
T
I
O
N

9
.
2

N
U
M
E
R
I
C
A
LT
E
C
H
N
I
Q
U
E
SANDA
D
D
I
T
I
O
N
A
LE
X
A
M
P
L
E
S 3
6
1

"

S
t
e
p2
.Choosearandomp
o
i
n
tont
h
es
u
r
f
a
c
eo
fau
n
i
tn
d
i
m
e
n
s
i
o
n
a
lh
y
p
e
r
s
p
h
e
t
oe
s
t
a
b
l
i
s
has
e
a
r
c
hd
i
r
e
c
t
i
o
nS

{
X
o
)
;
S
t
e
p3
.C
a
l
c
u
l
a
t
eVj

Example9
.
1 T
h
i
si
st
h
ef
u
n
c
t
i
o
nu
s
e
di
nS
e
c
l
i
o
n9
.
1
.Theg
l
o
b
a
lminimumi
sa
tX
=
[
4
4
]w
i
t
ha
n0t
i
v
ev
a
l
u
ej
"=-19.6683.I
nef
i
g
u
r
t
h
ec
o
n
t
o
u
r
ss
u
g
g
e
s
ta
s
t
r
o
n
gg
l
o
b
a
lminimume
n
c
i
r
c
l
e
dbya
l
t
e
m
a
t
el
o
c
a
lmaximumsa
n
dminimums.F
o
r
mosto
ft
h
er
u
n
ses
t
a
r
l
i
n
gv
a
l
u
ewasf
a
rfromt
h
eg
l
o
b
a
lminimum.

I
fVj
{Xo
)<0
t
h
e
ncomputeI

D.

S
; =Xo+AX; Xo i+1
AX=
)~ 0
{
Xo
I
fVj
3
1
F
3
2
q
p

-pd

4
f
=f(
+f(); a=m
(

1
)
p=e~1lf
Arandomnumberr
;(
0Sr<1
)i
sg
e
n
c
r
a
t
e
d
.

I
frSpenes
t
e
pi
sa
c
c
e
p
d
anded
e
s
i
g
nv
e
c
t
o
ri
su
p
d
a
t
e
d
.
X; i i
+1
AX=aS; =Xo+d

5
2
5

1
2
2
2
2

HKZ

E
l
s
ec
o
n
t
i
n
u
ei
i+1
00t
oStep2

>
>

help SimulatedAnnealing

i
nt
h
ecommandwindow(
t
h
ec
u
r
r
e
n
td
i
r
e
c
t
o
r
ymustc
o
n
t
a
i
nt
h
ef
i
l
e
s
)
Th
e a
d
e
ri
ss
t
r
o
n
g
l
ye
n
c
o
u
r
a
g
e
dt
or
u
nt
h
eexample00h
i
sc
o
m
p
u
t
e
r
.I
twouldb
e
u
l
t
sp
r
e
s
e
n
t
e
dh
e
r
ea
r
ed
u
p
l
i
c
a
t
e
d
.Running
as
i
n
g
u
l
a
rc
o
i
n
c
i
d
e
n
c
ei
fanyo
ft
h
e s
n
u
m
e
r
i
c
a
le
x
p
e
r
i
m
e
n
t
si
ss
i
m
p
l
yam
a
t
t
c
ro
fc
a
l
I
i
ngt
h
eSAa
l
g
o
r
i
t
h
mandt
a
k
i
n
gal
o
o
k
l
ot
)
.Th
ed
e
s
i
g
nc
h
a
n
g
e
sa
r
ei
l
l
u
s
t
r
a
t
e
du
s
i
n
gc
o
l
o
r
s
:
a
tt
h
er

u
l
t
s(
e
s
p
e
c
i
a
l
l
yep
gd
a
s
i
n
gslopes-andblue-forp
o
s
i
t
i
v
es
l
o
p
e
s tl
e
a
dt
o
r
e
d
r
e
p
r
e
s
e
n
d
e
s
i
g
nc
h
a
n
g
e
s
.Th
en pmenti
nt
h
el
o
c
a
lminimuma
n
dt
h
ee
s
p
ear
c1e
a
r
l
y
i
I
lu
s
t
r
a
t
e
dont
h
ec
o
n
t
o
u
rp
l
o
t
s
.Thec
o
n
t
o
u
rp
l
o
t
s ef
i
l
l
e
dc
o
l
o
rp
l
o
t
st
oi
d
e
n
t
i
f
ya
r
e
a
s
ont
h
ef
i
g
u
r
ewheret
h
ef
u
n
c
t
i
o
nh
a
sal
o
c
a
lminimumo
rmaximum.D
a
r
k
e
rc
o
l
o
r
s
p
r
e
s
e
n
tminimumsandl
i
g
h
t
e
rc
o
l
o
r
st
h
emaximum.Thee
x
a
m
p
l
e
swerer
u
nu
s
u
a
l
l
y
f
o
r1000i
t
e
r
a
t
i
o
n
s
.I
ti
sc
o
n
c
e
i
v
a
b
l
e t
h
e
yc
o
u
l
dh
a
v
eb
e
c
ns
t
o
p
p
e
de
a
r
l
i
e
r
.Ona
200
H
zP
I
Il
a
p
t
o
pi
tr
e
q
u
i
r
e
dl
e
s
s am
i
n
u
t
ef
o
re
a
c
hr
u
n
.A l
a
r
g
enumbero
f
i
t
e
r
a
t
i
o
n
si
su
s
e
f
u
lt
oi
n
d
i
c
a
t
ei
ft
h
es
o
l
u
t
i
o
ne
s
c
a
p
e
st
h
eg
l
o
b
a
lminimumonceh
a
v
i
n
g
.
tWhil
ei
twasn
o
tc
v
i
d
c
n
ti
n i
sc
x
n
p
l
e
i
ti
sl
i
k
e
l
yp
r
o
b
l
e
md
c
p
e
n
d
e
n
l
o
b
t
a
i
n
e
di
scxampleI
h
eg
l
o
b
a
lminimumwass
i
g
n
i
f
i
c
a
n
t
l
yl
a
r
g
e
ra
nt
h
cl
o
c
a
l
s
i
n
c
ei
ni

J7P912

ti
sc
a
l
l
e
dt
h
r
o
u
g
h
The a
l
g
o
r
i
t
h
mi
sa
v
a
i
l
a
b
l
ci
n SimulatedAnneallng.m. I
l
g
o
r
i
t
h
mi
sa
p
p
l
i
e
dt
oExample9
.
1
whichwasu
s
e
df
o
ri
I
lu
s
t
r
a
t
i
o
n
Call_SA.m.Thea
i
nS
c
c
t
i
o
n9
.
1
.Toe
x
e
c
u
t
ct
h
ea
l
g
o
r
i
t
h
mt
h
ef
o
l
l
o
w
i
n
gm
f
i
l
e
s(
i
n
t
r
o
d
u
c
e
di
nC
h
a
p
t
e
r
6
)a
r
en
c
c
c
s
s
a
r
y
: UpperBound_nVar
GoldSection_nVar.m.Followingthe
v
a
r
i
a
b
l
ep
r
o
b
l
c
r
n
s

twof
i
l
e
sa
r
en
e
c
e
s
s
a
r
yf
o
r
p
r
a
c
t
i
c
eo
fd
r
a
w
i
n
gc
o
n
t
o
u
r
sf
o
rt
W
(
d
e
s
c
r
i
b
i
n
g.
eexample
t
h
es
e
c
o
n
dp
r
i
m
i
l
yt
oq
u
i
c
k
l
ydrawac
o
n
t
o
u
rp
l
o
tu
s
i
n
g
i
x c
a
l
c
u
l
a
t
i
o
n
s
. I
n t
h
i
s c
a
s
e t
h
e f
i
l
e
s a
r
e Example9_1.m a
n
d
ma
Example9_1plo
t
.m.Thel
a
t
t
e
rf
i
l
enamemustb
es
e
ti
nSimulatedAnnealing.m.
Thec
a
l
l
i
n
gs
t
a
t
e
m
e
n
tandv
a
r
i
a
b
l
e
sa
r
ec
x
p
l
a
i
n
e
di
nSlmulatedAnneallng.m.T
h
y
c
a
nber
e
a
dbyt
y
p
i
n
g

m
i
n
i
m
u
r
n
s
.S
e
v
e
r
a
lf
e
a
t
u
r
e
so
ft
h
eSAa
l
g
o
r
i
t
h
ma
r
ei
I
lu
s
t
r
a
t
e
du
s
i
n
gsomen
u
m
e
r
i
c
a
l
e
x
p
e
n
m
e
n
t
s
.

M
i
n
i
m
i
z
e

s
i
n(
0
.
1+ (
X
I-4
):
L+(
'
1
.-4
)

f
(
x
.

x
2
0 0.
2)--4)l
l
'~2'
-~
1+
(
x -4
)
:
L+
(
x
2
1

(
9
.
5
)

F
i
v
en
u
m
e
r
i
c
a
le
x
p
e
r
i
m
e
n
t
sa
r
ep
r
e
s
e
n
t
e
db
e
l
o
w
.I
ts
h
o
u
l
dbeemph
i
z
e
dt
h
a
te
s
e
r
u
n
sc
a
n
n
o
tbed
u
p
l
i
c
a
t
e
d
.O
v
e
r
a
l
l

t
h
ea
l
g
o
r
i
t
h
mworkedv
e
r
yw
e
l
l
.F
o
re
x
a
c
t
l
y
sames
t
a
r
t
i
n
gv
a
l
u
e
sa
n
dnumbero
fi
t
e
r
a
t
i
o
n
s(
c
a
l
l
i
n
gCall_SA.mw
i
t
h
o
u
tany
)eg
l
o
b
a
lminimumwasf
o
u
n
d95%o
ft
h
et
i
m
e
.
changes-about40s
Run1
:S
t
a
r
t
i
n
gd
e
s
i
g
n=[
5
5
]
; s
=
2.7567; Numberofiterations=1000
U
s
e
f
u
lnumbero
fi
t
c
r
a
t
i
o
n
s(whend
c
s
i
g
nc
h
a
n
g
e
so
c
c
u
r
r
e
d
)=383
F
i
n
a
ld
e
s
i
g
n=[
3
.
9
9
9
2
3
.
9
9
9
7
]
; F
i
n
a
lO
b
j
e
c
t
i
v
e=
19.6683
.I
te
s
c
a
p
e
st
h
es
t
a
r
t
i
n
gr
e
l
a
t
i
v
e
F
i
g
u
r
e9
.
6i
1
lu
s
t
r
a
t
e
st
h
et
r
a
c
eo
ft
h
ed
e
s
i
g
nc
h
a
n
g
p
p
e
dbyad
i
f
f
e
r
e
n
tl
o
c
a
lminimum.I
ti
sa
b
l
et
oe
s
c
a
p
e
minimumf
a
i
r
l
ye
a
s
i
l
yt
og
e
t .
t
ot
h
eg
l
o
b
a
lminimumwherei
ti
ss
t
r
o
n
g
l
ya
t
t
r
a
c
dt
oi
t
.be.
c
e
si
nt
h
el
o
c
a
l
minimuma
p
p
e
a
rt
op
r
e
f
e
rt
os
t
a
yi
nt
h
el
o
c
a
lminimums
u
g
g
e
s
t
i
n
gt
h
es
o
l
u
t
i
o
n
sa
r
e
n
o
tb
e
i
n
ge
n
c
o
u
r
a
g
e
d1
0e
s
c
a
p
e
.

Run2
:S
t
a
r
t
i
n
gd
e
s
i
g
n=[
4
4
]
;

0.0362; Numberofi
t
e
r
a
t
i
o
n
s=1
0
0
0

U
s
e
f
u
lnumbero
fi
t
e
r
a
t
i
o
n
s(whend
e
s
i
g
nc
h
a
n
g
so
c
c
u
r
r
e
d
)=9
9
1
F
i
n
a
ld
e
s
i
g
n=[
1
0
0
.
7
21
2
7
.
7
5
]
; F
i
n
a
lO
b
j
e
c
t
i
v
e=0.
128
F
i
g
u
r
e9
.
7i
I
Iu a
t
e
st
h
ct
r
a
c
eo
fes
e
a
r
c
hf
o
rt
h
eg
l
o
b
a
lminimum.Mosto
ft
h
ep
o
i
n
t
s
a
r
en
o
ti
nt
h
ep
l
o
ta
sc
a
nb
es
e
e
nfromt
h
es
o
l
u
t
i
o
n
.T
h
i
ss
o
l
u
t
i
o
nwass
t
a
r
t
e
da
tt
h
e
g
l
o
b
a
lminimum.Th
es
e
a
r
c
hp
r
o
c
e
s
sa
p
p
r
st
owandera
l
lo
v
e
rt
h
ep
l
a
c
ea
n
df
i
n
a
l
l
y
a
p
.T
h
e
r
ei
snop
a
t
t
e
mi
nt
h
l
i
m
i
t
e
dp
o
i
n
t
ss
e
c
nont
h
ef
i
g
u
r
e
e
x
c
e
p
te
o
f
f m
tt
h
es
e
a
r
c
hi
se
n
c
o
u
r
a
g
e
dtocape.Thi
sc
a
nb
et
i
c
dt
ot
h
elow
s
u
g
g
e
s
t
i
o
na
.
m
a
g
n
i
t
u
d
eo
fs

Run3
:S
t
a
r
t
i
n
gd
e
s
i
g
n=[
4
4
]
; s
=1
; Numbero
fi
t
e
r
a
t
i
o
n
s=1
0
U
s
e
f
u
lnumbero
fi
t
e
r
a
t
i
o
n
s(whend
e
s
i
g
nc
h
a
n
g
e
so
c
c
u
r
r
e
d
)=553
F
i
n
a
ld
e
s
i
g
n=[
3
.
8
8
3
2
4
.
3
7
0
9
]
; F
i
n
a
lO
b
j
e
c
t
i
v
e=
19.
173

F
i
g
u
r
e9
.
8i
l
Iu
s
t
r
a
t
h
et
r
a
c
eo
ft
h
ed
e
s
i
g
ns
e
a
r
c
h
.I
nt
h
i
sr
u
nt
h
ev
a
l
u
eo
fs
wass
e
l
t
o1
.Oncea
g
a
i
nt
h
es
t
a
r
t
i
n
gv
a
l
u
eo
ft
h
ed
e
s
i
g
nwast
h
eg
l
o
b
a
lminimumi
t
s
e
l
f
.From
t
h
ef
i
g
ei
ti
sc
1
e
a
rt
h
es
e
a
r
c
hi
sn
o
ta
l
l
o
w
e
dt
oe
s
c
a
p
eeg
l
o
b
a
lminimumt
h
a
t
e
a
s
i
l
y
.Thec
h
o
i
c
eo
ft
h
ec
o
n
t
r
o
lp
a
r
a
m
e
t
e
rwasa
r
b
i
y

a
n
ds
e
v
e
r
a
le
x
n
p
l
e
sh
a
v
e
sv
a
l
u
ei
sp
r
o
b
l
e
mi
n
d
e
p
e
n
d
e
n.
tTh
ev
a
l
u
eo
fIi
st
y
p
i
c
a
l
l
ya
t
ob
es
t
u
d
i
e
d1
0s
e
ei
fi

3
6
2 G
L
O
B
A
LO
P
T
I
M
I
Z
A
T
10
N

9
.
2N
U
M
E
R
I
C
A
LT
E
C
H
N
I
Q
U
E
SA
N
DADOmONALE
X
A
M
P
L
E
S 3
6
3

1
0

a
B

gagNM

-6

2
X1

v
a
l
u
e
s

1
0

F
l
g
u
r
e
9
.
6S
l
m
u
l
a
t
e
da
n
n
e
a
Ji
n
gE
x
a
m
p
l
e9
.
1
c
o
n
v
e
r
g
e
ds
o
l
u
t
i
o
n
.
g
o
o
dc
h
o
i
c
ef
o
ro
p
t
i
m
i
z
a
t
i
o
na
1
g
o
r
ims
p
u
l
a
r
l
yf
o
raw
e
U
s
c
a
1
e
dp
r
o
b
l
e
m
.What
a
1u
eo
fs
o
ft
h
eo
r
d
e
ro
f1i
m
p
l
e
m
e
n
t
s
i
sc
l
e
a
rf
r
o
mt
h
er
u
n
ss
of
a
ri
st
h
a
tan
e
g
a
t
i
v
ev
a
1a
t
c
t
o
r
.
al
o
c
Run4
:S
t
a
r
t
i
n
gd
e
s
i
g
n=5

5
]
;s
=-2.7567;Numberofiterations=1000
5
5
8
U
s
e
f
u
ln
u
m
b
e
ro
fi
t
e
r
a
t
i
o
o
s(
w
h
e
nd
e
s
i
g
nc
h
a
n
g
e
so
c
c
u
r
r
e
d
)=
F
i
n
a
ld
e
s
i
g
n=
[
3
.
3
3
9
7
6
.
2
0
1
9
]
;F
i
n
a
1O
b
j
e
c
t
i
v
e=
2
.
5
6
7
4
Th
i
si
sar
e
p
e
a
to
fRun1
:F
i
9
.
9t
r
a
c
e
st
h
es
e
a
r
c
hf
o
rt
h
eg
l
o
b
a
1minimumf
o
rt
h
i
s
ta
p
p
e
a
r
sa
t1
0
0
0i
t
e
r
a
t
i
o
n
si
sn
o
ts
u
f
f
i
c
i
e
o
tt
od
e
t
e
r
m
i
o
et
h
eg
l
o
b
a
lm
i
n
i
m
u
m
.
r
u
n
.I

W
h
i
l
et
h
es
o
l
u
t
i
o
nh
a
se
s
c
a
p
e
dt
h
es
t
a
r
t
i
n
gl
o
c
a
1minimum
i
th
a
sn
o
ty
e
te
s
c
a
p
e
de
a
11
e
y
s(
l
o
c
a
1m
i
n
i
m
u
m
s
)i
nt
h
i
se
x
a
m
p
l
ea
r
ec
i
r
c
u
l
a
r

n
e
x
tl
o
c
a
lm
i
n
i
m
u
m
.S
i
n
c
et
h
ev
n
o
t
ea
tt
h
es
o
l
u
t
i
o
nmovesa
r
o
u
n
di
nt
h
eva
11
e
yb
u
th
a
sn
o
tmanagedt
o

c
a
p
e
.
55
]
;s
=2
.
7
5
6
7
;Numbero
fi
t
e
r
a
t
i
o
n
s=1
0
0
0
0
Run5
:S
t
a
r
t
i
n
gd
e
s
i
g
n=[
3
8
1
3
U
s
e
f
u
ln
u
m
b
e
ro
fi
t
e
r
a
t
i
o
n
s(
w
h
e
nd
e
s
i
g
nc
h
a
n
g
e
so
c
c
u
r
r
e
d
)=
[
4
.
0
0
3
7
3
.
9
9
9
6
]
;F
i
n
a
lO
b
j
e
c
t
i
v
=1
9
.
6
6
8
2
F
i
n
a
ld
e
s
i
g
n=
Nof
i
g
u
r
ei
sp
r
e
s
e
n
t
e
d
.Thes
o
l
u
t
i
o
nc
o
n
v
e
r
g
e
st
ot
h
eg
l
o
b
a
lm
i
n
i
m
u
m
.

2
- UA

-6

-4

'

l
u

F
1g
u
r
e
9
.
7S
i
m
u
l
a
t
e
da
n
n
e
a
l
i
n
g
o
w
p
.

ntheNumericalExperiments:
'
h
ef
o
l
l
o
w
i
n
go
b
s
e
r
v
a
t
i
o
n
sa
r
e
Discussiono
b
a
s
e
do
nt
h
ee
x
p
e
r
i
e
n
c
eo
fa
p
p
l
y
i
n
gSA(
A
9
.
2
)t
oE
x
a
m
p
l
e9
.1
.T
h
e
yc
o
r
r
e
s
p
o
n
dt
o
er
e
p
o
r
t
e
de
x
p
e
r
i
e
n
c
ew
i
t
hSAi
neg
e
n
e
r
a
ll
i
t
e
r
a
t
u
.

)i
s i
m
p
o
r
t
a
n
tc
o
n
s
i
d
e
r
a
t
i
o
ni
nS
A
.S
t
a
r
t
i
n
g
A
n
n
e
a
l
i
n
gs
c
h
e
d
u
l
e(
c
h
a
n
g
i
n
gs
w
i
t
hal
o
wv
a
1u
eec
o
n
t
r
o
lp
m
e
r
smustbeincreascd-nottoof
a
s
to
ri
t
w
i
l
lg
e
t p
p
e
d al
o
c
a
1m
i
n
i
m
u
m
.
i
sd
e
v
e
l
o
p
e
dt
h
r
o
u
g
ht
r
ia
1a
n
de
r
r
o
r
i
n
s
i
g
h
t i
sn
o
t
A
n
n
e
a
l
i
n
gs
c
h
e
d
u
l
o
b
v
i
o
u
s
a
n
db
a
s
e
do
np
r
o
b
a
b
l
eb
e
h
a
v
i
o
ro
ft
h
es
h
.
r
o
b
l
e
mn
e
e
d
st
ob
es
e
a
r
c
h
c
d
L
e
n
g
t
ho
ft
i
m
e(
o
rn
u
m
b
e
ro
fi
t
e
r
a
t
i
o
n
s
)t
h
a
tep
i
sd
i
f
f
i
c
u
l
tt
od
t
e
r
m
i
n
e
.W
h
i
l
ei
sc
a
nb
eb
a
s
e
do
nm
o
n
i
t
o
r
i
n
g
g
i
v
e
nt
h
e

i
tc
a
n
n
o
tb
ea
s
s
e
r
t
e
dt
h
a
teg
l
o
b
a
lminimumh
a
s
s
t
o
c
h
a
s
t
i
cn
a
t
u
r
co
ft
h
es
e
a
r
c
h
1
i
s
h
e
d
.TheSAa
l
g
o
r
i
t
h
mi
t
s
e
l
fc
a
nb
ee
x
p
e
c
t
e
dt
oc
o
n
v
e
r
g
ew
h
i
l
et
h
i
s
b
e
e
ne
s
t
a
b
p
r
o
p
e
r
t
yi
st
o
os
l
o
wf
o
rp
r
a
c
t
i
c
a
lu
s
e[
6
]
.Al
a
r
g
e
rn
u
m
b
e
ro
fi
t
e
r
a
t
i
o
n
si
sm
o
r
e
a
1s
o
l
u
t
i
o
n
.
l
i
k
e
l
yt
oy
i
e
l
dt
h
eg
l
o
b
SAs
e
a
r
c
hd
o
e
sn
o
ts
u
b
s
c
r
i
b
et
oi
t
e
r
a
t
i
v
ei
m
p
r
o
v
e
m
e
n
t
.I
ti
5a
ne
x
a
m
p
l
eo
fa
ei
m
p
l
e
m
e
n
t
a
t
i
o
nu
s
c
dh
e
r
e P
s
e
o
t
sa
s
t
o
c
h
a
s
t
i
ce
v
o
l
u
t
i
o
n
a
r
yp
r
o
c
e
s
s
.Th
m
e
m
o
r
y
l
e
s
ss
e
.

"

3t

364

9
.
2 N
U
M
E
R
I
C
A
LT
E
C
H
N
I
Q
U
E
SANDA
D
D
I
T
I
O
N
A
LE
X
A
M
P
L
E
S 365

G
L
O
B
A
LOP
T
lM
I
Z
A
T
I
O
N

-4

-2

0
2

v
a
l
u
e
s

1
0

F
l
g
u
r
e
9
.
8S
c
r
e
a
s
d
opp

u
n
i
t
yt
oe
s
c
a
p
e
.
l
m
u
l
a
l
e
da
n
n
e
a
l
l
n
g
:I
n
c
r
e
a
s
l
n
g-de

R
e
s
t
t
i
n
gfromd
i
f
f
e
r
c
n
tds
i
g
nmusts
t
i
Ub
ei
m
p
l
e
m
e
n
t
e
dt
oovercomes
t
r
o
n
g
l
o
c
a
li
n
f
l
u
e
n
c
e
s
.
i
sa
1s
oi
m
p
o
r
t
a
n
tf
o
re
s
t
a
b
l
i
s
h
i
n
g
R
e
s
t
a
r
t
i
n
g fromt
h
esamei
n
i
t
i
a
l de
c
o
n
v
c
r
g
e
n
c
ep
r
o
p
e
e
sa
si
si
sas
t
o
c
h
a
s
t
i
cs
e
a
r
c
hp
r
o
c
e
s
s
.

ConstrslnedProblems:Thi
sc
h
a
p
t
c
rd
o
e
sn
o
td
e
a
lw
i
t
hc
o
n
s
t
r
a
i
n
l
so
rd
i
s
c
r
c
t
c
e
r
ca
r
ctwowayso
fh
a
n
d
l
i
n
gc
o
n
s
t
r
a
i
n
t
sd
i
r
c
c
t
l
y
p
r
o
b
l
c
m
s
. l
1
. l
ft
h
enewd
e
s
i
g
np
o
i
n
ti
sn
o
tf
e
a
s
i
b
l
e
i
ti
sn
o
ta
c
c
e
p
t
e
dandnod
c
s
i
g
nc
h
a
n
g
e
t
a
k
p
l
a
c
e
.Th
i
si
sq
u
i
t
e1 c
t
i
v
cf
o
rrandoms
hd
i
r
c
c
t
i
o
n
s
.
ef
u
n
c
t
i
o
na
td
r
i
v
c
st
h
ea
1
g
o
r
i
t
h
md
u
r
i
n
ge
a
c
hi
t
e
r
n
t
i
o
ni
st
h
eonew
i
t
ht
h
e
2
. l
l
a
r
g
c
s
tv
i
o
l
a
t
i
o
n
.I
ft
h
ec
u
r
r
e
n
tp
o
i
n
ti
sf
e
a
s
i
b
l
e
t
h
e
nt
h
eo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
nw
i
l
l
e
a
r
c
h
.Th
i
smakes
e
n
s
e ep
r
o
c
e
s
si
sl
a
r
g
e
l
ys
t
o
c
h
a
s
t
i
candi
s
d
r
i
v
e s
.e
x
p
e
c
t
e
dt
owandera
r
o
u
n
d ed
e
s
i
g
ns
p
a
c
e
.
Cons
n
cdp
r
o
b
l
c
m
sc
b
ee
x
p
e
c
t
e
dt
ot
a
k
eas
i
g
n
i
f
i
c
a
n
t
l
yI
a
r
g
e
rnumbcro
f
S
r
c
h
e
s
.Th
e
i
sc
u
r
r
e
n
t
l
ynop
r
'
u
r
et
oi
d
e
n
t
i
f
yc
o
n
v
e
r
g
e
n
c
cp
r
o
p
e ofthcSA
p
t
t
oc
o
n
s
t
r
a
i
n
e
do
p
t
i
m
i
z
a
t
i
o
n
.
w
i
t
h s

g
a
E
'
N
H

-8

bp qdNJStAqdfd
2404

gaE'
-8

-6

-4

-2

-val
S

1
0

l
m
u
l
a
l
e
da
n
n
e
a
l
i
n
g
:n
o
n
F
l
g
u
r
e
9
.
9S
o
n
v
e
r
g
e
n
c
e-ln
s
u
l
f
i
c
i
e
n
l
l
l
e
r
a
l
l
o
n
s
.

D
l
s
c
r
e
t
eProblems: I
nd
i
s
c
r
c
t
ep
r
o
b
l
e
m
s
t
h
eSAa
l
g
o
r
i
t
h
mw
i
l
lu
s
ead
i
s
c
r
e
t
e
d
e
s
i
g
nv
e
c
t
o
rd
e
t
u
neds
t
o
c
h
a
s
t
i
c
a
l
l
yomt
h
ea
va
i
1a
b
l
es
e
tEache
l
e
m
e
n
to
fe
d
e
s
i
g
nv
e
c
t
o
ri
sr
a
n
d
o
m
l
yc
h
o
s
e
nfromep
e
r
m
i
s
s
i
b
l
es
e
to
fv
a
l
u
e
s
.Th
er
e
m
a
i
n
i
n
g
e
l
e
m
e
n
t
so
ft
h
SAa
l
g
o
r
i
t
h
ma
r
ed
i
r
e
c
t
l
ya
p
p
l
i
c
a
b
l
e
.I
feo
b
j
e
c
v
ed
e
c
r
c

c
s
t
h
e
n
t
h
ed
e
s
i
g
nc
h
a
n
g
ei
sa
c
c
e
p
t
e
d
.l
fi
ti
n
c
r
e
a
s
c
st
h
e
nac
o
n
d
i
t
i
o
n
a
lp
r
o
b
a
b
i
l
i
t
yo
f
a
c
c
c
p
t
a
n
c
ei
sd
e
t
e
r
m
i
n
e
da
l
l
o
w
i
n
gt
h
es
o
l
u
t
i
o
nt
oe
s
c
a
p
et
h
cminimum.T
h
i
si
sl
c
f
tt
o
a
d
e
r
.Th
es
i
g
n
i
f
i
c
a
n
te
f
f
o
r
ti
st
h
ei
d
e
n
t
i
f
i
c
a
t
i
o
no
ferandomd
e
s
i
g
nv
e
c
t
o
r
.
Example9
.
2 T
h
i
si
saf
u
n
c
t
i
o
no
ftwov
a
r
i
a
b
l
e
st
h
a
tMATLABu
s
e
sf
o
ri
I
lu
s
t
r
a
t
i
n
gt
h
e
g
r
a
p
h
i
c
sc
a
p
a
b
i
l
i
t
yo
f s
o
f
t
w
a
r
e
.Typingp
e
a
k
sa
tt
h
eCommandpromptw
i
l
ldraw
t
.Typingpep
e
a
k
s
.
ma
tt
h
eCommandpromptw
i
l
lp
r
o
v
i
d
ee
t
h
es
u
r
f
a
c
ep
l
o
u
a
t
i
o
n
st
h
a
twereu
s
c
dt
oo
b
t
a
i
nt
h
ec
u
r
v
e
.T
h
i
sf
u
n
c
t
i
o
nu
s
e
da
seo
b
j
e
c
t
i
v
ci
s
a
v
a
i
l
a
b
l
ei
nExample9_2.mandExample9_2plo
t
.m.SAi
su
s
e
d~o i
d
e
n
t
i
f
yt
h
e
g
l
o
b
a
lminimum.F
i
g
u
r
e9
.
1
0i
l
l
u
s
t
r
a
t
e
st
h
ec
o
n
t
o
u
rp
l
o
ta
10ngwi changesi
ne
d
e
s
i
g
n
.Th
ef
u
n
c
t
i
o
nh
a
stwol
o
c
a
lminimumsandt
h
r
l
o
c
a
lmaximums.Th
e
f
o
l
l
o
w
i
n
gr
e
p
r
c
s
e
n
t
sac
o
n
v
e
r
g
e
dr
u
n
.
g
n=[
2
2
]
;2.
5
;Numbcrofi
t
e
r
a
t
i
o
n
s=1
0
0
0
S
t
a
r
t
i
n
g i
U
s
e
f
u
lnumbero
fi
t
e
r
a
t
i
o
n
s(whend
e
s
i
g
o.
c
h
a
n
g
c
so
c
c
u
r
r
c
d
)=
3
3
8

366 G
L
O
B
A
LOP
T
lM
I
Z
A
T
lO
N

-4

-3

-2

9
.
2N
U
M
E
R
I
C
A
LT
E
C
H
N
I
Q
U
E
SANDA
O
O
I
T
I
O
N
A
LE
X
A
M
P
L
E
S 367

X
1

v
a
l
u
e
s

F
l
g
u
r
e9
.
1
0S
f
m
u
l
a
t
e
da
n
n
e
a
l
i
n
g
:E
x
a
m
p
l
e9
.
2
.

F
i
n
a
ld
e
s
i
g
n [
0
.
0
2
2
7
9
1
.6
2
5
9
]
;F
i
n
a
lO
b
j
e
c
t
i
v
e 6
.
5
5
1
1

T
h
isi
st
h
eg
l
o
b
a
lminimumo
ft
h
ep
r
o
b
l
e
m
.
Th
e
r
ewasas
i
g
n
i
f
i
c
a
n
tc
h
a
n
g
emadet
oeSAa
l
g
o
r
i
t
h
mb
a
s
e
doni
n
i
t
i
a
l s
t
s
w
i

b
u
t
f
o
rd
i
s
c
o
v
e
r
i
n
goptimums
o
l
u
t
i
o
n
s
.H
o
l
l
a
n
d[
7
]wast
h
ef
i
r
s
tt
ou
s
ct
h
et
e
c
h
n
i
q
u
e
iusea
sa
no
p
t
i
m
i
z
a
t
i
o
nt
o
o
lb
e
g
a
ni
ne
a
m
e
s
ti
nel
a
t
c1
9
8
0
8

d
e
v
e
l
o
p
e
dmomentum
i
nt
h
em
i
d
1
9
9
0
s
a
n
dc
o
n
t
i
n
u

t
oa
t
t
r
a
c
ts
e
r
i
o
u
si
n
t
e
r
e
s
tt
o
y
.Th
et
e
r
me
v
o
l
u
t
i
o
n
a
r
y
i
ss
u
g
g
e
s
t
i
v
eo
fen
a
t
u
r
a
lp
r
o
c
e
s
sa
s
s
o
c
i
a
t
e
dwi b
i
o
l
o
g
i
c
a
le
v
o
l
u
t
i
o
n
p
r
i
m
a
r
i
l
y
t
h
eO
a
r
w
i
n
i
a
nr
u
l
o
ft
h
es
e
l
e
c
t
i
o
no
ft
h
ef
i
t
t
e
st
.Thi
st
a
k
e
sp
l
a
c
eb
e
c
a
u
s
eec
o
n
s
t
a
n
t
o
p
u
l
a
t
i
o
ny
i
e
l
dab
e
t
t
e
rg
e
n
e
m
u
t
a
t
i
o
nandr
e
c
o
m
b
i
n
a
t
i
o
no
ft
h
echromosomesi
nep
S
t
r
u
c
t
U
I
. A
1
1o
ft
h
e
s
et
e
r
m
sa
p
p
e
i
nt
h
ed
i
s
c
u
s
s
i
o
n
sr
e
l
a
t
e
dt
oe
v
o
l
u
t
i
o
n
a
r
y
h
isb
a
s
i
cp
r
o
s
si
sa
l
s
of
o
l
d
e
di
n
t
oev
a
r
i
o
u
sn
u
m
e
r
i
c
a
lt
e
c
h
n
i
q
u
e
s
.
c
o
m
p
u
t
a
t
i
o
n
.T
Thed
i
s
c
u
s
s
i
o
n
sc
a
nbecomec
l
e
a
ri
f t
e
r
mc
h
r
o
m
o
s
o
m
e
sc
a
nb
ea
s
s
o
c
i
a
t
e
dw
i
t
h
d
e
s
i
g
nv
a
r
i
a
b
l
e
s
.Overt
h
ey
E
v
o
l
u
t
i
o
n
a
r
yC
o
m
p
u
t
a
t
i
o
nh
a
si
n
c
l
u
d
e
dG
e
n
e
t
i
c
Programming-whichd
e
v
e
l
o
p
sp
r
o
g
r
a
r
n
s r
n
i
r
n
i
c
kee
v
o
1
u
t
i
o
n
a
r
yp
r
o
c
;
G
e
n
e
t
i
cA
l
g
o
r
i
t
h
m
s
o
p
t
i
r
n
i
z
a
t
i
o
no
fc
o
m
b
i
n
a
t
o
r
i
a
V
d
i
s
c
r
e
t
ep
r
o
b
1
e
m
s
;E
v
o
l
u
t
i
o

r
y
Programming-optimizing c
o
n
t
i
n
u
o
u
s f
u
n
c
t
i
o
n
s w
i
t
h
o
u
t r
e
c
o
m
b
i
n
o
n
;
E
v
o
l
u
t
i
o
n
a
r
yS
t
r
a
t
e
g
i
e
s
o
p
t
i
r
n
i
z
i
n
gc
o
n
t
i
n
u
o
u
sf
u
n
c
t
i
o
n
sw
i
t
hr
e
c
o
m
b
i
n
a
t
i
o
n[
8
]
.
Th
i
sc
1a
s
s
i
f
i
c
a
t
i
o
ni
sq
u
i
t
el
o
o
s
e
.F
o
rexample

GAsh
a
v
eb
e
e
nu
s
e
d

a
n
dc
o
n
t
i
n
u
et
o
o
n
a
r
y
b
eu
s
e
de
f
f
e
c
t
i
v
e
l
yt
os
o
l
v
ec
o
n
t
i
n
u
o
u
so
p
t
i
m
i
z
a
t
i
o
np
r
o
b
1
e
m
s
. Evoh
a
l
g
o
r
i
t
h
m
sh
a
v
eb
e
e
ns
u
c
c
e
s
s
f
u
l
l
ya
p
p
l
i
e
dt
oo
p
t
i
m
i
z
a
t
i
o
np
r
o
b
1
c
m
si
ns
e
v
e
r
a
la
r
e
a
s
:
e
n
g
i
n
e
c
r
i
n
gd
e
s
i
g
n
p
a
r
a
r
n
e
t
e
rf
i
t
t
i
n
g
k
n
a
p
s
a
c
kp
r
o
b
l
e
m
s
t
r
a
n
s
p
o
r
t
a
t
i
o
np
r
o
b
1
e
m
s
imagep
r
o
c
e
s
s
i
n
g
t
r
a
v
e
l
i
n
gs
a
l
e
s
m
a
n
s
c
h
e
d
u
l
i
n
g
.ands
oo
n
.Asc
a
nb
ei
n
f
e from
el
i
s
t
.as
i
g
n
i
f
i
c
a
n
tnumbera
r
eomt
h
ca
r
c
ao
fd
i
s
c
r
c
t
cp
r
o
g
r
a
m
m
i
n
g
.
sbooki
n
c
l
u
d
eo
n
l
yGAsthougha
l
1o
ft
h
ee
v
o
l
u
t
i
o
n
a
r
y
T
h
i
ss
e
c
t
i
o
nandi
a
l
g
o
r
i
t
h
m
ss
h
a
r
e many common f
e
a
t
u
r
e
s
. GAs g
e
n
e
r
a
t
e and u
s
es
t
o
c
h
a
s
t
i
c
i
n
f
o
r
m
a
t
i
o
ni
nt
h
e
i
ri
m
p
l
e
m
e
n
t
a
t
i
o
n and t
h
e
r
e
f
o
r
ec
a
n bec
o
n
s
i
d
e
r
e
dg
l
o
b
a
l
l
1
yu
s
e
f
u
lf
o
rh
a
n
d
l
i
n
gi
l
lbehaved
o
p
t
i
m
i
z
a
t
i
o
nt
e
c
h
n
i
q
u
e
s
.Theya
r
ee
x
c
e
p
t
i
o
n
a
d
i
s
c
o
n
t
i
n
u
o
u
s
.andn
o
n
d
i
f
f
e
r
e
n
t
i
a
b
l
ep
r
o
b
l
e
m
s
.U
n
l
i
k
et
h
eSA
c
o
n
v
e
r
g
e
n
c
et
ot
h
e
g
l
o
ba
1optimumi
nGAsc
a
no
n
l
ybeweak1ye
s
t
a
b
l
i
s
h
e
d
.GAsg
e
n
e
r
a
t
eandu
s
e
p
o
p
u
l
a
t
i
o
no
fs
o
l
u
t
i
o
n
s
d
e
s
i
g
nv
e
c
t
o
r
s(X)-u
n
1
i
ke SAwhereas
e
a
r
c
h
d
i
r
e
c
t
i
o
n(
8
) was g
e
n
e
r
a
t
e
d
. Thi
sd
i
s
t
i
n
c
t
i
o
ni
se
s
s
e
n
t
i
a
lt
ou
n
d
e
r
s
t
a
n
dt
h
e
d
i
f
f
i
c
u
l
t
yo
fh
a
n
d
l
i
n
gc
o
n
t
i
n
u
o
u
sp
r
o
b
l
e
m
si
nOAs.Fore
x
a
m
p
l
e
.u
s
i
n
grandom
numberst
ogn
e
r
a
t
eav
a
l
u
ef
o
rXlot
h
ev
a
l
u
e
so
f3
.
1
4
3
.
1
4
1
o
r3
.
1
4
1
5w
i
l
lal
1be
d
i
s
t
i
n
c
tandhavenor
e
l
a
t
i
o
nt
oe
a
c
ho
t
h
e
r
.

eex
u
t
i
o
n
.Att
h
es
t
a
r
t
i
n
gd
e
s
i
g
nt
h
ed
e
s
i
g
nwasl
a
r
g
e
l
yp
o
s
i
t
i
v
e(
8
.
9
5
)
.Changing
s
o
l
u
t
i
o
nt
oe
s
c
a
p
et
h
el
o
c
o
lrninimumo
rt
o
i
tt
oan
e
g
a
t
i
v
enumberd
i
dn
o
tp
e
r
m
i
tt
h
swingo
r
o
u
n
dt
h
en
e
a
r
b
yl
o
c
a
lmaximum.N
u
m
e
r
i
c
a
le
x
p
e
r
i
m
e
n
l
ss
u
g
g
e
s
t e
c
o
n
s
t
a
n
t3
fbemaintainedbetween

2.
5s
3
fS-1

(
9
.
6
)

O
t
h
e
r
w
i
s
enoc
h
a
n
g
ewasmadet
oea
l
g
o
r
i
t
h
m
.
9
.
2
.
2 GeneticAlgorithm(GA)
l
eG
e
n
e
t
i
cA
l
g
o
r
i
t
h
mi
sa
ni
m
p
o
r
t
a
n
tp
a
r
to
f0newa
r
e
ao
fa
p
p
l
i
e
dr
e
s
e
a
r
c
ht
e
r
m
e
d

E
v
o
l
u
t
i
o
n
a
r
yC
o
m
p
u
t
a
t
i
o
n
.Thcsea
r
ea
c
t
u
a
l
l
ys
e
a
r
c
hp
r
o
c

s
e
sandn
a
t
u
r
a
l
l
yu
s
e
f
u
l

GenerlcGenetlcAlgorlthm(GGA)(
A
9
.
3
)
S
t
e
p1
:S
e
tupa
ni
n
i
t
i
a
lp
o
p
u
l
a
t
i
o
nP(O)-ani
n
i
t
ia
1s
e
to
fs
o
l
u
t
i
o
no
rchromosomes
E
v
a
l
u
a
t
e i
n
i
1s
o
l
u
t
i
o
nf
o
rf
i
t
n
e
s
s
d
i r
e
n
t
i
a
t
e
.c
o
l
l
a
t
eandr
a
t
e
s
o
l
u
t
i
o
n
s
G
e
n
e
r
a
t
i
o
ni
n
d
e
xI=
0
S
t
e
p2
: Useg
e
n
e
l
i
c
:o
p
e
r
a
l
o
r
st
og
e
n
e
r
a
t
et
h
es
e
to
fc
h
i
l
d
r
e
n(
c
r
o
s
s
o
v
e
r

m
u
t
a
t
i
o
n
)

l~m;;ii;~

368

GL08ALOPTIMIZATION

Addanews
e
to
frandomlyg
e
n
e
r
a
t
e
dp
o
p
u
l
a
t
i
o
n(
i
m
m
i
g
r
a
l
l
t
s
)

R
e
e
v
a
l
u
a
t
e population-fitness
P
c
r
f
o
r
mc
o
m
p
e
t
i
t
i
v
es
e
l
e
c
t
i
o
n

whichmcmberswi
l
1bep
a
r
to
fn
e
x
t
g
e
n
e
r
a
t
i
o
n
S
e
l
e
c
t
p
o
p
a
t
i
o
nP
(
t+l)-samenumberofmembers

I
fnotconvergedtt+1
GoToStep2
Thev
a
r
i
o
u
st
e
r
m
sa
r
ee
x
p
l
a
i
n
e
db
e
l
o
w
.

9
.
2 NUMERICALTECHNIQUESANDADDITION
AlEXAMPLES

G
e
n
e
t
i
cO
p
e
r
a
t
o
r
s
:l
e
s
ep
r
o
v
i
w
ays o
fd
e
f
i
n
i
n
g new p
o
p
u
l
a
t
i
o
n
sf
r
o
m
e
x
i
s
t
i
n
go
n
e
s
.lIsi
sa
l
s
owheret
h
ee
v
o
l
u
t
i
o
nt
a
k
e
sp
l
a
c
e
.Twot
y
p
e
so
fo
p
e
r
a
t
i
o
na
r
e
r
e
c
o
g
n
i
z
e
d
.Thef
i
r
s
ti
sc
r
o
s
s
o
v
e
ro
rr
e
c
o
m
b
i
n
a
t
i
o
n
andt
h
es
e
c
o
n
dm
u
t
a
t
i
o
n
.
r
es
e
v
e
r
a
lt
y
p
e
so
fc
r
o
s
s
o
v
e
rs
t
r
a
t
e
g
i
e
s
.A s
i
m
p
l
e
C
r
o
s
s
o
v
e
rO
p
e
r
a
t
o
r
s
:Therea
c
r
o
s
s
o
v
e
ri
si
l
I
t
r
a
t
e
di
nF
i
g
u
r
e9
.
11
.Hereap
i
e
c
eo
f echromosome-an
a
1
1
e
l
e
i
se
x
c
h
a
n
g
e
db
e
t
w
e
e
ntwop e
n
t
sP
" P2t
op
r
o
d
u
c
ec
h
i
l
d
r
e
nC"C2 Th
e
l
e
n
g
o
ft
h
ep
i
e
c
ei
su
s
u
a
l
l
yd
e
t
e
r
m
i
n
e
dr
a
n
d
o
m
l
y
.M
a
t
h
e
m
a
t
i
c
a
l
l
y

i
ft
h
ep
a
r
e
n
t
s e
]a
X [x"X2
.
.

Xn
ndY(
d
e
f
i
n
e
ds
i
m
i
l
a
r
l
y
)
andri
st
h
er
a
n
d
o
m
l
ys
e
l
e
c
t
e
dt
r
u
n
c
a
t
i
o
n
i
n
d
e
x
.e
nt
h
ec
h
i
l
d
r
e
nU(=C1
)a
ndV(=C2)
bys
i
m
p
l
ec
r
o
s
s
o
v
e
ra
r
ed
e
f
i
n
e
da
s[
1
1
]

x
r
t11r14llL
.

i
f
i<r
oe
r
w
i
s

i
fi<r

(
9
.
7
)

o
t
h
e
r
w
i
s
e

A
r
i
t
h
m
e
t
i
cc
r
o
s
s
o
v
e
ri
so
b
t
a
i
n
e
dbyaI
in
e
a
rc
o
m
b
i
n
a
t
i
o
no
ftwop
a
r
e
n
tchromosomes
yieldtwoc
h
i
l
d
r
e
n[
9
]
.I
fI
l
.1a
nd^
'
za
r
erandomlyg
e
n
e
r
a
t
e
dnumberst
h
e
nt
h
e
c
h
i
l
d
r
e
nc
a
nb
ed
e
f
i
n
e
d

yy

++

hh

xx

hL.

qq

v
a
l
u
e
.As
e
l
e
c
t
i
o
nfromt
h
i
sp
o
p
u
l
a
t
i
o
ni
sn
e
c
e
s
s
yf
o
r(
1
)i
d
e
n
t
i
f
y
i
n
gp
a
r
e
n
t
sand(
2
)
i
n
g
i
fe
ywi
1
1bepromodtoenextg
e
n
e
r
a
t
i
o
n
.Therea
r
ev
a
r
i
o
u
ss
e
l
e
c
t
i
o
n
i
d
e
n
t
i
s
c
h
e
m
e
s
.A f
r
a
c
t
i
o
no
ft
h
eb
e
s
ti
n
d
i
v
i
d
ua
1scanbechosenf
o
rr
e
p
r
o
d
u
c
t
i
o
na
n
d
p
r
o
m
o
t
i
o
n
t
h
er
e
m
a
i
n
i
n
gp
o
p
u
l
a
t
i
o
nb
e
i
n
gmadeupo
fnewi
m
m
i
g
r
a
n
t
s
.Th
i
si
sc
a
11
d
t
o
u
r
n
a
m
e
n
ts
e
l
e
c
t
i
o
n
.A m
o
d
i
f
i
c
a
t
i
o
no
fi
sp
r
o
c
e
d
u
r
ei
st
oa
s
s
i
g
nap
r
o
b
a
b
i
l
i
t
yo
f
s
e
l
e
c
t
i
o
nf
o
r o
r
d
e
r
e
ds
e
to
ft
h
ep
o
p
u
l
a
t
i
o
nf
o
rbo
r
e
p
r
o
d
u
c
t
i
o
nandp
r
o
m
o
t
i
o
n
.
Th
i
sp
r
o
b
a
b
i
l
i
t
yc
a
nbeb
a
s
e
donr
o
u
l
e
t
t
ew
h
e
e
l

e
l
i
t
i
s
ts
e
l
e
c
t
i
o
n
l
i
n
e
a
rorg
e
o
m
e
t
r
i
c
r
a
n
k
i
l
l
g[
10
]
.Thei
m
p
l
e
m
e
n
t
a
t
i
o
ni
nt
h
i
sbooku
s
e
sr
a
n
k
i
n
gw
i
t
h
o
u
ta
s
s
i
g
n
i
n
ga
p
r
o
b
a
b
i
l
i
t
yo
fs
e
l
t
i
o
n
.1
tu
s
e
st
h
eo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
nv
a
l
u
et
or
a
n
ks
o
l
u
t
i
o
n
s
.

uv

n
c
t
i
o
n
.

b
ep
o
p
u
l
a
t
i
o
nc
a
nber
a
n
k
e
d
l
s
o
r
t
e
da
c
c
o
r
d
i
n
gt
oeo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
n

--

Chromosomes:Whilei
ti
sa
p
p
r
o
p
r
i
a
t
et
or
e
g
a
r
de
a
c
hchromosomea
sad
e
s
i

v
e
c
t
o
r

t
h
e
r
ei
s i
s
s
u
eo
fr
e
p
r
e
s
e
n
t
a
t
i
o
no
ft
h
ed
e
s
i
g
nv
e
c
t
o
rf
o
rh
a
n
d
l
i
n
gbyg
e
n
e
t
i
c
op
r
a
t
o
r
s
.I
ti
sp
o
s
s
i
b
l
et
oworkw
i
t
h d
e
s
i
g
nv
e
c
t
o
rd
i
r
e
c
t
l
y(
a
si
nei
n
c
l
u
d
e
d
e
x
a
m
p
l
e
)o
ru
s
esomek
i
n
do
fmappingl
t
1(
r
e
a
le
n
c
o
d
i
n
g
)o
rb
i
n
a
r
y(
b
i
n
a
r
y
e
n
c
o
d
i
n
g
)
.E
a
r
l
i
e
rworki
nGAu
s
e
db
i
n
a
r
ye
n
c
o
d
i
n
g
.I
ni
sbookt
h
ed
e
s
i
g
nv
e
c
t
o
r
i
su
s
e
dd
i
r
e
ct
1yf
o
ra
11t
r
a
n
s
a
c
t
i
o
n
s
.R
e
a
lc
n
c
o
d
i
n
gi
s commended.A p
i
e
c
eo
ft
h
e
chromosomei
sc
a
l
l
e
da
na
l
l
e
l
e
.
F
i
t
n
e
s
sandS
e
/
e
c
t
i
o
n
:The f
i
t
n
e
s
sf
u
n
c
t
i
o
nc
a
nb
er
e
l
a
t
e
dt
ot
h
co
b
j
e
c
t
i
v
e

(
9
.
8
)

I
f

1;

i' 0
i
ti
saconvexI i
n
a
t
i
o
n

I
n
l
t

I
P
o
p
u
l
a
t
i
o
n
:Theparametrherei
st
h
enumbero
fi
n
i
t
ia
1d
e
s
i
g
nv
e
c
t
o
r
s
(
c
h
r
o
m
o
s
o
m
e
s
)f
o
res
t
a
r
t
i
n
gg
e
n
e
r
a
t
i
o
n
Np' numberi
su
s
u
a
l
l
yk
e
p
tt
h
esame
i
ns
u
c
c
e
s
s
i
v
eg
e
n
e
r
a
t
i
o
n
s
.U
s
u
a
l
l
yanorm
a
1randomg
e
n
e
r
a
t
i
o
no
ft
h
ed
e
s
i
g
nv
e
c
t
o
r
s
d
o
p
i
n
g
addingsomegoods
o
l
u
t
i
o
n
st
oe
i
srecommended.I
nsomei
n
s
t
a
n
c
e
s
p
o
p
u
l
a
t
i
o
ni
sp
e
r
f
o
r
m
e
d
.I
ti
sd
i
f
f
i
c
u
l
tt
oe
s
t
a
b
l
i
s
hhowmanymembersmustbe
c
o
n
s
i
d
e
r
e
dt
or
e
p
r
e
s
e
n
tt
h
ep
o
p
u
l
a
t
i
o
ni
ne
v
e
r
yg
e
n
e
r
a
t
i
o
n
.

E
v
a
l
u
a
t
i
o
n
:Thev
a
r
i
o
u
sd
e
s
i
g
nv
t
o
r
sh
a
v
et
ob
ee
va
1u
a
t
e
dt
os
i
feycanbe
u
s
e
df
o
rc
r
e
a
gc
h
i
l
d
r
e
na
tc
a
nbecomep
a
r
to
ft
h
en
e
x
tg
e
n
e
r
a
t
i
o
n
.F
i
t
n
e
s
s
a
t
t
r
i
b
u
t
e
sc
a
ns
e
r
v
w
e
l
lf
o
rt
h
i
spu
o
s
e
.Fort
h
ec
a
s
eo
fu
n
c
o
n
s
t
r
a
i
n
e
doptimum

t
h
e
o
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
nJ
(
X)i
sagoodm
e
a
s
u
r
e
.S
i
n
c
e g
l
o
b
a
lminimumi
sunknown
a
p
r
o
g
r
e
s
s
i
v
em
e
a
s
u
r
emayb
eu
s
e
d
.

369

Pt
P
a
r
e
n
t
s
g
ure9

P2

Ct

C2

Childrcn

...

JA4 1~'r.!.iIt";1:.;L'.:m~~

l
J
u
U
F
I
F
t
i
j
i
n
J
r
RHU--

1."1':':W ~r~ 1:.:c.

370

GLOBALOPTIMIZATION

I
ft
h
e
r
ea
r
en
or
e
s
t
r
i
c
t
i
o
n
so
n
'
s

i
ti
sa
nq
n
ec
o
m
b
i
n
a
t
i
o
n
.F
o
rr
e
a
l

'
si
ti
sal
i
n
e
a
r
c
o
m
b
i
n
a
t
i
o
n
. An i
m
p
o
r
t
a
n
tc
r
o
s
s
o
v
e
ro
p
e
r
a
t
o
ri
s ad
i
r
e
c
t
i
o
n
b
a
s
e
dc
r
o
s
s
o
v
e
r
o
p
e
r
a
t
o
r

i
f
J
t
Y
)i
sl
e
s
st
h
a
nJ
t
X)a
n
dri
sau
n
i
tr
a
n
d
o
mn
u
m
b
e
r
:

c=r(Y-X)+Y
M
u
t
a
t
i
o
n
:M
u
t
a
t
i
o
nr
e
f
e
r
st
oer
e
p
l
a
c
e
m
e
n
to
fas
i
n
g
l
ee
l
e
m
e
n
to
fad
e
s
i
g
nv
e
c
t
o
r
b
yar
a
n
d
o
m
l
yg
e
n
e
m
t
e
dv
a
l
u
e
.Thee
l
m
e
n
ti
sa
l
s
or
a
n
d
o
m
l
yc
h
o
s
e
n
.F
o
r d
e
s
i
g
n
v
e
c
t
o
rX
f
o
rar
a
n
d
o
m
l
ys
e
l
e
c
t
e
de
l
e
m
e
n
tk

c
=
J
Xi*k
j

i=k
I
X
k

T
h
e
r
ea
r
eo
t
h
r
t
y
p
e
so
fm
u
t
a
t
i
o
ni
n
c
l
u
d
i
n
gn
o
n
u
n
i
f
o
r
mm
u
t
a
t
i
o
na
n
dd
i
r
e
c
t
i
o
n
a
l
m
u
t
a
t
i
o
n
.Ther
e
a
d
e
ri
sr
e
f
e
r
r
e
dt
oR
e
f
e
r
e
n
c
e9f
o
rar
e
l
e
v
a
n
td
e
s
c
r
i
p
t
i
o
n
.
Im
m
.
'
f
a
n
t
s
:T
h
i
si
sea
d
d
i
t
i
o
no
fas
e
t0'
1l
'
a
n
d
o
m
l
yg
e
n
e
r
a
t
e
dp
o
p
u
l
a
t
i
o
ni
ne
a
c
h
e
x
tg
e
n
e
r
a
t
i
o
n
.Thes
e
a
r
c
hp
r
o
c
e
s
si
n
g
e
n
e
r
a
t
i
o
nb
e
f
o
r
et
h
es
e
l
e
c
t
i
o
ni
smadef
o
ren
GAi
sb
r
o
a
d
l
yc
l
a
s
s
i
f
i
e
d p
l
o
r
a
t
i
o
na
n
de
x
p
l
o
i
t
a
t
i
o
n
.E
x
p
l
o
r
a
t
i
o
ni
m
p
l
i
e
s
s
e
a
r
c
h
i
n
ga
c
r
o
s
st
h
ee
n
t
i
r
ed
e
s
i
g
ns
p
a
c
e
.E
x
p
l
o
i
t
a
t
i
o
ni
sf
o
c
u
s
i
n
gm
o
r
ei
nt
h
ea
r
e
a
c
r
o
s
s
o
v
e
ra
n
dm
u
t
a
t
i
o
na
r
ew
a
y
st
on
a
r
r
o
wt
h
es
e
a
r
c
h
p
r
o
m
i
s
i
n
gas
o
l
u
t
i
o
n
.T
y
p
i
c
a
l
l
y
b
a
s
e
do
n
ec
h
a
r
a
c
t
e
r
i
s
t
i
c
so
ft
h
ec
u
r
r
e
n
tp
o
p
u
l
a
t
i
o
n
t
oa
na
r
e
ao
fd
e
s
i
g
ns
p
a
e
a
r
c
ho
p
e
nf
o
rt
h
eg
l
o
b
a
lmInImum
o
t
h
e
r
s
u
g
g
e
s
t
i
n
gal
o
c
a
lm
i
n
i
m
u
m
.Tok
e
e
pes
a
r
e
a
sm
u
s
tb
ep
r
e
s
e
n
t
e
df
o
re
x
p
l
o
r
a
t
i
o
n
.

l
i
sI
sa
c
h
I
c
v
e
db
yb
r
i
n
g
i
n
gi
na
nu
n
b
I
a
s
e
d
e
x
tg
e
n
e
r
a
t
i
o
na
l
o
n
gw
i
p
o
p
u
l
a
t
i
o
na
n
dh
a
v
i
n
gt
h
e
mc
o
m
p
e
t
ef
o
rp
r
o
m
o
t
i
o
nt
oen
t
h
e
c
n
ts
e
.
tI
nmanyv
a
r
i
a
t
i
o
n
st
h
e
s
ea
r
ea
l
l
o
w
e
dt
ob
r
e
e
dw
I
t
hec
u
r
r
e
n
tg
o
o
d
s
o
l
u
t
i
o
n
sb
e
f
o
r
ee
v
a
l
u
a
t
i
o
n
.
Thei
d
e
a
sp
r
e
s
e
n
t
c
dh
e
r
ea
r
eb
a
s
i
ct
oa
l
lo
fe
v
o
l
u
t
i
o
n
a
r
yc
o
m
p
u
t
a
t
i
o
n ds
p
e
c
i
f
i
c
.T
h
e
ya
r
ea
l
s
os
i
m
p
l
eb
u
tc
o
m
p
l
e
t
e
.F
o
re
a
c
hc
a
t
e
g
o
r
yl
i
s
t
e
da
b
o
v
et
h
e
r
ea
r
e
t
oGA
ta
r
eb
e
i
n
gr
e
s
e
a
r
c
h
e
da
n
dd
e
v
e
l
o
p
e
da
n
der
e
a
d
e
ri
s
m
o
r
es
o
p
h
i
s
t
i
c
a
t
e
dm
o
d
e
l
s a
e
n
c
o
u
r
a
g
e
dt
os
e
e
ki
n
f
o
r
m
a
t
i
o
nf
r
o
mt
h
ec
u
r
r
e
n
tl
i
t
e
r
a
t
u
r
e
.Thcs
p
e
c
i
f
i
cGAa
l
g
o
r
i
t
h
r
n
i
m
p
l
e
m
e
n
t
e
di
n b
o
o
ki
n
c
o
r
p
o
r
a
t
e
sa
l
lo
fee
l
e
m
e
n
t
sa
b
o
v
ei
nas
i
m
p
l
e
o
n
.
i
m
p
l
e
m
e
n
GenetlcAlgorlthm(
A
9
.
4
)Thea
l
g
o
r
i
t
h
ma
c
t
u
a
11
yi
m
p
l
e
m
e
n
t
e
di
sg
i
v
e
nb
e
l
o
w
.I
n
I
tI
sp
r
o
v
I
d
e
da
sawayt
o
l
i
n
ew
i
t
hmanyo
t
h
e
ra
l
g
o
r
i
t
h
m
sa
v
a
I
l
a
b
l
eI
nt
h
I
st
e
x
t
ti
sa
1
s
om
e
a
n
tt
op
r
o
v
i
d
eas
t
a
r
t
i
n
g
u
n
d
e
r
s
t
a
n
da
n
da
p
p
r
e
c
i
a
t
et
h
ew
o
r
k
i
n
go
ft
h
eGA.I
p
o
i
n
tf
o
rt
h
eu
s
e
rt
oi
m
p
l
e
m
e
n
th
i
sownv
a
r
i
a
t
i
o
no
ft
h
eGAs
i
n
c
et
h
e
s
et
e
c
h
n
i
q
u
e
sa
r
e
p
e
c
i
f
i
cp
r
o
b
l
e
mb
e
i
n
gs
d
i
e
d
.A
n
o
t
h
e
ri
m
p
o
r
t
a
n
tf
e
a
t
u
r
em
I
s
s
i
n
g
s
e
n
s
i
t
i
v
et
oes
omI
si
m
p
l
e
m
e
n
t
a
t
i
o
ni
sa
ne
n
c
o
d
i
n
gs
c
h
e
m
e
.Thed
e
s
i
g
nv
e
c
t
o
ri
sh
a
n
d
l
e
dd
i
r
e
c
t
l
y
t
op
r
o
v
I
d
ead
i
r
e
c
tf
c
e
d
b
a
c
k
.

9
.
2 NUMERICALTECHNIQUESANDADDITIONALEXAMPLES

3
7
1

SP1
.S
e
tu
pa
ni
n
i
t
i
a
lp
o
p
u
l
a
t
i
o
n
.Tens
e
t
so
fd
e
s
i
g
nv
e
c
t
o
r
sa
r
er
a
n
d
o
m
l
y
c
r
e
a
t
e
d
.Fromt
h
i
ss
e
t
a
nI
n
i
t
i
a
1g
e
n
e
r
a
t
i
o
no
ft
w
ov
e
c
t
o
r
si
ss
e
l
e
c
t
e
d
.T
h
i
si
s
p
o
o
la
v
a
i
l
a
b
l
ef
o
rp
r
o
p
a
g
a
t
i
o
n
.T
h
i
si
sa
l
s
ot
h
emInImumt
op
e
r
f
o
r
m
b
j
e
c
t
I
v
ef
u
n
c
t
I
o
n
.
c
r
o
s
s
o
v
e
r
s
.Thef
i
t
n
e
s
sf
u
n
c
t
i
o
nI
seo
G
e
n
e
r
a
t
i
o
nI
n
d
e
xt=
1
Maximumg
e
n
e
m
t
i
o
nI
n
d
e
x=t
m
a
x
S
t
e
p2
.G
e
n
e
t
i
cO
p
e
r
a
t
i
o
n
s
TwoS
i
m
p
l
ec
r
o
s
s
o
v
e
r
s
TwoA
r
it
h
m
e
t
i
cc
r
o
s
s
o
v
e
r
s(
c
o
n
v
e
x
)
OneD
i
r
e
c
t
i
o
n
a
lc
r
o
s
s
o
v
e
r
M
u
t
a
t
i
o
no
fa
l
ls
e
v
e
nd
e
s
i
g
nv
e
c
t
o
r
ss
of
a
r
F
o
u
r
i
m
r
g
S
e
l
e
c
t
i
o
no
ft
h
eb
e
s
tt
w
od
e
s
i
g
nv
e
c
t
o
r
s(
f
r
o
m1
8v
e
c
t
o
r
s
)f
o
rp
r
o
m
o
t
i
o
n
t
oen
e
x
tg
e
n
e
r
a
t
i
o
n
t
o
p
I
f1=1m S
I
n
c
r
e
m
e
n
tg
e
n
e
r
a
t
i
o
ni
n
d
e
x
:It+1
Got
oS
t
e
p
2
Example9
.
3 Th
ee
x
a
m
p
l
ef
o
ri
ss
e
c
t
i
o
ni
saB
e
z
i
e
rc
u
r
v
ef
i
t
t
i
n
gp
r
o
b
l
e
m
.G
i
v
e
nas
e
t
a
t
a
f
i
n
dt
h
ef
i
f
t
h
o
r
d
e
rB
e
z
i
e
rc
u
r
v
eatw
e
da
.l
eo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
ni
s
o
f d
l
e
t
s
q
u
a
r
e
r
r
o
r
b
e
t
w
n
e
o
r
i
g
i
n
a
ld
a
t
aa
n
dt
h
ed
a
t
ag
e
n
e
r
a
t
e
db
yt
h
eB
e
z
i
e
r
c
u
r
v
e
.
s
p
r
o
b
l
e
mh
a
sb
e
e
na
d
c
l
r
e
s
s
e
de
a
r
l
i
e
rr
o
u
g
hac
o
n
t
i
n
u
o
u
sd
e
t
e
r
r
n
i
n
i
s
t
i
ct
e
c
h
n
i
q
u
ei
n
a
ne
a
r
l
i
e
rc
h
a
p
t
e
r
. e
r
e
a
r
ee
h
t
d
e
s
i
v
a
r
i
a
b
l
e
sa
s
s
i
a
t
e
dw
i
t
ht
h
ep
r
o
b
l
e
m
.l
e
r
e
a
r
e
s
e
v
e
r
a
lr
e
a
s
o
n
sf
o
rt
h
ec
h
o
i
c
eo
ft
h
ep
r
o
b
l
e
m
.F
i

t
h
eg
l
o
b
a
ls
o
l
u
t
i
o
n

i
fi
te
x
i
s
t
s

m
u
s
t
b
ec
l
o
s
ezero
d c
o
n
d
i
sp
r
o
v
I
d
e
saway
d
e
s
c
r
i
b
ee
i
g
h
td
e
s
i
g
nv
a
r
i
a
b
l
e
si
n
n
s
i
o
n
a
ls
p
a
c
e(
p
l
o
t
)
. l
eb
e
s
tv
a
l
u
ef
o
rt
h
ed
e
s
i
g
nv
e
c
t
o
r e
v
e
r
yg
e
n
e
r
o
n
i
s
t
w
o
d
i
m
p
l
o
t
t
e
dt
os
e
ehowt
h
ec
u
r
v
ef
i
ti
m
p
r
o
v
e
s
.S
i
n
c
ee
n
c
o
d
i
n
gI
sn
o
tb
e
i
n
gu
s
e
di
nt
h
e
ar
e
a
s
o
n
a
b
l
yl
o
n
gd
e
s
I
g
nv
e
c
t
o
ri
sn
e
c
e
s
s
a
r
y

o
b
s
e
r
v
et
h
c
r
o
s
s
o
v
e
r
s
.l
e
a
l
g
o
r
i
t
h
m
l
I
o
w
i
n
gf
u
n
c
t
i
o
n
:
o
r
i
g
i
n
a
ld
a
t
aa
r
eo
b
t
a
i
n
e
df
r
o
mt
h
ef
o
xc
J
t
x
)=1+O
.
2
5
x+2
e
o
s3
x

(
9
.
9
)

F
i
g
u
r
e9.
12i
l
l
u
s
t
r
a
t
e
sesePo
ft
h
ee
x
a
m
p
l
ei
n
c
l
u
d
i
n
gt
h
ee
i
g
h
td
e
s
i
g
nv
a
r
i
a
b
l
e
s
w
h
i
c
ha
r
et
h
exa
n
dycompo
o
ft
h
ei
n
t
e
m
a
lv
e
r
t
i
c
e
s td
e
t
e
r
m
i
n
eef
i
t
t
e
d
c
u
r
v
ec
o
m
p
l
e
t
e
l
y
.
'
b
ef
u
n
c
t
i
o
ni
sh
a
r
dc
o
d
e
di
nCa
l
I
GA.m.T
h
e
r
ea
r
es
e
v
e
r
a
lr
e
a
s
o
n
s
whyi
sI
sn
o
teb
e
s
te
x
a
m
p
l
ef
o
ri
l
l
u
s
t
r
a
t
i
o
no
fGA
.F
I
r
s
t
i
tI
s e
x
a
m
p
l
ei
n
c
o
n
t
i
n
u
o
u
sv
a
r
i
a
b
l
e
s
.S
e
c
o
n
d

l
o
c
a
lminimumsa
r
en
o
tv
e
r
yd
i
s
t
i
n
g
u
i
s
h
a
b
l
ea
n
dt
h
e
r
e
a
r
ei
n
n
u
m
e
r
a
b
l
em yo
ft
h
e
m
.Theg
l
o
b
a
lminimumi
sn
o
ts
u
b
s
t
a
n
t
i
a
l
l
yd
i
f
f
e
r
e
n
t
f
r
o
mt
h
el
o
c
a
lminimumv
a
l
u
et
ow r
a
n
te
x
p
l
o
m
t
i
o
no
ft
h
eg
l
o
b
a
1s
o
l
u
t
i
o
n
.T
h
i
r
d

t
h
e
p
r
o
b
l
e
mI
sn
o
ti
l
lb
e
h
a
v
e
d
.Them
o
s
ti
m
p
o
r
t
a
n
tr
e
a
s
o
nd
e
p
e
n
d
sn
o
to
nt
h
ea
l
g
o
r
i
t
h
m
o
m
p
u
t
a
t
I
o
no
ft
h
eo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
n
c
a
1
c
u
l
a
t
I
n
gt
h
el
e
a
s
ts
q
u
a
r
e
de
r
r
o
r
.
b
u
t
o
nec

9
.
2N
U
M
E
I
C
A
LT
E
C
H
N
I
Q
U
E
SA
N
DA
D
O
I
T
I
O
N
A
LE
X
A
M
P
L
E
S 3
7
3

L
O
B
A
LO
P
T
I
M
I
Z
A
T
lO
N
3
7
2 G

F
i
g
u
r
e
s9
.
1
3t
o9
.
1
7c
a
p
t
u
r
ct
h
ca
p
p
1
i
c
a
t
i
o
nofGAt
oE
x
a
m
p
l
e9.
3t
h
r
o
u
g
h2
0
0
e
a
d
e
rs
h
o
u
l
db
ea
b
l
et
or
u
nt
h
ec
x
a
m
p
l
ef
r
o
mt
h
ec
o
d
ed
i
r
c
c
t
o
r
y
g
e
n
e
r
a
t
i
o
n
s
.Thcr
o
fC
h
a
p
t
e
r9
.B
e
c
a
u
s
eo
ft
h
eh
e
u
r
i
s
t
i
cn
a
t
u
r
eo
ft
h
eg
c
n
c
r
a
t
i
o
no
fs
o
l
u
t
i
o
n
st
h
e
s
o
l
u
t
i
o
nt
r
a
i
lw
i
l
lb
ed
i
f
f
e
r
e
n
tf
r
o
ma
tshownb
e
l
o
w
.R
u
n
n
i
n
gt
h
ee
x
a
m
p
l
c
s
e
v
e
r
a
lt
i
m
e
sb
yi
n
v
o
k
i
n
gCallGAf
r
o
mt
h
eCommandwindows
h
o
u
l
di
l
l
u
s
t
r
a
t
e
t
h
e GA a
l
w
a
y
sm
a
n
a
g
e
st
oc
l
o
s
c
l
ya
p
p
r
o
x
i
m
a
t
et
h
cc
u
r
v
ea
f
t
e
r2
0
0
a
1u
ei
sn
o
tp
r
e
d
i
c
t
a
b
l
ea
n
dc
a
n
n
o
tb
c
g
e
n
e
r
a
t
i
o
n
s
.Themaximumg
c
n
e
r
a
t
i
o
nv
e
s
t
a
b
1
i
s
h
e
dw
i
t
hc
e
r
t
a
i
n
t
y
.I
tr
e
l
i
e
so
nn
u
m
e
r
i
c
a
1e
x
p
e
r
i
m
e
n
t
a
t
i
o
n
.Thi
sn
u
m
b
e
r
a
p
p
e
a
r
sc
a
s
i
l
yt
ob
eo
v
e
r1
0
0
0i
nmanyi
n
v
e
s
t
i
g
a
t
i
o
n
s
.Ther
e
a
d
e
ri
su
r
g
e
dt
o
lGA.mt
h
eb
c
s
t
e
x
p
c
r
i
m
c
n
tw
i
t
ht
h
ee
x
a
m
p
l
et
oc
x
p
l
o
r
et
h
i
si
s
s
u
e
.I
nr
u
n
n
i
n
gCaJ
v
a
l
u
e
so
ft
h
ed
e
s
i
g
nv
e
c
t
o
ra
n
dt
h
cc
o
r
r
e
s
p
o
n
d
i
n
go
b
j
c
c
t
i
v
ef
u
n
c
t
i
o
nv
a
l
u
ca
r
e
c
a
p
t
u
r
c
di
nt
h
ev
a
r
i
a
b
l
eXbest-ana
r
r
a
yo
f2
0
1r
o
w
sa
n
d9c
o
l
u
m
n
s(
a
s
s
u
m
i
n
g
2
0
0g
e
n
e
r
a
t
i
o
n
s
)
.Thef
i
r
s
t8e
o
l
u
m
n
sc
a
r
r
yt
h
eb
e
s
td
e
s
i
g
nf
o
rt
h
eg
e
n
e
r
a
t
i
o
na
n
d
t
h
el
a
s
tc
o
l
u
m
ni
st
h
ec
o
r
r
e
s
p
o
n
d
i
n
go
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
nv
a
l
u
e
.
F
i
g
e9
.
1
3i
seb
e
s
td
e
s
i
g
nf
r
o
mt
h
es
t
a
r
t
i
n
gg
e
n
e
r
a
t
i
o
n
.I
th
a
s o
b
j
e
c
t
i
v
e
a
1u
co
f3
2
.
13
.l
i
si
secors
q
u
a
r
e
do
v
e
r1
0
0d
a
t
ap
o
i
n
t
so
ft
h
eo
r
i
g
i
n
a
1
f
u
n
c
t
i
o
nv
d
a
t
aa
n
dec
u
r
v
eg
e
n
e
r
a
t
e
db
ycd
e
s
i
g
nv
a
r
i
a
b
l
e
s
.Thcd
e
s
i
g
nv
a
r
i
a
b
l
e
sa
r
ce
v
e
r
t
i
c
i
I
Iu
s
t
r
a
t
c
di
nt
h
ef
i
g
u
r
c
.Theb
a
s
i
cc
u
r
v
eo
fF
i
g
u 9
.t
3i
so
b
t
a
i
n
e
du
s
i
n
g

B
e
z
i
e
rc
u
r
v
e
s

O
r
f
g
l
n
D
a
l
a
--R
l
l
e
d
d
a
l
a
oV
e
r
t
l
c
e
s
2
.
5

(
X
s
'
X
)
o

.
.
.

g
~ 1
.5

/ 0
"

/.}

12

1
.
4

x
z

0
.
5

02

0
.
4

g
u
r
e9
.
1
2

0
.
6

0
.
8

x
d
a
t
a

1
.
6

1
.
8

G
e
n
e
t
l
cA
l
g
o
r
i
t
h
m
.d
e
s
c
r
i
p
t
i
o
no
fExa
m
p
l
e9
.
3
.

I
tu
s
e
saN
e
w
t
o
n
R
a
p
h
s
o
nt
e
c
h
n
i
q
u
ea
ti
sp
r
o
n
et
of
a
i
1
u
r
ci
ft
h
cc
u
r
v
ef
o
l
d
so
v
e
r
s
e
v
e
r
e
l
y
o
rcv
e
r
t
i
c
e
sa
r
ev
e
r
yc
10
s
et
o
g
e
e
r
.Onewayt
oo
v
e
r
c
o
m
et
h
i
si
st
oc
a
u
s
e
r
t
i
c
e
st
om
o
n
o
t
o
n
i
c
a
l
l
yi
n
c
r
e
a
s
e(
w
h
i
c
hi
si
m
p
l
e
m
c
n
t
c
di
ne
t
h
ex
I
o
c
a
t
i
o
no
ft
h
ev
p
o
p
u
l
a
t
i
o
ng
c
n
c
r
a
t
i
o
nf
u
n
c
t
i
o
n
)
.T
h
i
su
p
s
e
t
smanyo
ft
h
ec
r
o
s
s
o
v
e
ro
p
e
r
a
t
i
o
n
sS
I
O
C
C
a
renowco
a
l
l
yr
a
n
d
o
mi
n
d
ofp
l
yr
a
n
d
o
m
.
l
ef
o
l
l
o
w
i
n
gf
i
l
e
sa
r
en
c
c
e
s
s
a
r
yf
o
r s
s
p
i
f
i
ce
x
n
p
l
er
e
l
a
t
i
n
gt
ot
h
eh
a
n
d
l
i
n
g
o
fB
e
z
i
c
rc
u
r
v
e
s
:c .m
F
a
c
t
o
r
l
a
l
.
m
Comblnatlon.m
Bez_SCLError.ma
n
d
a
r
ea
v
a
i
l
a
b
l
ei
nt
h
ec
o
d
ed
i
r
c
c
t
o
r
yf
o
rt
h
ec
h
a
p
t
e
r
.
Th
eGAi
nA
I
g
o
r
i
t
h
r
n(
A
9
.
4
)i
si
m
p
l
e
m
e
n
t
e
du
s
i
n
gt
h
ef
o
l
l
o
w
i
n
gmf
l
Ie
s
.
S
e
t
t
i
n
gu
pp
a
r
a
m
e
t
e
r
sa
n
dd
a
t
af
o
rt
h
ea
1
g
o
r
i
t
h
m
Cal
IGA.m
D
i

c
t
s
t
h
e
a
l
g
o
r
i
t
h
m
GenetlcAlgorlthm.m
G
e
n
c
r
a
t
e
sr
a
n
d
o
md
e
s
i
g
nv
c
c
t
o
r
p
o
p
u
l
a
t
o
r
.
m
c
r
f
o
r
m
ss
i
m
p
l
ec
r
o
s
s
o
v
e
rb
e
t
w
e
e
nt
w
op
a
r
e
n
t
s
SlmpleCrossover.m P
v
e
xc
r
o
s
s
o
v
c
rb
e
t
w
n
t
w
op
a
r
e
n
t
s
e
r
f
o
r
m
sn
m
e
t
l
c
C
r
o
s
:
I
v
e
r
.
mP
A
r
i
e
r
f
o
r
m
so
n
ed
i
r
c
c
t
i
o
n
a
lc
r
o
s
s
o
v
e
r
D
i c
t
i
o
n
a
l
C
r
o
s
s
o
v
e
r
.
mP
P
e
r
f
o
r
m
sm
u
t
a
t
i
o
no
fed
e
s
i
g
nv
e
c
t
o
r
s
M
u
t
a
t
l
o
n
.
m
D
r
a
w
s
t
h
c
c
u
r
v
e
f
o
r
e
a
c
h
d
e
s
i
g
n
v
c
c
t
o
r
DrawCurve.m

2
.
5

.
.
1
.
5

>0

0
.
5

0
.
5

0
.
5

d
a
t
a

1.
5

n
e
r
a
t
l
o
nt 1
.
e
s
tf
l
t
a
t
g
F
l
g
u
r
e
9
.
1
3B
"

2
.
5

374

9
.
2 NUMERICALTECHNIQUESANDADDITIONALEXAMPLES

GLOBALOPTIMIZAnON

375

35

30

2
s

"

"....

.
.
"""' .

.
.
.
.

.
.
.
.

.
4

....'+

1
0

-0
.
5

a4u P 3

g0
.
5

Z
O
L

0
.
4

0
.
6

O
.

x d
a
t
a

1
.
2

1
.
4

1
.
6

1
.
8

xd
a
t
a
g
u
r
e9
.
1
5

f
:
tg
e
n
e
r
a
t
i
o
nt
=200.
F
l
g
u
r
e9
.
1
4B
e
s
ti

D wCurve.mf
u
n
c
t
i
o
n
.Todrawt
h
ei
n
i
t
ia
1c
u
r
v
e(
ef
i
g
u
r
ei
se
d
i
t
e
ds
u
b
s
e
q
u
e
n
t
l
y
w
i
t
hp
l
o
t
e
d
i
t
)i
nt
h
eCommandwindow:
>> DrawCurve(Xbest(1
1:8))

i
na
1g
e
n
e
r
a
t
i
o
n(
g
e
n
e
r
a
t
i
o
nnumber 2
).
F
i
g
u
r
e9
.
1
4i
st
h
eb
e
s
td
e
s
i
g
na
tef
l
eo
b
j
c
c
t
i
v
ef
u
n
c
t
i
o
ni
s0
.
0
7
2
9
.1
ti
sc
l
e
a
rfromt
h
ef
i
g
u
r
e GAi
sa
t
t
e
m
p
t
i
n
g
t
of
i
n
dt
h
es
o
l
u
t
i
o
n
.Th
i
sc
a
nb
ee
s
t
a
b
l
i
s
h
e
dbyI
kinga
tF
i
g
u
r
e9.
15whichi
se
i
t
e
r
a
t
i
o
n
l
g
e
n
e
r
a
t
i
o
nh
i
s
t
o
r
yo
ft
h
eGAf
o
rt
h
i
se
x
a
m
p
l
e
.Th
eo
b
j
c
c
t
i
v
ef
u
n
c
t
i
o
ni
s
r
a
11
yd
c
c
r
e
a
s
i
n
gt
h
r
o
u
g
heg
e
n
e
r
a
t
i
o
n
s
.F
i
g
u
r
9
.
1
6i
sas
c
a
1
edr
e
p
r
e
s
e
n
t
a
t
i
o
nt
o
g
e
n
e
d
i
s
p
l
a
yt
h
eb
e
h
a
v
i
o
ro
ft
h
eGAa
tl
a
t
r
i
t
e
r
a
t
i
o
n
s
.Fromed
a
t
ai
ti
sa
l
8
0a
p
p
a
r
e
n
tt
h
n
t
i
scxamplcn
e
e
d
st
or
u
nf
o
rmorc 200g
e
n
e
r
a
t
i
o
n
s
.Thef
i
na
1f
i
g
u
r
e

F
igure9.17
r
e
p
r
e
s
e
n
t
s movcmentandt
h
ec
l
u
s
t
e
r
i
n
go
ft
h
ev
e
r
t
i
c
e
st
h
r
o
u
g
ht
h
eg
e
n
e
r
a
t
i
o
n
s
.I
t
i
ss
u
g
g
e
s
t
i
v
co
fl
o
ca
1r
e
f
i
n
i
n
g i
sa
s
s
o
c
i
a
t
e
dw
i
t
ht
h
eGA.

SuggestedExt
ensions:Most of t
h
ee
n
g
i
n
r
i
n
gd
e
s
i
g
np
r
o
b
l
e
m
si
n
v
o
l
v
c

i
n
t
s
.C
u
r
r
c
n
t
l
yt
h
emostp
o
p
u
l
a
rwayt
oh
a
n
d
l
ec
o
n
s
t
r
a
i
n
t
si
st
or
e
f
o
r
m
u
l
a
t
et
h
c
c
o
n
s
p
r
o
b
l
e
mu
s
i
n
gp
c
na
1t
yf
u
n
c
t
i
o
n
s(
C
h
a
p
t
e
r7
)
.
'
h
e
r
ea
r
es
c
v
c
r
a
lo
t
h
ra
l
t
e
m
a
t
i
v
e
st
o

5
0

2
0
0

250

O
b
j
e
c
l
i
v
ef
u
n
c
t
i
o
nf
o
re
v
e
r
yg
e
n
e
r
a
t
i
o
n
.

i
sp
r
'
u
r
e
.1
ti
sp
o
s
s
i
b
l
et
os
c
p
a
r
a
t
e
l
ya
c
c
u
m
u
l
a
t
eo
n
l
yd
e
s
i
g
n
s a
r
ef
e
a
s
i
b
l
e

w
h
i
l
ec
o
n
t
i
n
u
i
n
gt
oi
m
p
l
e
m
e
n
tt
h
es
t
a
n
d
a
r
dp
r
o
c
e
d
u
r
e
.Sccond

u
s
eas
t
r
a
t
e
g
yt
oi
v
e
t
h
es
e
a
r
c
hp
r
o
c
e
s
st
h
r
o
u
g
hc
o
n
s
t
r
a
i
n
ts
a
t
i
s
f
a
c
t
i
o
nandd
e
c
r
e
a
s
i
n
geo
b
j
c
c
t
i
v
c
s
i
m
u
l
t
a
n
e
o
u
s
l
yo
rs
e
p
a
r
a
t
e
l
y
.Thca
l
g
o
r
i
t
h
m
si
n
d
i
c
a
t
c
di
nt
h
i
sc
h
a
p
t
e
rc
a
ne
a
s
i
l
y
g
e
s
.
i
n
c
o
r
p
o
r
a
t
c n dedch
Thea
1g
o
r
i
t
h
m
i
cp
r
o
c
e
d
u
r
ea
simplementeddo
n
o
tr
e
a
l
1yc
a
r
ei
f eproblem
i
sc
o
n
t
i
n
u
o
u
so
rd
i
s
c
r
e
t
e
.1
tj
u
s
tworksonac
u
r
r
e
n
ts
e
tofv
a
l
u
e
sf
o
rt
h
ed
e
s
i
g
n
v
a
r
i
a
b
l
e
s
.D
is
c
r
e
t
eva
1u
e
swi
l
1p
r
i
m
a
r
i
l
yr
e
q
u
i
r
eminorc
h
a
n
g
e
si
nt
h
cwayt
h
e
p
o
p
u
l
a
t
i
o
ni
sg
e
n
e
r
a
t
e
d
.A1
s
ot
h
ci
m
p
l
c
m
c
n
t
a
t
i
o
no
ft
h
ec
r
o
s
s
o
v
e
randm
u
t
a
t
i
o
n
o
p
e
r
a
t
o
r
swi
l
1needt
obec
h
a
n
g
e
d
.Thesechangesi
nnowayd
e
t
r
a
c
tfromt
h
emo
r
i
d
e
ab
c
h
i
n
dt
h
eGA.

CloslngComments:I
n summaryf
o
rah
e
u
r
i
s
t
i
ct
e
c
h
n
i
q
u
ceGA i
sv
e
r
y
imp
s
s
i
v
e
.l
Csamcc
a
nb
es
a
i
df
o
r SA
.I
nc
l
u
d
i
n
gt
h
ef
a
c
ta
tp
r
a
c
t
i
c
a
l
e
n
g
i
n
e
e
r
i
n
gp
r
o
b
l
e
m
smustr
e
q
u
i
r
ep
a
r
t
l
yd
i
s

e
t
er
e
p
s
e
n
t
a
t
i
o
net
c
c
h
n
i
q
u
e
si
n
C
h
a
p
t
e
8 d9a
r
ev
e
r
yr
e
l
e
v
a
n
tt
oep
r
a
c
t
i
o
fd
e
s
i
g
nanddevelopmen
.
tWith
t
h
eenormousc
o
m
p
u
t
a
t
i
o
na
1r
e
s
o
u
ei
n ed
e
s
k
t
o
p
sand l
a
p
t
o
p
sa
ne
f
f
e
c
t
i
v
c
programmingl
a
n
g
u
a
g
eI
i
keMATLABandawidev
a
r
i
c
t
yo
fa
l
g
o
r
i
t
h
m
sa
ta
r
c
i
ti
sp
o
s
s
i
b
l
cnowt
od
e
s
i
g
na
ni
n
d
i
v
i
d
u
a
la
p
p
r
o
a
c
ht
omanyo
p
t
i
m
i
z
a
t
i
o
n
a
v
a
i
l
a
b
l
e

376 G
L
O
B
A
LO
P
T
I
M
I
Z
A
T
I
O
N

R
E
F
E
R
E
N
C
E
S 377
L
o
c
a
t
l
o
no
fd
e
s
l
g
nv
a
r
l
a
b
l
e
so
v
e
r2
0
0g
e
n
e
r
a
t
i
o
n
s

0
.
5
0
.
4
5
s
c
a
l
e
df
o
rl
o
w
e
s
tv
a
l
u
e
s

0
.
4
0
.
3
5

e2Z

t-t
n
u

-0
.
5

0
.
1
5
1
0
.
1
1
.
5

0
.
0
5

2
0

4
0

6
0

8
0

1
0
0 1
2
0
xd
a
t
a

1
4
0

1
6
0

1
8
0

2
0
0

F
l
g
u
r
e9
.
1
6 Ob
t
i
v
ef
u
n
c
t
l
o
ns
c
a
l
e
d1
0s
h
o
wc
o
n
t
l
n
u
a
ld
e
c
r
e
a
s
el
o
rl
a
l
e
rg
e
n
e
r
a
t
l
o
n
.

2
0
.
4

0
.
6

0
.
8

g
u
r
e9
.
1
7

1
.
2
1
.
4
1
.
6
xv
a
l
u
eo
fv
e
r
t
e
x

1
.
8

2
.
2

2
.
4

T
r
a
c
e0
1I
h
ed
e
s
l
g
nv
a
r
l
a
b
l
e
sI
h
r
o
u
g
hg
e
n
e
r
a
t
f
o
n
.

REFERENCES
p
r
o
b
l
e
m
s
.Th
i
si
si
m
p
o
r
t
a
n
tt
op
r
a
c
t
i
c
i
n
ge
n
g
i
n
e
e
r
swhoc
a
nb
r
i
n
gi
n
l
u
i
t
i
o
na
n
di
n
s
i
g
h
t
yh
a
v
et
os
o
l
v
ei
nat
i
g
h
tt
i
m
ef
r
a
m
e
.I
nt
h
ee
n
g
i
n
e
e
r
i
n
g
f
o
rt
h
ep
a
r
t
i
c
u
l
a
rp
r
o
b
l
e
me
a
d
i
t
i
o
n
a
ld
e
s
i
g
no
p
t
i
m
i
z
a
l
i
o
nc
o
u
r
s
e
sl
e
n
dt
od
e
a
lo
n
l
yw
i
c
u
r
r
i
c
u
l
u
r
d
e
t
e
s
t
i
cc
o
n
t
i
n
u
o
u
sa
l
g
o
r
i
t
h
m
s
.Th
a
ti
ni
t
s
e
l
fwasmo
t
h
et
i
m
ea
v
a
i
l
a
b
l
e
t
eo
p
t
i
m
i
z
a
t
i
o
no
rg
l
o
b
a
lo
p
t
i
m
i
z
a
t
i
o
na
r
en
e
e
e
s
s
a
r
i
l
ys
k
i
l
l
s
f
o
ri
n
s
t
r
u
c
t
i
o
n
.D
i
s
c
m
edbyp
e
r
s
o
n
a
le
n
d
e
a
v
o
r
. l
Co
v
e
r
c
m
p
h
a
s
i
sond
e
t
e
r
m
i
n
i
s
t
i
cp
r
o
g
r
a
m
m
i
n
gn
e
e
d
s
l
e
v
i
s
i
t
e
di
nt
h
el
i
g
h
to
fc
u
r
r
e
n
td
e
v
e
l
o
p
i
nea
r
c
ao
fo
p
t
i
m
i
z
a
t
i
o
n
.T a
y
t
ober
r
i
n
gd
e
p
a
r
e
n
t
sa
n
di
t
sr
e
i
n
f
o
r
c
c
m
e
n
ti
su
s
u
a
l
l
y
c
o
m
p
u
t
e
rp
r
o
g
r
a
m
m
i
n
gi
ne
n
g
i
n
u
r
r
i
c
u
l
u
m
.Newp
r
o
g
r
a
m
m
i
n
gp
l
a
t
f
o
r
m
sI
i
k
e
a
d
j
u
n
c
l1
0o
t
h
e
rd
e
v
e
l
o
p
m
e
n
t
si
nec
sd
i
s
a
d
v
a
n
l
a
g
e
.Newa
l
g
o
r
i
t
h
m
sa
l
s
or
e
f
l
e
c
la
MATLABo
f
f
e
rawayt
or
e
s
o
l
v
ci
i
l
y s
l
a
t
e
da
n
da
p
p
l
i
e
d
.
s
i
m
p
l
i
c
i
t
yt
h
a
tc
a
nb
e s
i
sbookh
a
sb
e
e
na
b
o
u
tp
r
o
g
r
a
m
m
i
n
gi
nMATLAB
l
e
a
m
i
n
ga
b
o
u
tt
r
a
d
i
t
i
o
n
a
la
n
d
nd
e
s
i
g
no
p
t
i
m
i
z
a
t
i
o
n
and t
r
a
n
s
l
o
t
i
n
gt
h
e
s
eo
p
t
i
m
i
z
a
t
i
o
n
c
u
r
r
e
n
tt
e
c
h
n
i
q
u
e
si
i
q
u
e
si
n
t
oa
p
p
l
i
.
t
i
o
n
s
.l
ti
sa
l
s
oa
b
o
u
te
x
p
l
o
r
a
t
i
o
na
n
dd
e
v
e
l
o
p
i
n
ga
ni
n
s
i
g
h
t o
o
d
e i
sp
r
o
v
i
d
e
di
so
n
l
ym
e
a
n
tt
ob
ea
t
h
ep
r
o
c
e
s
so
fd
e
s
i
g
no
p
t
i
m
i
z
a
t
i
o
n
.Thec
s
t
a
r
t
i
n
gp
o
i
n
.
t

1
.Aror
a

J
.S
.
l
n
t
r
o
d
u
c
t
i
o
n10O
p
t
i
m
a
lD
e
s
i
g
n
McGraw-H
I
i NewY
o
r
k
1
9
8
9
.
2.K
'
k
p
a
c
k

S
.
G
e
)
a
t
t

G
.C
.
l
I
n
dV
e
c
c
h
i
M.P
.
Opt
i
m
i
z
a
t
i
o
nb
yS
i
m
u
l
a
t
e
dAnnea1i
n
g
S
c
i
e
n
c
e

Vo
l
.2
2
0

N
o
.4
5
9
8
May1
9
8
3
.
p
o
l
i
s
N
.R
o
s
e
n
b
l
u
t
h
A
.R
o
s
e
n
b
l
u
t
hM
.
T
e
l
l
eA
.
a
n
dT
e
l
l
e
r

E
.
Equ
l
l
t
i
o
n
so
f
3
.Me
1
c
u
l
a
t
i
o
n
sb
yF
a
s
tC
o
m
p
u
t
i
n
gM
a
c
h
i
n
e
s
JoumalofC
h
e
m
i
c
a
lP
h
y
s
i
c
s

Vo
.
l2
1
S
a
t
eCa
p
p
.1
0
8
7
1
0
9
21
9
5
3
.

1
.0.
J
o
h
n
s
o
n
M.E
"a
n
dM
y
r
o
n
L
.S
.G
e
n
e
r
a
l
i
z
e
dS
u
l
a
dAn
n
e
a
1
i
n
g
4
.B
o
c
h
a
v
e
s
k
y
T
e
c
l
m
o
m
e
t
r
i
c
s

Vo
.
l2
8
N
o
.3
1
9
8
6
.
f
o
rF
u
n
c
t
i
o
nO
p
t
i
m
i
z
a
t
i
o
n
A
.E
.W.
a
n
dF
o
r
b
e
s
G
.W.AnA
d
a
p
t
i
v
eS
i
m
u
l
a
t
e
dAnn
e
a
1
i
n
gAlg
o
r
i
t
h
mf
o
r
5
.J
o
n
e
s
Jou
:
r
1
o
f
G
l
o
b
a
lO
p
l
i
m
i
l
.
C
lt
i
o
n
Vo
l
.6

G
l
o
b
a
lO
p
t
i
m
i
z
a
t
i
o
no
v
e
rC
o
n
t
i
n
u
o
u
sV
a
r
i
a
b
l
e
s
p
p
.1
3
7
1
9
9
5
.
H
.
a
n
dF
i
e
l
d
i
n
g
M.
S
i
m
u
l
a
t
e
dA
n
n
e
a
l
i
n
g
:S
e
a
r
c
h
i
n
gf
o
ra
nO
p
t
i
m
a
1T
empe
u
r
e
6
.C
o
h
n

SIAMJoumalonO
p
t
i
m
i
z
a
t
i
o
n
VoL9

N
o
.3

p
p
.7
7
9
7
8
2
1
9
9
9
.
S
c
h
e
d
u
l
e
J
.H
.
O
u
t
l
i
n
eo
faL
o
g
i
c
a
lTh
e
o
r
yo
fA
d
a
p
t
i
v
eS
y
s
t
e
m
s
Joumalof
ACM

V
o
l
.
7
.H
o
l
l
a
n
d
3

p
p
.
2
9
7
3
1
41
9
6
2
.


"
^
F

378 G
L
O
B
A
LO
P
T
I
M
I
Z
A
T
I
O
N
8
.D
u
m
i
t
r
e
s
c
uD
.Lazz
e
r
i
n
iB
.J
a
i
n
L
.C
.a
n
dD
u
m
i
t
r
c
s
c
uA
.E
v
o
l
u
t
i
o
n
a
r
y
I
n
t
e
m
a
t
i
o
n
a
1S
e
r
i
e
so
nC
o
m
p
u
t
a
t
i
o
na
1I
n
t
e
l
l
i
g
e
n
c
e
L
.C
.J
a
i
n
(
e
dCRC
C
o
m
p
u
t
a
t
i
o
n
BocaR
a
t
o
n
FL

2
0
0
0
.
P
r
e
s
s
9
.Gen

M.
a
n
dCheng
R
.
G
e
n
e
t
i
cA
l
g
o
r
i
t
h
m
sandE
n
g
i
n
e
e
r
i
n
gO
p
t
i
m
i
z
a
t
i
o
n
W
i
l
e
yS
e
r
i
e
s
H
.R
.
P
a
r
s
a
c
i(
s
c
r
i
c
sc
d
.
)
W
i
l
c
y
-I
nt
c
r
s
c
i
c
n
c
e

i
nE
n
g
i
n
e
e
r
i
n
gD
e
s
i
g
na
n
dA
u
t
o
m
a
t
i
o
n
NewYork

2
0
0
0
.
1
0
.G
o
l
d
b
c

D
.
G
e
n
e
t
i
cA
l
g
o
r
i
t
h
m
sUI S
e
a
r
c

hO
p
t
i
m
i
z
a
r
i
o

na
n
dM
a
c
h
i
n
eLe
am
U
l
g
A
d
d
i
s
o
n
W
e
s
l
e
y
R d
i
n
gMA
1
9
8
9
.
11
.Houck
C
.R
.J
o
i
n
e
sJ
.A
.
a
n
dKa
y
M.G
.A G
e
n
e
t
i
cA
l
g
o
r
i
t
h
mf
o
rF
u
n
c
t
i
o
n
e
ra
c
c
o
m
p
a
n
y
i
n
gt
h
e GAOT (
G
e
n
e
t
i
cA
l
g
o
r
i
t
h
mO
p
t
i
m
i
z
a
t
i
o
n
O
p
t
i
m
i
z
a
t
i
o
np
b
o
x
)
No
C
a
r
o
l
i
n
aS
t
a
t
eU
n
i
v
e
i
t
y
.
TI

PROBLEMS
er
e
a
d
e
ri
se
n
c
o
u
r
a
g
e
dt
od
e
v
e
l
o
ph
i
sown
Nof
o
n
n
a
lp
r
o
b
l
e
m
sa
r
es
u
g
g
e
s
t
e
d
. l
l
g
o
r
i
t
h
m
si
n
c
l
u
d
i
n
gf
o
n
n
a
lo
n
e
sa
v
a
i
l
a
b
l
ei
nes
u
g
g
e
s

v
a
r
i
a
n
t
so
f twoa
l
i
t
e
r
a
t
u
r
e
.Hei
sa
l
s
oe
n
c
o
u
r
a
g
e
dt
oc
o
n
s
i
d
e
re
x
t
e
n
s
i
o
n
st
oc
o
n
s
t
r
a
i
n
e
dandd
i
s
c
r
e
t
e
v
a
r
i
o
u
sexamplesf
o
u
n
d
p
r
o
b
l
e
m
s
.Hei
srecommendedt
ou
s
eh
i
sv
e
r
s
i
o
n
st
oe
x
p
l
o
e
t
e
n
n
i
n
i
s
t
i
ct
e
c
h
n
i
q
u
e
s(
u
n
t
i
lC
h
a
p
t
e
r8
)
i
nt
h
el
i
t
e
r
a
t
u
r
e
.I
nc
o
m
p
a
r
i
s
o
nw
i
t
hed
t
h
e
s
h
e
u
r
i
s
t
i
cmethodsa
r
et
r
u
l
yf
a
s
c
i
n
a
t
i
n
gandi
l
1u
s
t
r
a
t
ec
o
n
s
i
d
e
r
a
b
l
er
o
b
u
s
t
n
e
s
s
andwhati
scvenmorec
h
a
l
l
e
n
g
i
n
g10o
u
rp
c
r
c
e
p
t
i
o
n
t
h
e
yc
a
ns
o
l
v
eanyproblemw
i
t
h
nor
e
s
e
r
v
a
t
i
o
n
swhatsoever.heya
r
ei
n
c
r
e
d
i
b
l
yb
a
s
i
ci
nc
o
n
s
t
r
u
c
t
i
o
ns
i
m
p
l
et
o

ande
a
s
yt
od
e
p
l
o
y
.Theyo
n
l
yr
e
q
u
i
r
eal
o
toft
i
m
et
os
e
a
r
c
ht
h
ed
e
s
i
g
ns
p
a
.
program

10
O
P
T
I
M
I
Z
A
T
I
O
NTOOLBOX
FROMM
A
T
L
A
B

leMA
11.ABi
f
l
yo
fp
r
o
d
u
c
l
si
n
c
l
u
d
e
st
h
eO
p
t
i
m
i
z
a
t
i
o
nToolbox
whichi
sal
i
b
r
a
r
v

f
i
l
. nbeu
s
e
dt
o vep
l
emsi
n aofo
p
t
i
m
i
z
o
n
.1
tcon
ofm
manyoft
h
emethodst
h
a
thaveb
e
e
ne
x
p
l
o
r
e
di
ni
sbook
ands
e
v
e
r
a
la
ta
r
en
o
t
s
c
o
n
s
i
d
e
r
i
n
gt
h
ep
r
a
c
t
i
c
eofd
e
s
i
g
no
p
t
i
m
i
z
a
t
i
o
n
a
v
a
i
l
a
b
l
ei
nt
h
eT
o
o
l
b
o
x
.Foro
t
h
eO
p
t
i
m
i
z
a
t
i
o
nToolboxi
saw
o
r
t
h
yi
n
v
e
s
t
m
e
n
.
tManyofthet
e
c
h
n
i
q
u
e
s
s
e
r
i
o
u
s
l
y
l
1u
s
t
r
a
t
e con

p
t
s de
m
f
i
l
e
s
andd
i
s
c
u
s
s
i
o
n
si
nt
h
ebookwerek
e
p
ts
i
m
p
l
et
oi
wereu
s
e
dt
op
r
o
v
i
d
eMA
11.ABp
rogramminge
x
p
e
r
i
e
n
c
e
.Mostoft
h
em
f
i
l
e
sa
v
a
i
l
a
b
l
e
willneedtobed lopedfurtherforscriousprac application.
one w
An
o
t
h
e
ri
m
p
o
r
t
a
n
tr
e
a
s
o
nf
o
rh
a
v
i
n
ga
c
c
e
s
st
oeO
p
t
i
m
i
z
a
t
i
o
nToolboxi
s i
t
1
1
.
A
B
'So
pens
o
u
r
c
ep
o
l
i
c
ya
ti
st
oh
a
v
e c
e
s
s
a
l
l
o
w
sone1
0t
a
k
ea
d
v
a
n
t
a
g
eofMA
1
0t
h
es
o
u
r
c
ef
i
l
e
s
.Youmaybea
b
l
et
omodifyande
x
t
e
n
dt
h
o
s
efilesmeety
o
u
r
u
i
r
m
e
n
t
s
.Whilesomeoft
h
eprogramsfromt
h
eToolbox
s
p
e
c
i
a
lo
p
t
i
m
i
z
a
t
i
o
n q
nu
s
e
do
rr
e
f
e
r
e
n
c
e
de
l
s
e
w
h
e
r
ei
ni
sbooki
sc
h
a
p
t
e
rc
o
l
l
e
c
t
sthemi
na
h
a
v
eb
yr
e
f
e
r
e
n
c
e
.Asc
a
nbee
x
p
e
c
t
e
d
t
h
em
a
t
e
r
i
a
lofi
sc
h
a
p
t
e
rw
i
l
l
s
i
n
g
l
ep
l
a
c
ef
o
r s
h
ec
h
a
p
t
e
ri
sb
a
s
e
donMA
11.ABV
e
r
s
i
o
n5
.
3
.Asofi
sw
r
i
t
i
n
g
V
e
r
s
i
o
n
bed
a
t
e
d
.
'

t
h
eMATtABp
r
o
d
u
c
t
s
e
s
p
e
c
i
a
l
l
yt
h
eO
p
t
i
m
i
z
a
t
i
o
n
6
.
0
f
6
.
1
w
a
ss
h
i
p
p
i
n
g
.
I
ng
e
n
e
r
a
l
havereachedal
e
v
e
lofm
a
t
u
r
i
t
ys
u
c
h tnewerv
e
r
s
i
o
n
s

s
h

o
Toolbox
i

ss
e

on
.Wh
a
ts
h
o
u
l
dbeexp
t
e
d
however

i
s t
i
n
v
a
l
i
d
a
t
et
h
ei
n
f
o
n
n
a
t
i
o
ni
n
sc
h
a
p
t
e
rwi
1
1mostl
i
k
e
l
ybeu
p
g
r
a
d
e
d

enhanced

ande
x
t
e
n
d
e
dw
i
t
h
t
h
em
a
t
e
r
i
a
li
ni
e
a
c
hnewr
e
l
e eofMA
.

7Q

3
8
0 O
P
T
I
M
I
Z
A
T
I
O
NT
O
O
L
B
O
XF
R
O
MM
A
T
I
.
A
I
l
1
0
.
1

1
0
.
1 THEOP
T
lMIZA
T
lONTOOLBOX
Mosto
ft
h
eb
a
s
i
ci
n
f
o
r
m
a
t
i
o
ni
ni
sc
h
a
p
t
e
ri
sf
r
o
m u
s
e
r
'
sg
u
i
d
ef
o
rt
h
e
fyoua
r
eb
u
y
i
n
gas
p
e
c
ia
1l
yb
u
n
d
l
e
dCOt
h
a
ti
n
c
l
u
d
e
st
h
e
O
p
t
i
m
i
z
a
t
i
o
nT
o
o
1
b
o
x[
1
]
.I
O
p
t
i
m
i
z
a
t
i
o
nT
o
o
l
b
o
x
t
h
e
ny
o
umayn
o
th
a
v
eR
e
f
e
r
e
n
c
e1i
nb
o
o
kf
o
n
n
.Youmay
h
a
v
eR
e
f
e
r
e
n
c
e1a
saPOFf
i
l
eont
h
cCOo
ryoumayh
a
v
ea
c
c
e
s
st
oi
t
.F
o
re
x
n
m
p
l
e
o
neWindowsNTp
l
a
t
f
o
r
mf
o
rd
c
f
a
u
l
td
i
r
c
c
t
o
r
yi
n
s
t
a
1
1
a
t
i
o
nt
h
ePOFv
e
r
s
i
o
no
f
R
e
f
e
r
e
n
c
e1i
sa
v
a
i
1
a
b
l
ca
t

MATLABR1
1
h
e
l
p

p
d
C
d
o
c

o
p
t
i
m

o
p
t
i
m
_
t
b
.
p
d
f
C
:
R
e
f
c
r
e
n
c
e1p
r
o
v
i
d
c
se
x
a
m
p
l
c
so
fu
s
cd
i
s
c
u
s
s
e
ss
p
i
a
1c
a
s
e
sa
1
g
o
r
i
ms e
d
d
many0

rs
i
g
n
i
f
i
c
a
n
tt
o
p
i
c
s
.
i
sc
h
a
p
t
e
rd
o
e
sn
o
ta
t
t
e
m
p
tt
od
o
c
u
m
e
n
tt
h
e
m
.
L
i
m
i
dd
u
p
l
i
c
a
o
ni
s e
dt
ob
r
i
n
ga
t
t
e
n
t
i
o
nt
ot
h
em
o
s
tu
s
e
f
u
ll
f
t
u
r
e
s
.

b
e
1
l
u
s
t
r
a
t
i
n
gt
h
eu
s
eo
feT
o
o
l
b
o
xf
o
rv
a
r
i
o
u
se
x
a
m
p
l
e
si
n
d
u
d
emphisi
soni
t
h
eb
o
o
k
.I
nf
a
c

t al
n
r
g
enumbero
ft
h
et
e
c
h
n
i
q
u
e
si
nt
h
T
o
o
l
b
o
xa
r
en
o
tu
s
e
di
ne
ss
e
c
t
i
o
n
.Th
ee
x
p
e
r
i
e
n
c
ep
r
o
v
i
d
e
di
nt
h
i
sc
h
a
p
t
e
r
b
o
o
ka
n
da
r
en
o
tc
o
v
e
r
e
di
ni
e
c
t
i
v
c
l
y
.
s
h
o
u
l
db
es
u
f
f
i
c
i
e
n
tt
ou
n
d
e
r
s
t
a
n
da
n
du
s
et
h
enewo
n
e
sc
Tou
s
ecp
r
o
g
r
a
m
so
rm
.
f
i
l
c
si
nt
h
eT
o
o
l
b
o
x
t
h
r
e
i
m
p
o
r
t
a
n
ta
c
t
i
o
n
sm
u
s
tb
c
i
n
i
t
i
a
t
e
d
.One
f
o
r
m
u
l
n
t
eep
r
o
b
l
e
mi
nt
h
ef
o
r
m
a
te
x
p
c
c
t
e
db
yMATLABw
i
t
h p
c
c
t
t
o cp
a
r
t
i
c
u
l
a
rf
u
n
c
t
i
o
nt
h
n
tw
i
l
lb
cc
a
l
I
e
d
. Mos
t
1
yi
sw
i
l
li
n
v
o
l
v
u
s
i
n
g
a
p
p
r
o
p
t
ep
r
o
c
e
d
u
r
e
st
os
c
tt
h
er
i
g
h
t
.
h
a
n
ds
i
d
eo
ft
h
ec
o
n
s
t
r
n
i
n
t
st
oz
e
r
oo
rc
o
n
s
t
a
n
t
1
u
e
s
.Th
eT
o
o
l
b
o
xa
1
s
om
a
k
e
sap
o
i
n
to
fd
i
s
t
i
n
g
u
i
s
h
i
n
gb
e
t
w
e
c
nl
i
n
e
a
ra
n
d
va
n
o
n
l
i
n
e
a
rc
o
n
s
t
r
a
i
n
t
s
. Sc
o
n
ds
et
/
c
h
a
n
g
e ep
a
r
a
m
e
t
e

f
o
r eo
p
n
i
z
n
t
i
o
n
a
1g
o
r
i
t
h
m
s
.Th
i

u
s
et
h
ea
p
p
r
o
p
r
i
a
t
eo
p
t
i
m
i
z
e
ra
n
de
n
s
u
r
e ava
1i
ds
o
l
u
t
i
o
nhas
b
c
e
no
b
t
a
i
n
e
d
.A
I
Io
fs
ea
c
t
i
o
n
sn
a
t
u
r
a
1l
yd
e
p
e
n
do
nt
h
ea
c
t
ua
1m
a
t
h
c
m
a
t
i
c
a
lm
o
d
e
l
o
fe
p
r
o
b
l
c
m
.
0
.
1
.
1

Programs

Th
cd
i
f
f
e
r
e
n
tt
y
p
e
so
fo
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
e
m
s a
tc
a
nb
cs
o
l
v
e
dw
i
t
h
cv
e
r
s
i
o
no
f
t
h
eT
o
o
l
b
o
xi
n
s
t
a
l
l
e
don c
o
m
p
u
t
e
rc
a
nb
ed
i
s
c
o
v
e
r
e
di
nt
h
eCommandwindow

H
e
l
pwindow
o
ri
nt
h
eb
r
o
w
s
e
ru
s
i
n
gH
e
l
pO
e
s
k
.I
neMATLABCommandwindow
t
y
p
e

>>help

optim

Th
eo
u
t
p
u
ti
sr
e
p
r
o
d
u
c
d
b
e
l
o
w
.Theya
r
ea
1
1
o
ft
h
ev
a
r
i
o
u
sm
.
f
i
l
e
si
nt
h
eT
o
o
l
b
o
xa
n
d
t
h
e
i
ro
p
t
i
m
i
z
a
t
i
o
nf
u
n
c
t
i
o
ni
sa
1
s
oi
d
c
n
t
i
f
i
e
d
.

Optimization Toolbox.
Version 2.0 (Rl1) 09Oct1998
What's new.
Readme - New featuresbug fixesand changes in this
version.

T
H
EO
P
T
I
M
I
Z
A
T
I
O
NT
O
O
l
B
O
X 3
8
1

Non1inear minimization of functions.


faiabad--Scalar bounded nonlinear
minimization.

mi:22::

Mulimen

Eunction

constrained nonlinear

minsearch-Multidimensionalunconstrained n
onlinear
minimizationbyNelder-Meaddirect searchmethod.
fmiaunc-Multidimensional unconstrained nonlinear
minimization.
fgaminf-MIlltidimensional constrained minimization
semi-infinite constraints.

Nonlinear minimization of multi-objecive functions.


fgoa1at?ain-Multidime iona1 goa1 attainment
optimization.
minimax M
u1tidimensiona1 minimax optimization.
Linear 1east squares (
o
f matrixprob1ems).
1aqlin-Linea?leastsquares with linear constraints.
:t::-LLn
least ares with non ativity

Nonlinear least squares (


o
f functions).
?sqcuwefit-Nonlinear curvefittingvia least squares
(
w
thbo
l
d
s
).
lsqnonlin-Nonlinearleastsquareswithupperandlower
bounds.
Non1inear zero finding (equation solving).
~zero -S
calar nonlinear zero finding.
fsolve - Non1inear system of equations s01ve function
Solve.
Minimization of matrixproblems.
1inprog -Linear programming
quadprog -Quadratic programming.
Controlling defaults and options.
optimset
Create or a1ter optimization OPTONS
structure.
optimget - Get optimization parameters from OPTIONS
structure.
Demonstrations of large-sca1e methods.
circusteat-Quadratic programming tofind Shape of a
circus ent.
molecule-MOlecule conformation SOILltion using .
unconstrained nonlinear minimization.

P
T
I
M
I
Z
A
T
I
O
NT
O
O
L
B
O
X 383
1
0
.
1 THEO

382 OP
T
lM
I
Z
A
T
I
O
NTOOLBOXFROMMAn

optdeblur
mage deblurring using bounded linear
least-squares.

C(X) <= 0Ceq(X)


LB <= X <= UB

Demonstrations of medium-scale methods.


optdemo - Demonstration menu.
tutdemo - Tutorial walk-through.
bandemo - Minimization of banana function.
goaldemo - Goal attainment.
dfildemo - Finite-precision filter design (requires
Signal Processing Toolbox).
datdemo - Fitting data to a curve.

X=FMINCON(FUN
XO
A
B)
XO
A
B
Aeq
Beq)
X=FMINCON(FUN
X=FMINCON(FUN
XO
A
B
Aeq
Beq
LB
UB
XO
A
B
Aeq
Beq
LB
UB
NO
C
ON)
X=FMINCON(FUN
X=FMINCON(FUN
XO
A
B
Aeq
Beq
LB
UB
NONLCON
OPTONS)
X=FMINCON(FUN
XO
A
B
Aeq
Beq
LB
UB
NONLCON
OPTIONS
Pl
P2
.
.
.)
[
X
FVAL]=FMINCON(FUN
XO
.
..
)
[
X
FVAL
EXITFLAG]=FMINCON(FUN
XO.
.
.
)
[
X
FVAL
EXITFLAG
OUTPUT]=FMINCON(FUN
XO
.
.
.
)
[
X
FVAL
EXTFLAG
OUTPUT
LAMBDA]=FMINCON(FUN
XO
.
.
.
)
(
X
FVAL
EXTFLAG
OUTPUT
LAMBDA
GRAD]=FMINCON(FUN
XO.
.
.
)
[
X
FVAL
EXTFLAG
OUTPUT
LA
DA

GRAD
HESSIAN]=FMINCON
FUN
XO
.
.
.
).

r
e
s
o
f Tool

Thetutdemoi
srecommendcdf
o
rad
e
m
o
n
s
t
r
a
t
i
o
no
f u
Of t
h
e above i
t
e
m
soptimsetlinprogquadprog
fminunc and
sc
h
a
p
t
e
r
.
fmincona
r
ef
u
r
t
h
e
re
x
p
l
o
r
e
di
ni
u
n
c
t
i
o
n
sc
a
nbeo
b
t
a
i
n
e
dfromt
h
et
a
b
u
l
a
r
Ac
l
e
a
r
e
rp
i
c
t
u
r
eo
ft
h
ec
a
p
a
b
i
l
i
t
yo
fef
g
r
a
p
h
i
r
c
p
r
e
s a
1o
ft
h
cP lems
c
t
si
nt
h
ef
i
r
s
tc
h
a
p
l
e
ro
fR
e
f
e
r
e
n
c
e1
.F
o
r
exampleef
o
u
rt
c
c
h
n
i
q
u
e
sr
e
f
e
r
e
n
c
e
di
ni
sc
h
a
p
t
e
ra
r
ei
I
lu
s
t
r
a
t
e
di
nT
a
b
l
e1
0
.1
.
1
0
.
1
.
2 UslngPrograms
Eacbo
ft
h
cf
u
n
c
t
i
o
n
si
nt
h
eToolbox i
m
p
l
e
m
e
n
t
sa
no
p
t
i
m
i
z
a
t
i
o
st
c
c
h
l
l
i
q
u
ec
a
n
be e
di
ns
e
v
a
lw
a
y
s
.
'
h
e
s
ea
r
edocumentedi
nR
e
f
e
r
e
n
c
e1anda
r
ea
l
s
oa
v
a
i
l
a
b
l
e
o
n
l
i
n
et
h
r
o
u
g
honeo
ft
h
emechanismso
fob
i
n
i
n
gh
e
l
pi
nMATLABForexample
typMghelpmneoninthCommandwindowidentifiest
h
e
f
o
l
l
o
w
i
n
g
d
i
f
f
e
m
t
w
a
y
s
o
f
u
s
i
n
g f
u
n
c
t
i
o
n
l
p
r
o
g
r
a
m(
t
h
ef
o
l
l
o
w
i
n
gi
se
d
i
t
e
dt
oshowo
n
l
yt
y
p
e
so
fu
s
a
g
e
)
.
FMXNCON Finds the constrained minimum of a function of
several variables.
FMNCON solves problems of the form:
min F(X) subject to:
A*X <= BAeq*X Beq (linear constraints)
r
i
p
t
i
o
no
rSomeT
o
o
l
b
o
xF
u
n
c
t
i
o
n
s
T
a
b
l
e1
0
.
1 G
r
a
p
h
i
c
a
lD c
N
o
t
a
t
i
o
n

Typ

f
u
n
c
u
o
n
fminunc
fminsearch

i
n
.
f
(
x
)
U
n
c
o
n
s
a
i
n
c
dm
i
n
i
m
i
z
a
u
o
n m
Tx s
m
i
n
.
f
u
c
ht
h
a
t

linprog

L
in
e
a
rp
r
o
g
r
a
m
m
i
n
g

[
A
]
[

x
][
b
]
;[
A

[
]
x]=[b

]
; l~x u

Q
u
a
d
r
a
u
cp
r
o
g
r
a
m
m
i
n
g

[
AJ
[
x
]
[
b
]
;
[
A.
J
[

;
x
]=
.
[
b
.
q

l xSu
m
n

n
.
f
(
x

:
)
;such at c(x
)~O
Ceq(X:
)=0

quadprog

C
o
n
s
n
e
dm
i
n
i
m
i
z
a
u
o
n

[
AJ
[
.

X
]S[
b
)
;

fmincon

l+

[
A
.

q
U

x]=[b

]
; l~ xSu

0 (nonlinear constraints)

Th
e
r
e 1
2wayso
f t
h
e
n
i
n
c
o
nf
u
n
c
t
i
o
n
.1
Cd
i
e
r
e
n

sd
e
p
e
n
done
i
n
f
o
r
m
a
t
i
o
ni
nt
h
em
a
t
h
c
m
a
t
i
c
a
lmodela
ti
sb
c
i
n
gu
s
o
oa
n
dt
h
ep
o
s
t
o
p
t
i
m
i
z
a
t
i
o
n
c
t
i
o
nc
a
l
l
s
ym
i
s
s
i
n
gd
a
t
a
i
n
f
o
r
m
a
t
i
o
nt
h
a
ts
h
o
u
l
db
ed
i
s
p
l
a
y
e
d
.I
nu
s
i
n
gt
b
ed
i
f
f
e
r
c
n
tn
a
tdon
o
te
x
i
s
tf
o
rt
h
ep
a
r
t
i
c
u
1
a
re
x
a
m
p
l
e
sm
u
s
tb
ei
n
d
i

t
O
Ou
s
i
n
gan
u
l
lv
e
c
t
o
r(
0
)
.

b
e
r
ea
r
eeke
v
o
r
d
t
h
eabovel
i
s
to
ff
u
n
c
t
i
o
nc
a
l
l
s a
tn
e
e
ds
p
e
c
i
a
l
a
l
t
e
n
t
i
o
ni
nc
o
n
s
i
n
O
On
o
n
l
i
n
e
a
ro
p
t
i
m
i
z
a
t
i
o
n
.

si
n
d
i
c
a
si
ft
h
co
p
t
i
m
i
z
a
t
i
o
nwass
u
c
c
e
s
s
f
u.
l
EXITFLAG:F
o
rf
m
i
n
c
o
ni
If EXITFLAG is:
>0 then FMINCON converged to a solution X.
o then the maximumnumber of function evaluations was
reached.
< 0 then FMINCON did not converge to a solution.
OUTPUT: i
ss
p
e
c
i
a
1MATLABs
t
r
u
c
t
u
r
ee
l
e
m
e
nt
.S
t
r
u
c
t
u
r
ei
saprogramming
tu
s
O
Oi
nmanyh
i
g
h
e
r
I
e
v
e
l
l
a
n
g
u
a
g
e
s
.I
ti
ssim
i
1
a
rt
oa
r
r
a
yi
nMATLAB.An
c
o
n
s c
a
r
r
a
yw
i
l
lc
o
n
t
a
i
no
n
l
ys
i
m
i e
l
e
m
e
n
t
s ukca
l
lnumbcrso
ra
l
ls
t
r
i
n
g
s
.As
c
t
u
r
e
ont
h
eo
t
h
e
rh
a
n
d

c
a
nc
o
n
t
a
i
nd
i
s
p
a
r
a
t
ee
l
e
m
e
n
t
s
s
u
c
ha
ss
e
v
e
r
a
1numbersands
t
r
i
n
g
s
.
Thee
l
e
m
e
n
t
so
fes
t
r
u
c
t
u
r
ec
a
nbea
c
c
e
s
s
e
dbyu
s
i
n
gt
h
enameo
ft
h
e5
C

e
l
e
m
e
n
tf
o
l
l
o
w
e
dbyap
e
r
i
o
dandCP
I
e
f
i
n
O
Okeywordf
o
rt
h
ep
a
r
t
i
c
u
l
a
re
l
c
m
e
n.
t
Forc
x
a
m
p
l
e

twoo
ft
h
ee
l
e
m
e
n
t
so
ft
h
eOUTPUTa
r
e
:
OUTPUT.iterations:t
h
enumbero
ff
u
n
c
t
i
o
nc
v
a
l
u
a
t
i
o
n
s
OUTPUT.algorithm: thenameofea
l
g
o
r
i
t
h
mu
s
O
Ot
os
o
l
v
et
h
ep
r
o
b
l
c
m
LAMBDAITh
i
si
sa
l
5
0a
n
o
t
h
e
rs
t
r
u
c
t
u
r
ei
nMATLABandc
o
n
t
a
i
n
s L
a
g
r
a
n
g
e
mt
i
p
u
e
r
s
.1
ev
a
l
u
e
s
o
fe
sm
u
l
t
i
p
l
i
e
r
sa
r
eu
s
e
f
u
lf
o
rv
e
r
i
i
n
gt
h
es
a
t
i
s
f
a
c
t
i
o
no
f
en
e
c
e
s
s
a
r
yc
o
n
d
i
t
i
o
n
so
fo
p
t
i
m
i
z
a
t
i
o
n
.Th
ev
a
l
u
e
sc
a
nber
e
c
o
v
e
r
b
yesame
p
r
'Ci
n
d
i
c
a
t
e
df
o
rt
h
cOUTPUTs
t
r
u
c
t
u
r
c
.

384 O
P
T
I
M
I
Z
A
n
O
NT
O
O
L
B
O
XFROMM
A
n

LAMBDA.lower:v
a
l
u
eo
ft
h
em
u
l
t
i
p
l
i
e
r
sa
tt
h
el
o
w
e
rb
o
u
n
d
LAMBDA.ineqnonlin: v
a
l
u
eo
ft
h
em
u
l
t
i
p
l
i
e
r
st
i
e
dt
ot
h
ei
n
e
q
u
a
l
i
t
y
(s'si
no
k
)
c
o
n
sn
The
s
ci
t
e
m
sw
e
r
ca
s
s
o
c
i
a
t
e
dw
i
t
ht
h
ef
u
n
c
t
i
o
n
n
i
n
c
o
n
.O
t
h
e
rf
u
n
c
t
i
o
n
smayh
a
v
c
o
fi
n
f
o
r
m
a
t
i
o
na
v
a
i
l
a
b
l
.
d
i
f
f
e
r
e
n
t
l
a
d
d
i
t
i
o
n
a
lt
y
p
e

1
0
.
1
.
3 S
e
t
t
i
n
gO
p
t
l
m
i
z
a
t
i
o
nParameters
Th
e
r
ea
r
s
e
v
e
r
a
lp
a
r
a
m
e
t
c
r
st
h
a
ta
r
eu
s
e
dt
oc
o
n
t
r
o
lt
h
ea
p
p
l
i
c
a
t
i
o
no
fap
a
r
t
i
c
u
l
a
r
o
p
t
i
m
i
t
i
o
nt
e
c
h
n
i
q
u
e
. Thc
s
ep
a
r
a
m
e
t
e
r
sa
r
ea
v
a
i
l
a
b
l
et
h
r
o
u
g
ht
h
c OPTIONS
s
t
r
u
c
l
u
r
e
.AI
lofthema
r
ep
r
o
v
i
d
e
dw
i
t
hd
e
f
a
u
l
tv
a
l
u
e
s
.T
h
c
r
ew
i
l
lb
eo
c
c
a
s
i
o
n
swhen
c
e
s
s
a
r
yt
oc
h
a
n
g
et
h
ed
e
f
a
u
l
tv
a
l
u
e
s
.Onee
x
a
m
p
l
emayb
ea
ty
o
u
rc
o
n
s
t
r
a
i
n
t
i
ti
sn
r
er
e
q
u
i
r
e
dt
obel
o
w
e
rt
h
t
h
ed
e
f
a
u
l
tOrEXITFLAGrecommendsa
t
o
l
e
r
a
n
c
e
sa
b
l
e
m
.Th
e
s
ep
m
e
t
e
r
sa
r
es
e
t
l
a
r
g
e
rnumbero
ff
u
n
c
t
i
o
ne
v
a
l
u
a
t
i
o
n
st
os
o
l
v
eEo
a
m
e
/
v
a
l
u
ep
r
.Gen
c
r
a
l
l
yenamei
su
s
u
a
l
l
yas
t
r
i
n
gw
h
i
l
ct
h
ev
a
l
u
ei
s
u
s
i
n
gan
t
y
p
i
c
a
l
l
yanumb
r
.Thel
i
s
to
fp
a
r
a
m
e
t
e
r
si
sa
v
a
i
l
a
b
l
ei
nR
e
f
e
r
e
n
c
e1a
n
dc
a
nb
e
a
c
c
e
s
s
e
dm
o
s
te
a
s
i
l
yi
nt
h
eCommandwindowb
yt
y
p
i
n
g
>>

help optimset

1
0
.
2

E
X
A
M
P
L
E
S 385

TolCon
Termination tolerance on the constraint
violation [positive scalar J
TolFun. Termination tolerance on the function value
positive scalar J
TolX -Termination tolerance on X [positive scalar 1
Thep
a
r
a
m
e
t
e
rnamesa
r
ei
d
e
n
t
i
f
i
e
di
nb
o
l
df
o
n1
.Th
e
s
ep
a
r
a
m
e
t
e
r
sa
r
et
h
eo
n
e
s
wi
1
1c
o
n
1
as
t
a
n
d
a
r
do
p
t
i
r
n
i
z
a
t
i
o
na
p
p
l
i
c
a
t
i
o
n
.Thed
e
f
a
u
l
tv
a
l
u
e
sf
o
r p
a
r
a
m
e
r
a
r
eg
i
v
e
nw
i
t
h
i
nt
h
ec
u
r
l
yb
r
a
c
e
s
.F
o
re
x
a
m
p
l
e
:
>> op
ions
opimset(
Diagnostics''
o
n
''Display'
'iter''Maxlter''500''HessUpdate''dfp'
'TolCon'
'
l
.
Oe-OS')

Ut
h
ea
b
o
v
es
t
a
t
e
m
c
n
ti
su
s
e
dp
r
i
o
rt
oa
c
t
u
a
l
l
yc
a
l
l
i
n
ga
no
p
t
i
m
i
z
a
t
i
o
nt
e
c
h
n
i
q
u
e
e
d
e
f
a
u
l
tv
a
l
u
e
sf
o
rt
h
o
s
ep n
e
t
e
r
swi
1
1b
ec
h
a
n
g
e
di
nt
h
eOPTIONSs
t
r
u
c
t
u
r
e
.Th
e
d
e
f
a
u
l
tv
a
J
u
e
sf
o
rap
a
r
t
i
c
u
l
a
rt
e
c
h
n
i
q
u
ec
a
nbef
o
u
n
db
yt
y
p
i
n
gt
h
eCommandwindow
>>

options

opimset(
fmincon')

l
e
r
e

s
e
v
e
r
a
lo
t
h
c
rwayst
ou
s
et
h
eo
p
t
i
m
s
e
tp
r
o
g
r
a
mi
t
s
e
l
f
.F
o
rd
e
t
a
i
l
sc
o
n
s
u
l
t
R
e
f
e
r
e
n
c
e1
.

Ap
a
r
t
i
a
l
l
i
s
to
ft
h
ep
a
r
a
m
e
t
e
r
si
nt
h
eOPTIONSs
t
r
u
c
t
u
r
ec
o
p
i
e
df
r
o
mt
h
eCommand
w
i
n
d
o
w
.f
o
l
l
o
w
s
.

1
0
.
2 EXAMPLES

OPTIMSET PARAMETERS
Diagnostics print diagnostic information about the
function to be minimized or solved [on I{off}l
DiffMaxChange -Maximum change in variables for finite
difference gradients [positive scalar I
.
{
l
e
l
}1
DiffMinChange - Minimum change in variables for finite
difference gradients [positive scalar I{leS
}1
Display - Level of display [off Iiter I{final} ]
GradConstr - Gradients for the nonlinear constraints
defined by user [on I{off} 1
GradObj - Gradient(s) for the objective function(s)
defined by user [on I(off}l
Hessian -Hessian for theobjective functiondefinedby
user [on I{off} ]
HessUpdate -Quasi-Newtonupdatingscheme [{bfgs} I
dfp
Igillmurray Isteepdesc 1
LineSearch
Type
Line search algorithm choice
cubicpoly I{quadcubic} 1
MaxF
unEv
als - Maximum number of function evaluations
allowed [positive integer 1
Maxlter
Maximum number of iterations allowed
positive integer ]

1
e
x
a
m
p
l
e
si
1
1
u
s
a
t
e
di
n i
ss
e
c
t
i
o
na
r
ec
o
n
t
i
n
u
o
u
so
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
e
m
s
d
i
s
c
u
s
s
e
di
neb
o
o
k
.l
e
ya
r
ef
r
o
mI
in
e
a
rp
r
o
g
r
a
m
m
i
n
g
.q
u
a
d
r
a
t
i
cp
r
o
g
r
a
m
m
i
n
g
.

u
n
c
o
n
s
t
r
a
i
n
e
dm
i
n
i
m
i
z
a
t
i
o
n
.a
n
dc
o
n
s
t
r
a
i
n
e
dm
i
n
i
m
i
z
a
t
i
o
o
.O
n
l
yt
h
ea
p
p
r
o
p
r
i
a
t
e
m
a
t
h
e
m
a
t
i
c
a
lm
o
d
e
li
si
o
t
r
o
d
u
c
e
dp
r
i
o
rt
oi
1
1u
s
.
t
i
n
gt
h
e eo
ft
h
eT
o
o
l
b
o
xf
u
n
c
t
i
o
n
.

inearProgrammlng
1
0
.
2
.
1 L
TheLPe
x
a
m
p
l
eu
S
e
dh
e
r
ei
n
c
l
u
d
e
sa
l
lt
h
r
e
et
y
p
e
so
fc
o
n
s
t
r
a
i
n
t
s
.Thet
o
o
l
b
o
x
f
u
n
c
t
i
o
ni
sl
i
n
p
r
o
g
.
Example1
0
.
1

Maximizc

z=4
x
I+6x
x
4
2+7X3+8

S
u
b
j
e
c
t
t
o
: X
I+X2+x)+X4=
X4~

950
400

2
x1+3X2+4X3+7X4S4600

3
x
I+4x:

+6x
000
4S5

5X3

XloX2.X3.X4;
?
;

1
0
.
2E
X
A
M
P
L
E
S 387

386 O
P
T
1
M
1
Z
A
T
1
0
NTOOLBOXFAOMM
A
T
I
.
A
I
l

+
i
[!
]
[
:
J

l
el
i
n
p
r
o
gprograms
o
l
v
e
st
h
ef
o
l
l
o
w
i
n
gp
r
o
b
l
e
m
:

)
=[
12
]

u
b
j
e
c
tt
o[
A
]
(
x
]S[
b
]
Minimize fT
x; s
Theo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
nf
o
rt
h
eexamplcn
e
e
d
st
ob
c
h
a
n
g
e
dt
oam
i
n
i
m
i
z
a
t
i
o
n
p
r
o
b
l
e
m
.Th
eg
r
e
a
t
e
r o
re
q
u
a
lt
oc
o
n
s
t
r
a
i
n
tn
e
e
d
st
ob
echangedt
oal
e
s
s o
r
c
q
u
a
lt
oo
n
c
.Thel
i
n
p
r
o
gprograma
1lowse
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
t de
x
p
l
i
c
i
tboundson
v
a
r
i
a
b
l
e
s
.l
ee
q
u
a
l
i
t
yc
o
n
s
a
i
n
tc
a
nb
es
e
tupw
i
t
hAandb
Th
e
r
ea
r
en
o
tmany
p e
r
sa
tn
e
e
dt
obechangede
x
c
e
p
ts
w
i
t
c
h
i
n
go
ft
h
e'
L
a
r
ge
Sc
a
l
c
'problemt
o

o
f
f
.
'Foro
n
l
yf
o
u
rd
e
s
i
g
nv
a
r
i
a
b
l
c
sI
c
a
v
i
n
gi
tons
h
o
u
l
dbep
e
r
m
i
s
s
i
b
l
e
.Theformo
f
eu
s
e
di
s
l
i
n
p
r
o
gt
h
a
tw
i
l
lb
[
X
FVAL
EXITFLAG]

LINPROG(f
A
b
Aeq
beq
LB
UB)

1
ydocumentedi
nR
e
f
e
r
e
n
c
e1o
ri
nt
h
eh
e
l
pd
o
c
u
m
e
n
t
a
t
i
o
n
.
T
h
i
sformi
sn
o
te
x
p
l
i
c
it
Ex10_1.m1i
sem
f
t
l
et
h
a
ti
su
s
c
dt
os
e
tupandr
u
nt
h
ep
r
o
b
l
e
m
.I
ti
sn
o
td
i
s
p
l
a
y
e
d
h
e
r
eb
u
tr
u
n
n
i
n
gt
h
ef
i
l
ei
n Commandwindowp
r
o
d
u
c
e
st
h
f
o
l
l
o
w
i
n
go
u
t
p
u
t

2]
x

f
(
.
)
=[c]T[x]

.
X
]
T

i
n
i

z
z

e )=[ 1

23
]

T
M

S
u
b
j
e
c
t

x
3 2
.
5
2
x1-X2+3X3=1

x1+
2
x
2

X
1
.X
2
'x
)~

L1

1 1
1
2 ll
[
x
]

14
1

Theprogramt
ob
eu
s
e
dfromt
h
eToolboxi
sq
u
a
d
p
r
o
g
.Thes
t
a
n
d
a
r
df
o
r
m
a
tf
o
rt
h
i
s
program1
8

FVAL

T
T
M
i
n
i
mize f
(
x
)=cX+~ x

-6.6500e+003
EXITFLAG =<

Su
e
c
tt
o
: AxSb

A x=b.

Thes
o
l
u
t
i
o
nt
o problemi
s
x
;=0
;X
;=400; x;=150; x;=47."=6650

1
0
.
2
.
2 QuadraticProgrammlng
Q
u
a
d
r
a
t
i
cprogrammingr
e
f
e
r
st
oap
r
o
b
l
c
mw
i
t
haq
u
a
d
r
a
t
i
co
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
nand
I
i
n
e
a
rc
o
n
s
t
r
a
i
n
t
s
.Whiles
u
c
hap
r
o
b
lm c
a
nbes
o
l
v
e
du
s
i
n
gn
o
n
l
i
n
e
a
ro
p
t
i
m
i
z
a
t
i
o
n
t
c
h
n
i
q
u
e
so
fC
h
a
p
t
e
r7
.s
p
e
c
i
a
lqua
t
i
cprogramminga
1g
o
r
i
t
h
m
sb
a
s
e
dont
h
e
q
u
e
so
fI
in
e
a
rprogramminga
r
en
o
r
m
a
l
l
y e
da
se
ya
r
es
i
m
p
l
eandf
a
s
t
e
r
.I
f
t
e
c
h
t
h
eo
b
j
e
c
t
i
v
cf
u
n
c
t
i
o
ni
saq
u
a
d
r
a
t
i
cf
u
n
c
t
i
o
ni
nt
h
ev
a
r
i
a
b
l
e
s
t
h
e
ni
tc
a
nb
ec
x
p
r
e
s
s
e
d
inqua
f
o
r
m
.Forexamplc
f(x
1

)=x1+2x2 + +X2+~

c
a
nb
ec
x
p
r
e
s
s
e
d
l
c
s1
0bed
o
w
n
l
o
n
d
e
df
r
o
mt
h
cw
e
bs
i
l
en
r
ci
n
d
i
c
a
l
c
db
yb
o
l
d
f
a
c
e n
ss
e
r
i
ft
y
p
e
.

I
F
i

[
x
]

Exam
p
l
e1
0
.
2S
o
l
v
et
h
equa
a
t
i
cprogrammingp
r
o
b
l
e
m

Optimization terminated successfully.


0.0000
400.0000
150.0000
400.0000

LBSxSUB

Th
es
o
l
u
t
i
o
nc
a
nbef
o
u
n
di
nEx
10_2.m.Th
ed
e
f
a
u
l
to
p
t
i
m
i
z
a
t
i
o
np :
t
e
r
sa
r
e
u
s
e
ds
oa
tt
h
eo
p
t
i
o
n
ss
t
r
u
c
t
u
r
ed
o
e
sn
o
th
a
v
ct
ob
ec
h
a
n
g
e
d
.Thea
c
a
Ic
a
l
l
i
n
g
s
t
a
t
e
m
e
n
tu
s
e
di
s
[
x
fval
EXTFLAG]

quadprog(H
c
A
b
Aeq
beq
LB
UB)

Thes
o
l
u
t
i
o
ni
sf
o
r
m
a
t
t
e
dandd
i
s
p
l
a
y
e
di
nt
h
cCommandwindow.1
ti
sc
o
p
i
e
dfrom
t
h
eCommandwindow:
Optimization terminated successfully.
Final Values
optimum Design Variables

------

0.5818 1.1182 0.3182


Optimum function value
-1.9218


388 O
P
T
I
M
I
Z
A
T
I
O
NT
O
O
L
B
O
XFAOMM
A
T
W
I
Lagrange Multipliers for inequality constraint
0.0218
nequality constraint

2.5000
Lagrange Multipliers for equality constraint
0.0164
Equality constraint
1.0000
1
0
.
2
.
3 UnconstralnedOptimization
3i
su
s
e
d1
0i
l
I
u
s
t
r
a
l
e u
s
co
ff
m
i
n
u
n
cep
r
o
g
r
a
mt
ob
e e
df
o
r
Example1
0.
i
n
ge
x
a
m
p
l
c(
a1
0
1o
f
u
n
c
o
n
s
t
r
a
i
n
e
dm
i
n
i
m
i
z
a
t
i
o
n
.1
1i
sas
i
m
p
l
ec
u
r
v
u
n
c
o
n
s
t
r
a
i
n
e
dm
i
n
i
m
i
z
a
t
i
o
np
r
o
b
l
e
m
si
ne
n
g
i
n
e
e
r
i
n
ga
r
ec
u
r
v
ef
i
t
t
i
n
gu
s
e
df
o
r
d
e
v
e
l
o
p
i
n
gc
o
r
r
e
l
a
t
i
o
nf
r
o
me
m
p
i
r
i
c
a
ld
a
t
a
)
.Th
eT
o
o
l
b
o
xh
a
sp
r
o
g
r
a
m
ss
p
e
c
i
f
i
c
a
l
l
y
tmayb
eau
s
e
f
u
li
d
e
af
o
rt
h
er
e
a
d
e
rt
oc
o
m
p
a
r
et
h
es
o
l
u
t
i
o
n
su
s
i
n
g
f
o
rc
u
r
v
ef
i
t
t
i
n
g
.I
t
h
o
s
p
r
o
g
r
a
m
s i
sp
r
o
v
i
d
e
sa
no
p
p
o
r
t
u
n
i
t
yt
ou
s
ep
r
o
g
r
a
m
sf
r
o
meT
o
o
l
b
o
xn
o
t
t
e
di
nt
h
i
sc
h
a
p
t
e
r
.
i
l
l
u
s

E
x
t
u
n
p
l
e1
0
.
3 As
e
to
fd
a
t
ao
v
e
r1
0
0p
o
i
n
t
sr
e
p
r
e
s
e
n
gac
o
s
i
n
ef
u
n
c
t
i
o
ni
s
t
e
d
.
a
t
i
cp
o
l
y
n
o
m
i
a
li
su
s
e
d1
0f
i
tt
h
ed
a
t
a
.Anu
n
c
o
n
s
t
r
a
i
n
e
dm
i
n
i
m
i
z
a
t
i
o
n
Aq
u
a
o
ft
h
ee
r
r
o
rb
e
l
w
e
e
no
r
i
g
i
n
a
ld
a
t
aa
n
dt
h
e
p
r
o
b
l
e
mb
a
s
e
dont
h
esumo
fs
q
u
o
v
e
ra
l
led
a
t
ap
o
i
n
t
si
ss
o
l
v
e
df
o
rt
h
ec
o
e
f
f
i
c
i
e
n
t
so
ft
h
ep
o
l
y
n
o
m
i
a
l
.
p
o
l
y
n
o
m
i
a
l
i
sa
ne
x
n
p
l
eo
ft
h
ec
l
a
s
s
i
c
a
l
l
e
a
s
ts
q
r
ede

.
r
p
r
o
b
l
e
m
.
f
i
I
eEx10_3.m.R
u
n
n
i
n
gt
h
ce
x
a
m
p
l
ea
l
s
od
i
s
p
l
a
y
s
Thee
x
a
m
p
l
ei
ss
o
l
v
e
di
n mac
o
m
p
a
r
i
s
o
no
ft
h
ef
i
tonaf
i
g
u
r
ew
h
i
c
hi
sn
o
ti
n
c
l
u
d
e
dh
e
r
e
.Twos
o
l
u
t
i
o
n
sa
r
e
o
b
t
a
i
n
e
di
nt
h
ee
x
a
m
p
l
es
t
a
r
t
i
n
gf
r
o
mt
h
esamei
n
i
t
i
a
lg
u
e
s
s
.Thef
i
r
s
to
n
ei
ss
o
l
v
e
d
e
r
s
t
p
l
i
e
dg
r
a
d
i
e
.Th
es
e
c
o
n
di
st
h
ed
e
f
a
u
l
tf
i
n
i
t
ed
i
f
f
e
r
e
n
c
e d
i
e
n
t
s
.
u
s
i
n
g '
Ther
e
s
u
l
t
sa
r
c
o
p
i
e
df
r
o
mt
h
eCommandw
i
n
d
o
w
:
Final Values Oser specified Gradients
Optimum DesignVariables
-0.4309 -0.0365 1.0033
Optimum function value
2.4055e-004
Gradients of the function

1
0
.
2E
X
A
M
P
L
E
S 389
Fiaalvalue--Default Finite Difference Gradients
Optimum Design Variables

--

0
.4309 -0.0365 1.0033
Optimum function value

2.4055e-004
Gradients of the function

1.0e-004 *
-0.0121 -0.0270 0.5715
o
fg
r
a
d
i
e
n
t
swouJdJ
e
a
d1
0ai
f
t
e
ra
n
d
I
ti
sg
e
n
e
r
a
l
l
yb
e
l
i
e
b
e
t
t
e
r u
t
i
o
n
.F
o
ri
t
nunct
o g
r
a
d
i
e
n
t
sc
o
m
p
u
t
e
db
y c
h
a
n
g
c
sh
a
v
c
p
t
i
o
n
ss
c
l
u
w
e
l
la
st
h
ef
i
l
ewh
t
h
eo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
ni
s
t
ob
cmadet
oeo
i
(IJ10_3m).In onsstruct e'OradObj
'p
a
r
a
m
e
t
e
rs
h
o
u
l
db
c
c
s
c
tt
o'
o
n
'
i
n
I
S
I
SEmponmtO
t
h
e
r
w
i
s
eMATUBdomn
o
tc
x
p
E
C
tt
h
Eg
r
a
d
i
e
n
t
st
ob
E
p
r
o
v
i
d
b
yt
h
eu
s
e
r
.Theo
p
t
i
o
nf
o
rt
h
i
se
x
n
p
l
ei
ss
e
lb
y

;
?

TZ;;??s TolFun'

;
;
z
p
;
;
;
;
f
f
;

play'

;J:??

b
i
sc
a
u
s
e
sepro
mt
ol
o
o
kf
o
r d
i
e
n
es
e
c
o
n
de
l
e
m
o
f
t
h
er
e
t
u
m
v
a
l
u
e
.
T
h
ef
u
n
c
t
I
o
nS
t
a
t
e
m
e
n
tusdforo
b
j
1
0
3.mi
s
function [
fgrad]

objlO_3(x)

Bo
eo
c
c
t
i
v
ef
u
n
c
t
i
o
ndeg
r
a
d
i
e
(
g
r
a
d
)a
r
ec
a
l
c
u
l
a
t
e
di
nt
h
esame
i
Ie
.
f
u
n
c
t
i
o
nm-f
1
0
.
2
.
4 ConstrainedOptimizatlon

Exam
p
l
e1
0
.
4 Th
ee
x
a
m
p
l
ei
st
h
e same a
s Example 8.
4a
n
d P s
e
n
t
sl
h
e
c
o
n
s
t
r
a
i
n
e
de
x
n
p
l
ef
o
rt
h
i
ss
c
c
t
i
o
n
.

n
em
a
t
h
e
m
a
t
i
c
a
lmodeli
s

Min
i
m
i
z
e f
(
X
)=
pLAc
J
S
u
b
j
e
c
tt
o
: g
.
(
X
)
:WL
2E
IS0
.
2
5(
d
c
f
l
t
i
o
n
)
-1

g
2
(
X
)
:WL
x
.-8
/
1

yS0(normals
t
r
e
ss
)
g
3(
X
)
:WQc-2
1
x
J
'
t
yS0 (
s
h
e
a
rs
l
r
e
ss
)

1.0e-007*
0
.1324 -0.2198 -0.3275

Someg
c
o
m
c
t
r
i
cc
o
n
s
t
r
a
i
n
t
st
or
e
l
u
t
ct
h
ev
a
r
i
o
u
sd
e
s
i
g
nv
a
r
i
u
b
l
e
sa
r
e

liiIi

390 O
P
T
I
M
(
Z
A
T
I
O
NT
O
O
L
B
O
XFROMM
A
n.A
B

R
E
F
E
R
E
N
C
E 3
9
1

3X20

Final Values
Optimum Design Variables

g
4
(
X
)
:x

g
s
(
X
)
:2
x
2- x(S0

- -- - -

6.8671 2.2890 0.1802 0.2500


Optimum function value

g
6
(
X
)
:X
3-1
.
5
x
4S0
g
7
(
X
)
:0.
5
x
4-x
3S0

62.4905
Final Nonlinear Constraints

Thes
i
d
ec
o
n
s
t
r
a
i
n
t
so
nt
h
ed
e
s
i
g
nv
a
r
i
a
b
l
e
sa
r
e

- --

--

3Sx(S202Sx
50.
12
5SX
O.750
.
2
5SX4S1
.2
5
3S
2S1

I
lowingr
e
l
a
t
i
o
n
sd
e
f
i
n
et
h
ec
r
o
s
s
s
e
c
t
i
o
n
a
la
r
e
a(
A
)
'momento
fi
n
e
r
t
i
a(

a
n
d
Thef
o
c
f
i
r
s
tmomento
fa
r
e
aa
b
o
u
tt
h
ec
e
n
t
r
o
i
d(
Q
c
)
:
x
2
x
4+x
(
x

Z
Ac=2

t
x
( -xJ(x
-2xi
1
2

l +0.5x
Q
c
=
0
.
5
.

000
Lagrange Multipliers for Linear constraints

-- -

1
2

1.0e+009
-2.7304 -0.0025 -0.0001
Lagrange Multipliers for Nonlinear constrains

_XJ2

Thep
a
r
a
m
e
r
sf
o
r p
r
o
b
l
e
ma
r
w
=1
.
42000

L=9
6
E=2
9E+0
6
P=0.284
'y=36E+03

:
y=2
1E+0
3
.T
h
e
r
ea
r
et
h
r
e
en
o
n
l
i
n
e
a
ri
n
e
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
t
sa
n
df
o
u
r
l
i
n
e
a
ri
n
e
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
t
s
. l
ec
o
d
ef
o
rr
u
n
n
i
n
gt
h
ep
r
o
g
r
a
r
ni
sa
v
a
i
l
a
b
l
ei
n
Ex10_4.m.
Th
ep
r
o
g
r
a
mt
ob
eu
s
e
di
sf
r
n
i
n
c
o
n
.Thef
o
l
l
o
w
i
n
gOPTIONp
a
r
n
e
t
e
r
sa
r
e
c
h
a
n
g
e
d
.
options optimset(
LargeScale'
off'
Display'

iter'
TolFun'1
.Oe0
8
);
l
ec
a
l
l
i
n
gs
t
a
t
e
m
e
n
tu
s
e
df
o
rf
m
i
n
c
o
nw
i
t
ht
h
eo
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
na
v
a
i
l
a
b
l
ei
n
o
b
j
e
c
t
i
v
e
.
m
a
n
dt
h
en
o
n
l
i
n
e
a
rc
o
n
s
t
r
a
i
n
t
si
nn
o
n
l
i
n
_
c
o
n
s
.
mi
s

[
x fvalexitflagoutput lambda]
objective'
xO
A
B LB
UB

nonlin_cons'
fmincon(
options);
TheI
in
e
a
ri
n
e
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
t
sa
r
es
e
tupi
nt
h
em
a
t
r
i
xAa
n
dv
e
c
t
o
rB
.T
h
e
r
ea
r
e
noI
in
e
a
re
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
t
sa
n
dt
h
e
ya
r
ei
n
d
i
c
a
t
e
dbyt
h
en
uI
lm
a
t
r
i
c
c
s
.The
l
o
w
e
ra
n
du
p
p
e
rb
o
u
n
d
sa
r
ea
p
p
l
i
e
dt
h
r
o
u
g
ht
h
ev
e
c
t
o
r
s LB a
n
dUB. The
c
u
s
t
o
m
i
z
e
dp
a
r
to
feo
u
t
p
u
td
i
s
p
l
a
y
e
di
nt
h
eCommandwindowi
si
n
c
l
u
d
e
d
a
r
i
o
u
sS
t
r
u
c
t
u
r
ev
a
l
u
e
sw
i
l
lp
r
o
v
i
d
ed
e
t
a
i
l
sa
b
o
u
tt
h
e
b
e
l
o
w
.A
c
c
e
s
s
i
n
g v
p
e
r
f
o
r
m
a
n
c
eo
f a
l
g
o
r
i m
.

2.5879 0 0 0
Ther
e
a
d
e
ri
se
n
c
o
u
r
a
g
e
dt
or
u
nt
h
ee
x
a
m
p
l
ew
i
t
h
D
i
a
g
n
o
s
t
i
c
s
'o
ne
x
p
l
o
r
e
o
p
t
t
o
n
ss
t
r
u
c
t
u
r
e
a
n
deo
u
t
p
u
ts
u
c
t
u
r
e
.
REFERENCE
1
.
lmmJ-R.BmEI
M
.A
.
a
n
dG
r
n
c
e
A
.
O
p
t
i
m
i
z
a
t
i
o
nT
o
o
l
b
o
x
f
o
rU
s
ew
i
t
l
r
MATLAB
U
s
e
r
'
sG
u
i
d
e
V
e
r
s
i
o
n2

M
a
t
h
W
o
r
k
sI
n
c
1
9
9
9
.

INDEX

A
b
s
t
r
a
c
tm
a
t
h
e
m
a
t
i
c
a
1m
o
d
e
l
5
9-10
A
c
c
e
p
t
a
b
l
es
o
l
u
t
i
o
n
s1
7

2
9
3
A
c
t
i
v
ec
o
n
s
t
r
a
i
n
ts
e
t
A
c
r
o
s
p
a
c
ed
e
s
i
g
n1
A
l
g
o
r
i
t
h
m
s
2
1
s
e
eA
u
g
m
c
n
l
c
dLagmnge
ALMm
c
t
h
o
d
M
u
l
t
i
p
l
i
c
rm
e
t
h
o
d
A
n
a
1
y
t
i
c
a
1c
o
n
d
i
t
i
o
n
s
175-176
A
n
n
e
a
l
i
n
g8
c
h
c
d
u
l
c

3
5
9
Ani
f
i
c
i
a
1v
a
r
b
l
e
s
9
8
A
u
g
m
e
n
t
e
dLa
gmngeM
u
l
t
i
p
l
i
c
r(ALM)
mt
h
o
d
2
7
62
8
1
a
l
g
o
r
im
2
7
7
2
7
8
a
p
p
l
i
c
a
t
i
o
n
2
7
8
2
8
1
n
e
dp
r
o
b
l
e
m
u
n
c
o
n
s
t
r
u
n
s
f
o
r
m
i
o
n10
2
7
6
2
7
7
A
u
g
m
e
n
t
e
dm
a
t
r
i
x

1
0
3
1
0
4
Bez
.
ie
rp
a
r
a
m
e
cc
u
r
v
c
s
f
i
t
t
i
n
go
f

2
5
8
2
6
2
BFOSm
e
t
h
o
d
s
e
eB
r
o
y
d
o
n
F
l
e
t
c
h
e
r
O
o
l
d
f
a
r
b
S
h
a
n
n
ome
d
B
i
s
c
c
t
i
o
nt
c
c
h
n
i
q
u
e
2
0
9
2
1
0
Bmncha
n
dB
o
u
n
d
g
o
r
i
m
3
3
3
a
p
p
l
i
c
a
n

3
3
3
3
3
5
t
c
c
h
n
i
q
u
c
3
2
93
3
6
B
r
o
y
d
o
n
-F
1e
t
c
h
c
r
Gol
d
f
a
r
b
S
h
a
n
n
o
(
B
F
O
S
)m
e
t
h
o
d
n
1g
o
r
i
t
h
m
.2
5
0
a
p
p
l
i
c
a
t
i
o
n
.2
5
0
2
51

C
a
r
t

i
a
ns
p
a
c
e
s
.1
0
6

C(
I
a
n
g
u
a
g
e
)
2
4
C
l
a
s
s

o
f
p
r
o
b
l
c
m
s
.
2
1
C
o
m
p
u
t
a
t
i
o
na
1r
c
s
o
u
r
c
e
s
.2
g
a
t
eO
m
d
i
c
n
tm
c
t
h
o
d
.2
44-2
4
6
2
5
5
a
l
g
o
r
i
t
h
m2
4
4
2
4
5
a
p
p
l
i
c
a
t
i
o
n
2
4
5
2
4
6
C
o
n
s
t
m
i
n
e
do
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
e
m
s
.
2
6
5
3
1
53
8
9
3
9
1
A
u
g
m
e
n
t
c
dLagmngeM
u
l
t
i
p
l
i
e
r
(ALM)m
c
t
h
o
df
o
rs
o
l
v
i
n
g

2
76-2
8
1
beamd
e
s
i
g
n
3
1
0
3
1
2
c
o
n
t
r
o
l
o
p
t
i
m

1
a3
1
3
3
1
5
d
c
f
i
n
i
t
i
o
no
fp
r
o
b
l
c
m
.266
e
c
tm
e
t
h
o
d
sf
o
rs
o
l
v
i
n
g
2
8
1
3
0
2
Ex
t
c
r
i
o
rP
c
n
n
l
t
yF
u
n
c
t
i
o
nP
F
)

2
7
1
2
7
6
m
e
t
h
o
df
o
rs
o
l
v
i
n
g
f
l
a
g
p
o
l
ep
r
o
b
l
e
m3
0
7
3
1
0
f
o
r
m
u
l
a
t
i
o
no
f
p
r
o
b
l
e
m
2
6
6
2
6
7
G
e
n
c
r
n
l
i
z
e
dR
e
d
u
c
e
dG
m
d
i
e
n
t(GRG)
me
df
o
rs
o
l
v
i
n
g
2
9
7
3
g
e
n
e
r
i
ca
1g
o
r
i
t
h
mf
o
r

2
6
9
2
7
0
i
n
d
i
r
c
c
tm
e
t
h
o
d
sf
o
rs
o
l
v
i
n
g

2
702
8
1
r
yc
o
n
d
i
t
i
o
n
sf
o
r

2
6
7
2
6
9
S u
e
nt
i
a
1G
r
a
d
i
e
n
tR t
o
m
t
i
o
n
a
1g
o
r
i
t
h
m(SGRA)
3
3
0
7
S u
c
n
t
i
a
JL
in
e
a
rP
r
o
gmmming
(
S
L
P
)g
o
r
i
m
2
842
8
9
S u
e
n
t
i
a
lQ
u
a
d
m
t
i
cP
r
o
g
r
a
m
m
i
n
g
(
S
Q
P
)a
1g
o
r
im
2
8
9
2
9
6
C
o
n
s
i
n
tf
u
n
c
t
i
o
n
s
6
7
t
s1

5
C
o
n
sn
393

394 I
N
O
E
X
n
(
c
o
n
r
i
n
u
e
d
)
C
o
n
s
u
a
l
i
t
y
.
7
8
i
n
e
q
u
a
l
i
t
y
.8
i
nLPp
r
o
b
1
e
m
s
.9
8
i
nN
L
P
p
r
o
b
1
e
m
s
.1
5
7
1
5
9
.
1
7
2
1
7
5
.
1
7
9
1
8
1
s
i
d
e
.
8
C
o
n
t
i
n
u
o
u
sr
e
1
a
x
a
l
i
o
n
.3
2
1
3
2
2
C
o
n
l
i
n
u
o
u
sv
n
r
i
a
b
l
c
s
.2
0
1
.
0
p
n
a
l
3
1
3
3
1
5
Con
C
o
n
v
c
r
g
i
n
g1
0t
h
cs
o
l
u
t
i
o
n
.2
1
C
o
n
v
c
xf
u
n
c
t
i
o
n
s
.3
5
2
3
5
3
C
o
r
n
e
r
s
.
2
0

lc
l
c
h
e
r
P
o
w
e
U(
D
F
P
)m
e
t
h
o
d
.
D
a
v
i
d
o
n
-F
246-249
D
c
g
r
e
e
so
ff
r
e
e
d
o
m(0
0 .
1
0
4
D
c
s
c
e
n
tf
u
n
c
l
i
o
n
.2
9
2
g
n
.
1
.
2
D i
D
c
s
i
g
nf
u
n
c
t
i
o
n
s
.5-6
e
C
l
i
v
c
s
.
5
D
c
s
i
g
n
o
D
c
s
i
g
np
a
m
m
e
l
c
r
s
.5
D
e
s
i
g
ns
p
a
c
e
.9
D
e
si
g
nv
a
r
i
a
b
1
c
s
.4-5
D
e
s
l
g
nv
e
c
l
o
r
.5
e
t
h
o
d
.s
e
eD
a
v
i
d
o
n
-F
l
e
t
c
h
e
r
P
o
w
e
l
l
DfPm

d
me
D
i
g
r
a
p
h
s
.3
3
6
D
i
s
c
r
e
t
eo
p
t
i
m
i
z
i
o
np
r
o
b
1
e
m
s(
D
P
)
.
3
1
8
3
4
8
Bmnchn
n
dBoundl
e
c
h
n
i
q
u
ef
o
rs
o
l
v
i
n
g
.
3
2
9
3
3
6
c
o
n
t
i
n
u
o
u
sr
e
1
a
x
a
t
i
o
n

a
p
p
r
o
a
c
hf
o
r
s
o
l
v
i
n
g
.
3
2
1
3
2
2
d
i
s
c
r
e
t
ed
y
n
a
m
i
cp
r
o
g
r
a
m
m
i
n
gn
s
h
n
i
q
u
ef
o
rs
o
l
v
i
n
g
.3
3
6
3
4
1
I
f
o
r
E
x
h
n
u
s
t
i
v
eE
n
u
m
e
r
a
t
i
o
nt
e
c
h
n
i
q
u
s
o
l
v
i
n
g
.3
2
6
3
2
9
n
d
c
s
i
g
n
.
3
4
1
3
4
3
I
da
p
p
r
o
a
c
ht
os
o
l
v
i
n
g
.3
2
2
3
2
4
s
l
n
n
d
Z
e
r
o
O
n
cI
n
t
e
g
c
rP
r
o
g
r
a
m
m
i
n
g(
Z
I
P
)
.
3
4
33
4
8
t
ep
r
o
g
r
a
m
m
i
n
g
.20
D
i
D
i
s
c
r
e
t
ev
a
r
i
n
b
l
c
s
.20
DOF(
d
e
g
r
e
e
so
ff
r
e
c
d
o
m
)
.1
0
4
e
eDis
c
r
e
t
co
p
t
i
m
a
t
i
o
np
r
o
b
1
c
m
s
;
D
P
.s
D
y
n
a
m
i
cp
r
o
g
r
n
m
m
i
n
g
D
u
a
lp
r
o
b
l
c
m
.1
3
8
1
4
8

I
N
O
E
X 395
P
)
.3
363
4
1
D
y
n
a
m
i
cp
r
o
g
r
a
m
m
i
n
g(D
a
l
g
o
r
i

m
3
4
0
nf
o
r
.3
3
8
3
3
9
f
u
n
c
t
i
o
n
a
l U o
EBV(
E
n
l
e
r
i
n
gB
a
s
i
cV
a
r
i
a
b
l
c
)
.1
1
8
E
d
i
t
o
r(
M
A
T
L
A
B
)
.3
3
3
7
E
n
g
i
n
e
e
I
g
.
2
)
1
1
8
E
n
t
e
r
i
n
gB
a
s
i
cV
a
r
i
a
b
1
e(EBV
d
s
e
eEx
t
c
r
i
o
rP
e
n
a
l
t
yF
u
n
c
t
i
o
n
EPFm
c
t
ho
m
e
t
h
o
d
E
q
u
a
l
i
t
yn
s
t
r
a
i
n
t
s
.
7
8

dimes
i
o
n
l
l
1
.1O
E
u
c
l
i
d
c
n
ns
p
a
c
c
s
.
E
x
h
a
u
s
t
i
v
eE
n
u
m
c
r
a
L
i
o
nt
e
c
h
n
i
q
u
c
.3
2
6
3
2
9
a
l
g
o
r
i
t
h
m
.3
2
7
a
p
p
l
i
c
a
l
i
o
n
.
3
2
7
3
2
9

d
E
x
t
e
r
i
o
rP
e
n
a
l
t
yF
u
n
c
t
i
o
n(
E
P
F
)m
e
t
h
o
2
7
1
2
7
6
a
l
g
o
r
i
t
h
m
.
2
7
2
2
7
4
l
I
o
n
.
2
7
42
7
6
a
p
p
l
i
c
l
l
F
c
a
s
i
b
1
ed
o
m
a
i
n
.2
3
F
e
a
s
i
b
1
es
o
l
u
t
i
o
n
s
.1
7
F
i
r
s
t
o
r
d
e
rc
o
d
i
t
i
o
n
s(
F
O
C
)
.1
7
6
1
7
8
o
r
d
e
r
l
l
i
n
c
a
rv
a
r
i
a
t
i
o
n
.1
6
9
F
i
r
s
t
F
la
g
p
o
1
ep
r
o
b
1
e
m
.3
0
7
3
1
0
e
eF
i
r
s
t
o
r
d
e
rc
o
n
d
i
t
i
o
n
s
FOC.s
s
e
eP
r
o
b
1
e
m
F
o
r
m
u
1
a
t
i
o
no
fp
r
o
b
1
em

8
!
t
io
n
f
o
n
n
F
O
R
T
R
A
N
.
2
4
.
2
2
7
F
u
n
c
t
i
o
n
(
s
)
:

t6
.7
c
o
n
s
t
r
a
i
n
6
o
b
j
e
c
t
i
v
c
o
f
o
n
ev
a
r
i
u
b
l
e
.1
6
2
1
6
4
.1
6
9
1
7
0
a
b
1
c
s
.
56-6
3
.
164-1
7
1
oftwov
GA.s
e
eGen
c
t
i
cA
1
g
o
r
i
t
h
m
G
u
u
s
s
J
o
r
d
a
nc
1
i
m
i
n
a
t
i
o
n
.1
0
7
G
e
n
e
r
a
l
i
z
e
dR
e
d
u
e
c
dG
m
d
i
c
n
t(GRG)

d
2
9
7
3
0
2
me
a
1
g
o
r
i
t
h
m
.2
9
9
3
0
0
0
2
u
p
p
l
i
c
a
l
i
o
n
.3
3
s
t
e
p
s
i
z
e
.c
a
1
c
u
1
a
l
i
o
no
f
.2
9
9
Ge
n
e
r
i
cG
e
n
e
t
i
cA
1
g
o
r
i
m
(GGA).
3
6
7
3
7
0
G
c
n
c
t
i
cA
1
g
o
r
i
l
h
m(
G
A
)
.3
5
0
.3
6
6
3
7
7
g
c
n
e
r
i
cf
o
n
n
.3
6
7
3
7
0
s
u
g
g
c
s
l
e
de
x
t
e
n
s
i
o
n
s1
0
.3
7
4
.3
7
5
G
1
0
b
a
lo
p
t
i
m
i
z
a
t
i
o
n
.3
5
0
3
7
7

d
e
f
i
n
i
t
i
o
no
f
p
r
o
b
l
c
m
.3
5
1
3
5
7
L
e
a
v
i
n
gB
a
s
i
cV
a
r
i
a
b
1
e(
L
B
V
)
.1
1
8
1
g
o
r
i
l
h
m(
G
A
)f
o
r e
i
n
.
G
e
n
e
t
i
cA
L
in
e
a
r
d
n
d
e
n
c
e

01
3
6
6
3
7
7
L
in
e
a
rprogmmming(
L
P
)p
r
o
b
1
c
m
s
.2
3
2
4
.
g
l
o
b
a
l
.3
5
1
3
5
3
m
i
n
i
m
um
9
3
1
5
3
5
8
n
u
m
c
r
i
c
a
lt
e
c
h
n
i
q
u
c
sf
o
r
.3
56-3
n
t
si
n
9
8
c
o
n
s
n
e
a
l
i
n
g(
S
A
)
.3
5
8
3
6
6
S
i
m
u
l
u
t
e
dAn
0
2
1
0
4
d
e
l
e
r
m
i
n
a
n
t
s.
5
6
s
o
l
u
t
i
o
n
.g
l
o
b
a
l
.3
54-3
d
wi
.
38
1
4
8
d
u
a
lp
r
o
b
1
e
ma
s
s
o
c
i
a
G
o
l
d
e
ns
e
c
t
i
o
nm
e
t
h
o
d
.2
142
1
7
.2
1
9
2
2
0
.
i
n
t
s
e
x
n
m
p
l
ei
n
v
o
l
v
l
n
g
.
c
q
u
a
l
i
t
yc
o
n
s
b
2
2
3
2
2
5
13
0-134
G
m
p
h
i
c
a
1o
p
t
i
m
i
z
a
1
6-1
8
.
4
5
9
2
c
x
.
p
l
c
so
f
.11111
1
2
4-1
3
8
.
-48
d
e
f
i
n
i
t
i
o
no
fp
r
o
b
l
e
m
.45
3
8
5
3
8
6
n
n
df
o
r
m
a
tf
o
rg
m
p
h
i
c
a
ld
i
s
p
l
a
y
.4
7
4
8
1
3
4-1
3
8
f
o
u
r
v
a
r
i
a
b
l
ep
r
o
b
1
c
m
s
f
u
n
c
t
i
o
no
ft
w
ov
a
r
i
a
b
l
e
s
.5
6
6
3
0
7
1
1
5
g
r
a
p
h
i
c
a
lS
o
l
u
l
i
o
n
so
f
.
n
n
dg
r
a
p
h
i
c
a
1u
s
e
ri
n
t
e
r
f
a
c
e
.8
1
9
1
i
n
f
m
i
t
es
o
l
u
t
i
o
no
f
.1
1
4
H
n
n
d
l
cG
n
i
p
h
i
c
s
7
9
8
1
w
i
s
o
l
u
t
i
o
nu
s
i
n
g
1
2
0
1
2
4
MATLAB
a
tt
r
a
n
s
f
e
re
x
a
m
p
1
e
.7
3
7
9
m
o
d
c
l
i
n
gi
s
s
u
c
si
n

9
8
1
0
7
s
o
l
u
t
i
o
n
. l
p
h
i
1
.
4
8
5
6
i
g
nv
a
r
i
a
b
l
c
si
n
.9
8
n
e
g
a
t
i
v
ed
s
t
r
u
c
t
u
r
a
le
n
g
i
n
e
e
r
i
n
gd
e
s
i
g
ne
x
a
m
p
1
e
.

9
3
9
4
NLPp
r
o
b
l
e
m
sv
s
.
.20
647
3
o
nf
o
r
.1
14-1
1
5
n
o
s
o
l
G
r
a
p
h
i
c
a
ls
o
l
u
t
i
o
n
s
:
p
r
i
m
p
r
o
b
l
e
m
.1
3
8
l
i
n
e
a
rp
r
o
g
r
n
m
m
i
n
g(
L
P
)p
r
o
b
1
e
m
s
.
.
1
4
8
1
5
1
s
e
n
s
i
t
i
v
i
t
yn
n
a
l
y
s
i
sw
i

0
7-1
5
S
i
m
p
l
c
xm
e
t
h
o
df
o
r
s
o
l
v
i
n
g
1
1
5
1
2
4
n
o
n
l
i
n
e
a
rp
r
o
g
r
n
m
m
i
n
g(
N
L
P
)p
r
o
b
1
e
m
s
.
s
t
l
l
n
d
a
r
df
o
r
m
a
tf
o
r
.9
4
9
8
1
7
1
1
7
5
n
p
1
e
.
t
r
a
n
s
p
o
n
a
l
i
o
np
r
o
b
1
e
me
x
o
n
e
d
i
m
e
n
s
i
o
n
a
lp
r
o
b
l
c
m
s
.2
0
5
2
0
6
1
2
4
1
3
0
G
m
p
h
i
c
a
lU
s
e
rI
n
t
e
r
f
a
c
e(
G
U
I
)
.2
4
.4
8
.49

1
1
4
u
n
b
o
u
n
d
e
ds
o
l
u
t
i
o
no
f
7
9
.
8
1
9
1
u
n
i
q
u
es
o
l
u
t
i
o
no
f
.1
1
4
GRGm
e
t
h
o

ds
e
eG
e
n
c
r
a
l
i
z
e
dR
e
d
u
c
e
d
t
v

t
o
r
si
06-1
0
7
G
m
d
i
e
n
tm
e
t
h
o
d
Mninfmmec
o
m
p
u
t
i
n
gs
y
s
l
e
m
s
.2
4
e
eG
m
p
h
i
c
l
l
1U
s
e
rI
n
t
e
r
f
a
c
c
G
U
I
.s
M
u
l
d
r
u
w
.
4
7
Mathem8
!
t
ic
a
lm
o
d
e
.5
H
a
n
d
l
eG
m
p
h
i
c
s
.7
9
8
1
Mathem
8
!
t
ic
n
lm
o
d
e
1
s
/
m
o
d
e
l
i
n
g
.3
.1
0
1
6
Jc
s
.
4
9
Hnnd
n
c
.

2
s
M
a
t
h
W
o
r
k
sI
n
s
f
e
r
.7
3
7
9
H
e
a
ta
MATLAB.2
4-44
c
o
d
es
n
i
p
p
e
t
.c
r
e
a
l
i
n
ga
.3
7-40
I
n
u
a
l
i
t
yc
o
n
s
i
n
t
s
.
8
e
d
i
l
o
ri
n
.3
3
3
7
1
7
I
n
f
i
n
i
t
es
o
l
u
l
i
o
n
s

l
l
i
m
eu
s
eo
f
.2
7
.2
9
3
3
i
g
ns
o
l
u
t
i
o
n
s
.2
1
l
n
i
t
i
a
ld
p
h
i
c
sf
u
n
c
t
i
o
n
so
f
.4
8
5
0
b
i
g
h
J
e
v
e
1g
r
l
m
i
n
g
.
2
0
J
n
l
e
g
e
rp
r
o
g
r
11
a
t
i
o
no
f
.2
6
2
8
i
n
s
la
l
I
t
e
g
e
rv
a
r
i
a
b
1
c
s
.2
0
I
o
p
e
r
u

r
si
n
.3
0
3
1
l
l
c
r
a
t
i
v
et
e
c
b
n
i
q
u
c
s
.2
1
O
p
t
i
m
i
z
a
t
i
o
nT

1boxo
f
.1
2
3
1
2
4
.
3
7
9
3
9
1
K
u
h
n
Tuc
k
e
rc
o
n
d
i
t
i
o
n
s
.1
9
2
1
9
3
.
a
t
i
n
g

a40-44
p
r
o
g
r
n
m
.
c
62
0
0

2
6
7
2
6
9
1
9
S
i
m
p
1
e
xM
e
t
h
o
du
s
i
n
g
.1
2
0
1
2
4

1boxo
f
.1
5
9
1
6
1
S
y
m
b
o
l
i
cM
a
t
hT
L
a
g
r
n
n
g
em
u
1
t
i
p
l
i
e
r
s
.1
8
0
1
9
2

2
6
u
s
e
f
u
1
n
e
s
so
f
LAMBDAprogmm(
O
p
Z8
t
i
o
n
T
I
1
b
o
x
)
.
M
a
t
r
i
c

:
3
8
3
3
8
4
a
u
g
m
e
n
l

.
1
0
3
1
0
4

INDEX

397

396 INDEX
M
B
I
I
I

s(
c
o
n
t
i
n
u
e
.

r
a
n
ko
f
1
04-1
0
6
i
mump
r
o
b
l
e
m
s
6
Mu
i
1

2
7
M-f
M
i
n
i
m
i
z
a
t
i
o
np
r
o
b
l
e
m
s
6
d

2
5
2
2
5
3
M
o
d
i
f
i
e
dN
e
w
t
o
nme
M
u
1
t
i
d
i
s
c
i
p
l
i
n
a
r
yd
e
s
i
g
n
s
6
M
u
1
t
i
o
b
j
c
c
t
i
v
ed

i
g
n
s
6
n
d
i
m
e
n
s
i
o
n
a
lE
u
c
l
i
d
e
n
ns
p
a
:

51
0
6
2
0
6
2
0
9

2
5
2
N
e
w
t
o
n
R
a
p
h
s
o
nt
e
c
h
n
i
q
u
e
)p
r
o
b
1
c
m
s
N
o
n
1
i
n
c
a
rp
r
o
g
r
a
m
m
i
n
g(NLP
2
1
2
31
5
4
2
'

W
1
a
1y
t
i

1c
o
n
d
i
t
i
o
n
si
n1
7
5
1
7
6
a
i
n
t
si
n
1
5
7
1
5
91
7
2
1
7
5
c
o
n
s
1
7
9
1
9
1
942
e
x
a
m
p
1
e
so
f
.1
f
i
r
s
l
o
r
d
c
rn
d
i
t
i
o
n
si
n1
7
6
1
7
8
f
o
r
m
u
1
a
t
i
o
no
f
p
r
o
b
l
e
m
1
5
5
1
5
7
g
r
a
p
h
ic
a
1s
o
l
u
t
i
o
n
so
f
1
7
1
1
7
5
Tu
c
k
c
rc
o
n
d
i
t
i
o
n
sf
o
r
1
9
2
1
9
3
K
u
h
n
1
9
6-2
L
a
g
r
a
n
g
cm
u
l
t
i
p
l
i
e
rm
c
t
h
o
d
u
s
eo
f

1
80-1
9
2
09
3
9
4
LPp
r
o
b
1
e
m
sVS.2
o
n
ev
a
r
i
a
b
l
e

f
u
n
c
t
i
o
n
so
f
1
6
2
1
6
4
1
6
9
1
7
0

0
e
r
c
o
n
d
i
t
i
o
n
si
n1
7
8
1
7
9
mnd
5
9
1
6
1
s
y
m
b
o
l
i
cc
o
m
p
u
t
o
t
i
o
n1
6
9
1
7
1
T
a
y
1
0
rs
c
r
i
e
sf
o
ru
s
ei
ns
o
l
v
i
n
g1
v
a
r
i
l
l
b
l
c
s
f
u
n
c
t
i
o
n
so
f

twoormo
1
64-1
7
1

t
i
m
a
ls
o
l
u
1
7
l
-4
O
p
t
i
m
i
z
a
t
i
o
n
O
p
t
i
m
i
z
a
t
i
o
nT
o
o
l
b
o
x1
2
3
1
2
4
3
7
9
3
9
1
c
o
n
s
t
r
a
i
n
c
do
p
t
i
m
i
l
.
a
t
i
o
nu
s
i
n
g
3
8
9
3
9
1
e
x
a
m
p
l
e
su
s
i
n
g
3
8
5
3
8
6
1
i
n
e
a
rprogmmminge
x
a
m
p
l
cu
s
i
n
g
3
8
5
3
8
6
s
e
t
t
i
n
g
3
8
4
3
8
5
o
p
t
i
m
i
z
a
t
i
o
np
a
r
a
m
e
t
e
r
s
i
cp
r
o
g
r
u
n
i
n
gc
x
a
m
p
1
eu
s
i
n
g
q
u
a
3
863
8
8
i
n
e
do
p
t
i
m
i
z
a
t
i
o
nu
s
i
n
g
u
n
c
o
n
s
t
r
3
8
8
3
8
9
3
8
2
3
8
4
u
s
i
n
gprogmmsi
n
OUTPUTp
r
o
g
r
a
m(
O
p
t
i
m
i
z
a
t
i
o
nT
o
o
1
b
o
x
)
3
8
3

d
i
g

n5
P
a
r
a
m
e
1

P
a
r
t
i
a
ld
e
r
i
v
a
t
i
v
2
4
P
a
s
c
n
l
P
n
t
l
e
mS
c
a
r
c
hm
e
t
b
o
d
2
3
4
2
3
8
l
1
up
p
r
o
x
i
m
a
t
i
o
n
2
1
1
2
1
4
2
1
7
P
o
1
y
n
o
P
o
w
e
l
l
'
sm
c
t
b
o
d
2
3
8
2
4
0
a
1g
o
r
i
m
2
3
8
2
3
9
a
p
p
l
i
c
a
t
i
o
n
2
3
9
2
4
0
a
n
dc
o
n
j
u
g
a
c
y
2
4
0
P
r
i
m
a
lp
r
o
b
l
e
m1
3
8
P
r
o
b
l
c
md
e
f
i
o
n
a
n
d
f
i
r
m
u
1
a
t
i
o
n

4-1
0
_
S
e
ea
l
s
oM
a
t
b
e
m
a
t
i
c
a
lm
o
d
e
l
s
J
m
o
d
e
l
i
n
g
S n
e
do
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
e
m
s
c
o
2
6
6
2
6
7
n
s
t
r
n
tf
u
n
c
t
i
o
n
si
n7
d
e
s
i
g
nf
u
n
c
t
i
o
n
si
n
5-6
p
a
r
n
e
t
e
r
si
n
5
di
O
b
j
t
i
v
cf
u
n
c
t
i
o
n
s6
9
a
n
dd
e
s
i
g
ns
p
a

0c
c
to
r
i
e
n
t
c
dp
r
o
g
r
a
m g

2
4
d
e
s
i
g
nv
n
r
i
a
b
1
e
si
n
4
5
di
m
c
n
s
i
o
n
a
lp
r
o
b
l
e
m
s
2
0
3
2
2
5
One i
t
yn
s
t
r
i
7
-8

2
0
9
2
1
0
b
i
s
c
c
t
i
o
nt
c
c
b
n
i
q
u
ef
o
ru
s
ew
i
.
a
i
t
n

3
5
1
3
5
7
g
l
o
b
a
lo
p
t
i
m
i
l
n
e
d
2
0
42
0
5
c
o
n
s
t
r
5-4
8
g
r
u
p
h
i
c
u
1o
p
t
i
m
i
z
a
t
i
o
n4
d
c
f
i
n
i
t
i
o
no
fp
r
o
b
l
e
m
2
0
4
8
i
n
e
q
u
a
l
i
t
yc
o
n
s
t
r
a
i
n
t
si
n
g
o
l
d
e
ns
c
c
t
i
o
nm
c
t
h
o
df
o
ru
s
ew
i
t
h

.2
3
n
o
n
l
i
n
e
a
rp
r
o
b
l
e
m
s
2
1.
2
142
1
7
2
1
9
2
2
0

2
2
3
2
2
5
Ii
n
r
p
r
o
g
r
a
m
m
i
n
g(
N
L
P
)p
r
o
b
1
e
m
s
g M1s
o
l
u
t
i
o
no
f

2
0
5
2
0
6
1
5
5
1
5
9

2
1
7
2
1
9
i
m
p
o
r
t
a
n
c
eo
f
o
b
j
c
c
u
v
ef
u
n
c
t
i
o
ni
n
6
q
u
ef
o
ru
s
ew
i
t
h
N
e
w
t
o
n
R
a
p
b
s
o
nt
c
c
o
n
e
d
i
m
c
n
s
i
o
n
a
lp
r
o
b
1
e
m
2
0
4
2
0
5
2
0
6
2
0
9
s
a
i
n
t
si
n
8
s
i
d
en
p
o
l
y
n
o
m
i
a
1a
p
p
r
o
x
i
m
B
t
i
o
nw
i
t
h
2
1
1
2
1
4
s
t
a
n
d
a
r
df
o
r
m
a
tf
o
r

9
1
0
t
w
o
p
o
i
n
tb
o
u
n
d
a
r
yv
a
l
u
ep
r
o
b
l
e
m
u
n
c
o
n
s
t
m
i
n
c
do
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
e
m
s
2
2
02
2
3
2
2
7
2
3
0
O
p
e
r
a
t
o
r
s
3
0
-3
1

P
r
o
b
1
e
mr
c
l
a
x
a
t
i
o

n3
2
1
3
2
2
P
r
o
b
l
e
ms
o
l
u
t
i
o
n
s1
6-2
4
P
r
o
g
r
a
m
m
i
n
g
d
i
s
c
r
c
t
e
20
g
r
a
m
m
i
n
gi
n
t
c
g
2
0
2
4
P
r
o
g
r
a
m
m
i
n
gl
a
n
g
u
a
g

2
4
P
r
o
g
r
a
m
m
i
n
gp
a
m
d
i
g
m
s
2
3
0
P
s
e
u
d
o
r
n
n
d
o
mn
u
m
b
e
r
s

Q
u
a
d
r
a
t
i
cc
o
n
v
e
r
g
e
n
c
e
2
3
8
a
t
i
cp
o
1
y
n
o
m
i
a
l
s
2
3
8
Qua
a
t
i
cp
r
o
g
r
n
g
3
8
6
3
8
8
Qua
2
3
0
2
3
3
2
5
4
2
5
5
RundomWalkm
e
t
h
o
d
R
e
d
u
n
d
a
n
c
y1
0
1

m2
5
3
2
5
5
R
o
s
e
n
b
r
o
c
kp
r
o
b
l
e
I
gd
i
s
k
t
h
r
c
e
d
i
m
e
n
s
i
o
n
a
l1
1
0
wn
e
a
r

R
o
t
a
2
5
5
2
5
8
SA
s
e
eS
i
m
u
l
a
t
e
dAnn
e
n
l
i
n
g
i
I
e
s
2
7
S
c
r
i
p
tm-f
h
a
1g
o
r
im
2
1
8
S
S
e
a
r
c
hd
i
r
c
c
t
i
o
n
2
2
o
d
s
2
1
S
e
a
r
c
hme
o
r
d
e
rc
o
n
d
i
t
i
o
n
s(
S
O
C
)1
7
8
1
7
9
S
e
c
o
n
d
6
9
S
c

I
I
d
u
r
d
e
r
l
l
l
n
e
a
rv
a
r
i
a
t
i
o
n1
y
s
i
s1
4
8
1
5
1
S
e
n
s
i
t
i
v
i
t
ya
n
S
e
q
u
e
n
t
i
a
ld
c
c
i
s
i
o
np
r
o
b
l
e
m
s

3
3
6
1g
o
r
i
t
b
m
S
e
q
u
e
n
t
i
n
lG
r
a
d
i
e
n
tR
e
s
t
o
r
a
t
i
o
na
(SGRA)
3
0
2
3
0
7
n
l
g
o
r
i
t
b
m

3
0
5
n

3
0
5
3

7
a
p
p
l
i
3
0
3
3
0
4
g
r
a
d
i
e
n
tp
h
n
s
e
r
e
s
t
o
r
a
t
i
o
np
h
a
s
e

3
0
4
3
0
5
S
e
q
u
e
n
t
i
a
lL
i
n
e
a
rP
r
o
g
r
n
g(
S
L
P
)
2
8
4
2
8
9
g
o
r
im
2
8
5
2
8
6

2
8
62
8
9
a
p
p
l
i
c
a
t
i
o
n
a
t
l
cP
r
o
gmmming(
S
Q
P
)
S
e
q
u
e
n
t
i
a
lQua
2
8
9
2
9
6
2
9
4
a
1
g
o
r
i
t
h
m
2
9
5
2
9
6
a
p
p
l
i
c
a
t
i
o
n

d2
9
3
2
9
4
a
n
dBFGSm
e
t
h
o
c
a
l
c
u
l
a
t
i
o
no
f

2
9
3
m
u
l
t
i
p
l
i
e
r
s
s
t
e
p
s
i
z
e
c
a
1
c
u
l
a
t
i
o
no
f

2
9
2
2
9
3
S
e
q
u
e
n
t
i
a
lU
n
c
o
n
s
t
r
n
i
n
c
dM
i
n
i
m
i
z
a
t
i
o
n

2
7
0
2
7
1
T
c
c
h
n
i
q
u
e
s(SUM'
s
e
eS
e
q
u
e
n
t
l
u
lO
m
d
i
c
n
tR
e
s
t
o
r
a
t
i
o
n
SGRA
a
Jg
o
r
i
t
b
m
8
S
i
d
ec
o
n
s
t
r
a
i
n
t
s

1
5
1
2
4
S
i
m
p
l
c
xm
e
t
h
o
d1
1
1
71
2
0
a
p
p
l
i
c
a
t
i
o
no
f
E
BV)
1
1
8
E
n
t
e
r
i
n
gB
a
s
i
cV
a
r
i
a
b
l
e(
f
e
a
t
u
r
e
so
f
1
1
5
1
1
7
L
e
a
v
i
n
gB
a
s
i
cV
a
r
i
u
b
1
c(LBV)
1
1
8
MA
T
I
A
B
s
o
l
u
t
i
o
nu
s
i
n
g1
2
0
1
2
4
tw
o
-p
h
a
s
e
125-130
e
a
l
i
n
g(
S
A
)

3
5
0

3
5
8
3
6
6
S
i
m
u
l
l
l
t
e
dAnn
a
1t
e
m
a
t
i
v
ca
1g
o
r
i
m
3
5
9
3
6
5
b
a
s
i
ca
l
g
o
r
i
t
h
mo
f

3
5
8
3
5
9
n
s
t
r
a
i
n
t
s
h
a
n
d
1
i
n
g
3
6
4
6
5
3
6
6
i
nd
l
s
c
r
c
t
ep
r
o
b
l
c
m
s3
S
l
a
c
kv
a
r
i
a
b
1
e
s
9
5

s
e
eS
e
q
u
e
n
t
i
a
lL
in
c
a
r'Og
r
a
m
m
i
n
g
SLP
s
e
eS
c
c
o
n
d

o
r
d
c
rc
o
n
d
i
t
i
o
n
s
SOC
S
o
f
t
w
2
4
62
4
S
o
l
u
t
i
o
n
s1
7
a
c
c
e
p
t
o
b
l
c1
7
i
n
f
i
n
i
t
e1
2
1
i
n
i
t
i
a
ld
e
s
i
g
n

1
7
o
p
t
i
m
SQP

s
e
eS
e
q
u
e
n
t
i
a
lQ
u
a
d
r
n
t
i
cP
r
o
g
r
a
m
m
i
n
g
9
1
0
S
t
n
n
d
a
r
df
o
r
m
a
l
d

2
4
1
2
4
4
S
t
e
c
p
e
s
tDcentme

J
e
eS
c
q
u
e
n
t
i
U
n
c
o
n
s
t
r
a
i
n
e
d
SUMT
M
i
n
i
m
i
z
a
t
i
o
nT
c
c
h
n
i
q
u
e
s
S
y
m
b
o
1
i
cMathT

o
l
b
o
x(
M
A
T
L
A
B
)
1
5
9
1
6
1
T
a
y
l
o
rs
c
r
i
c

1
6
9
1
7
1
a
n
c
e
2
0
9
T
o
l
TPBVP

s
e
cTw
o
-p
o
i
n
tb
o
u
n
d
a
r
yy
a
l
u
e
p
r
o
b
l
e
m
np
r
o
b
l
c
me
x
a
m
p
l
e
1
2
4-1
3
0
T
r
a
n
s
p
o
i
n
tb
o
u
n
d
a
r
yv
a
1
u
ep
r
o
b
l
c
m
Twc
'PBVP)220-223
U
n
c
o
n
s
t
r
a
i
n
e
do
p
l
i
m
i
z
l
l
t
i
o
np
r
o
b
l
e
m
s
2
2
7
2
6

23
8
8
3
8
9
B
c
z
i
c
rp
a
r
a
m
c
t
r
i
c
c
u
r
v
e
s
f
i
t
t
l
n
g
2
5
8
2
6
2
1
e
t
c
h
e
r
G
o
1
d
f
a
r
l
:
S
h
a
n
n
o
B
r
o
y
d
o
n
-F
m
e
t
h
o
df
o
rs
o
l
v
i
n
g
2
4
9
2
5
1
u
g
a
t
eG
r
a
d
i
e
n
tm
c
t
b
o
df
o
rs
o
l
v
i
n
g
Co
44-2
4
6
2
1e
t
c
h
c
r
P
o
w
c
l
lm
e
l
h
o
df
o
r
D
a
v
i
d
o
n
-F
s
o
l
v
i
n
g
24249

2
2
7
2
3
0
d
e
f
i
n
i
t
i
o
no
fp
r
o
b
1
e
m
1g
o
r
i
t
b
mf
o
r
2
2
9
2
3
0
g
e
n
e
r
i
ca

396

I
N
D
E
X

I
N
D
E
X

(
c
o
n
t
i
n
u
e
d
)
M
a
l
r
i
rankof
l
04-106
M ump
r
o
b
l
c
m
s
.6
27
M
f
i
l

6
M
i
n
i
m
i
z
a
t
i
o
np
r
o
b
l
c
m
s

252-253
ModifiNewtonmethod
M
u
l
l
i
d
i
s
c
i
p
l
i
n
n
r
yd
e
s
i
g
n
s
6
M
u
l
t
i
o
b
j
e
c
l
i
v
ed
e
s
i
g
n
s

n
d
i
m
e
n
s
i
o
n
a
lE
u
c
l
i
d
e
a
ns
p
a
c
e
s
1
0
6
New
lO
n
R
a
p
b
s
o
t
e
c
h
q
uc
2
062
252
N
o
n
l
i
n
e
a
rprogramming(NLP)problem~
21-231
5
4200
a
n
a
l
y
t
i
c
a
lc
o
n
d
I
t
i
o
n
si
n175-176
c
o
n
s
t
r
l
l
I
n
t
si
n
157-159
172-175
1
7
9
1
9
1
194-200
cxampl
1
-0e
rc
o
n
d
i
l
i
o
n
si
n1
76
-1
7
8
m
u
l
a
t
i
o
no
fp
r
o
b
l
e
r
n
155-157
171-175
g
r
a
p
h
i
c
a
l
s
o
l
u
t
i
o
n
so
f
Kuhn

Tuckerc
o
n
d
i
t
i
o
n
sf
o
r
192-193
1
9
2
L
a
gmngem
u
l
l
I
p
l
i
e
rmelhod
u
s
eo
f

180-192
20

93-94
LPp
r
o
b
l
e
m
sv
s
.
o
n
ev
a
r
i
n
b
l
e
.f
u
n
c
l
I
o
n
so
f
.1
6
2
-1
6
4
.
169-170
s
e
c
o
n
d
o
r
d
e
rc
o
n
d
I
l
i
o
n
si
n
178-179
s
y
r
n
b
o
l
i
ccomputa
on
159-161
1
y
l
o
r
s
e
r
i
f
o
ru
s
ei
ns
o
l
v
i
n
g
.169-171
Tl

f
u
n
c
l
i
o
n
so
f

twoo
rmorov
a
r
I
a
b
l
e
s
1
64-1
7
1

O
b
j
e
c
t
i
v
ef
u
n
c
t
i
o
n
s
6
0
loricn programming24
203-225
O
n
e
d
I
m
e
n
s
I
o
n
a
lp
r
o
b
l
c
m
s
b
i
s
c
c
t
i
o
nt
e
c
h
n
i
q
u
ef
o
ru
s
ew
I
t
h

209-210
n
ed
204205
C
O
I

204
d
e
f
i
n
i
t
i
o
no
fp
r
o
b
l
c
m
g
o
l
d
s
c
c
t
i
o
nmethodf
o
ru
s
ew
i
t
h
.
217
219-220

223-225
2
14g
r
n
p
h
i
c
a
ls
o
l
u
t
i
o
no
f

2
0
5-206
i
m
p
o
r
t
a
n
c
eo
f

217-219
Ncw
lO
n
Ra
p
h
s
o
nl
e
c
h
n
i
q
u
ef
o
ru
s
ew
i
t
h
206-209
p
o
l
y
n
o
m
i
a
la
p
p
r
o
x
i
m
a
t
i
o
nW
i
l
h
211-214
t
w
o
p
o
I
n
tb
o
u
n
d
n
r
yv
a
l
u
cp
r
o
b
l
e
m
220-223
Ope
mlo
3
0-31

7
O
p
t
i
m
a
ls
o
l
u
t
i
o
n
s1
i
o
n
l
-4
Opt
i
m
i
z
a
t
i
o
nT
lbox
123-124
379-391

c
o
n
s
l
m
I
n
c
do
p
t
I
m
i
z
n
l
i
o
nu
s
i
n
g

389-391
e
x
a
m
p
l
e
su
s
i
n
g
385-386
n
p
l
eu
s
I
n
g
l
i
n
e
a
rp
r
o
g
r
a
r
m
n
i
n
gex
385-386
me
r
s
s
e
u
i
n
g
384

3
85
o
p
t
i
m
i
z
a
l
i
o
np
qua
.
t
i
cprogrammingc
x
a
m
p
l
eu
s
i
n
g
386-388
i
n
e
do
p
t
i
m
i
z
a
t
i
o
nu
s
E

un
388-389
u
s
i
n
gp
r
o
g
r
a
m
sI
n
382-384
0rpUTprogmm(
O
p
t
i
m
i
z
a
t
i
o
nT

lbox).
383
d
e
s
i
g
n

5
P
a
r
a
m
e
t
e
r
s
164
P
a
n
i
a
ld
e
r
i
v
a
t
i
v
e
s
P
a
s
c
a
l
24
r
c
hmcthod
234-238
PattcmS
P
o
l
y
n
o
m
i
a
lo
p
p
r
o
x
i
m
a
t
i
o
n
2
1
1
2
1
4
.217
Powe
l
l
'sme
o
d

238-240
a
1
g
o
r
i
m
238-239
a
p
p
1 n.239-240

240
a
n
dc
o
n
j
u
g
a
c
y
P
r
i
m
a
lp
r
o
b
l
e
m
138
P
r
o
blemd
e
f
i
n
i
t
i
o
na
n
df
o
r
m
u
1a
t
i
o
n
4-1
0
.
S
e
ea
l
s
oMalhematicalmode1s
1
m
o
d
e
l
i
n
g
do
p
l
i
m
i
z
a
t
i
o
Dp
r
o
b
l
e
m
s
c
o
n
s
266-267
nsuaintfunctionsi
n
7
d
e
s
i
g
nf
u
n
c
t
i
o
n
si
n
5-6
d
e
s
i
g
np
a
r
a
m
e
t
e
i
n

5
9
andd
e
s
i
g
ns
p
a
c
e
45
d
e
s
i
g
nv
n
r
i
a
b
l
c
si
n
e
q
u
a
l
i
t
yc
o
n
s
u
a
i
n
t
si
n

7-8
p
t
a
t
i
o
n

351-357
g
l
o
b
a
l
g
r
a
p
h
i
c
a
lo
p
l
i
m
i
z
a
t
i
o
n
45
-48
1
it
yn
s
u
a
i
n
t
si
n
8
i
n
e
q
ua
n
o
n
l
i
n
e
a
rp
r
o
b
l
e
m
s
2
1 23
n
o
n
l
i
n
e
a
rprogramming(NLP)p
r
o
b
l
e
m
s

155-159
t
i
v
ef
u
n
c
t
i
o
Di
n
6
0
o
n
e
d
i
m
e
n
s
i
o
n
a
lp
r
o
b
l
e
m

204-205
8
s
i
d
ec
o
n
s
t
m
i
n
t
si
n

9-10
s
t
a
n
d
o
r
df
o
r
m
n
tf
o
r
i
n
c
do
p
t
i
m
i
z
a
t
i
o
np
r
o
b
l
e
m
s
u
n
c
o
n
s
227-230

P
r
o
b
l
e
m
r
e
l
a
t
i
o
n

321-322
Problems
o
l
u
t
i
o
n
s
1
6
24
Progmmming
d
i
s
c
t
e

20
P
r
o
gramming
i
n
t
e
g
c
r

20
P
r
o
gn
u
n
o
Unglanguages
24
P
r
o
gn
u
n
o
Ungparadigms
24
230
Pseudorandomn
u
m
b
e
r
s
Qua
d
r
a
t
i
cc
o
n
v
e
r
g
e
n

238
Q
u
a
d
r
a
t
i
cp
o
l
y
n
o
m
i
a
l
s
238
Quadmticprogramming
386-388
2
3
0-233
254-255
RandomWa1kmethod
n
d
a
n
c
y1
0
1
R u

253-255
R
0
5
c
n
b
r
o
c
kp
r
o
b
l
e
m
R
o
t
a
t
i
n
gd
i
s
k

t
h
r
c
d
i
m
e
n
s
i
o
n
a
lf
l
o
wn
e
a
r

5-258
SA

.
r
t
tS
i
m
u
l
a
t
e
dA
n
n
e
a
l
i
n
g
S
c
r
i
p
tm
f
i
l
c
s
27
1
g
o
r
i
t
h

m 218
S
e
a
r
c
ha
S hdi
t
i
o
n
22
c
hm
e
t
h
o
d
s
2
1
S r
Secondo
r
d
c
rc
o
n
d
i
t
i
o
n
s(SOC)
178-179
Secondo
r
r
/
l
i
n
e
a
rv
a
r
i
n
t
i
o
n1
6
9
S
e
n
s
i
t
i
v
i
l
y
a
n
n
l
y
s
i
s
148-151
e
n
t
i
a
ld i
S
I
O
Dp
r
o
b
l
e
m
s
336
S u
r
a
d
i
e
n
tR
e
s
t
o
r
a
t
i
o
na
1
g
o
r
i
t
h
m
S
e
q
u
c
n
t
i
a
lG
(SGRA)
302-307
a
1
g
o
r
i
m
305
305-307
a
p
p
l
i
c
a
t
i
o
n

303-304
g
r
a
d
i
e
n
tphe
04305
r
e
s
t
o
r
a
t
i
O
Dp
h
a
s
e
.3
S
e
q
u
e
n
t
i
a
lL
i
n
e
a
rProgramming(
S
L
P
)
2
84289

285-286
a
l
g
o
r
i
t
h
m
a
p
p
1
i
c
a
t
i
o
n
286-289
Sq
u
e
n
t
i
a
lQu
a
d
r
a
t
i
cP
r
o
gramming(SQP)
289-296
a
l
g
o
r
im.294
295-296
a
p
p
l
i
c
n
t
i
o
n
o
d
293-294
andBFGSme
m
u
1
t
i
p
l
i
e
r
s

I
i
onof

293

c
a
l
c
u
l
a
t
i
o
no
f

292-293
s
t
e
p
s
i
z
e
I
m
i
7
.
ot
i
o
n
S
e
q
u
e
n
t
i
a
lU
n
c
o
n
s
u
a
i
n
c
dI
M
T
>
270-271
T
e
c
h
n
i
q
u
e
s(
SU
soeS
c
q
u
c
n
t
ia
JG
r
a
d
i
e
n
lR
e
s
[
o
r
a
t
i
o
n
SGRA
n
1g
r
i
t
h
m
s
n
t
s
8
S
i
d
cn

397

115-124
S
i
m
p
l
e
xmethod
p
l
i
c
a
t
i
o
no
f
117-120
1
1
8
E
n
t
e
r
i
n
gB
a
s
i
cV
a
r
i
a
b
l
e(EBV)
f
e
a
t
u
r
e
so
f

115-117
Le
a
v
i
n
g8
a
s
i
cV
a
r
i
a
b
l
e(LBV)
1
1
8
MA
TlA
B
s
o
l
u
l
i
o
nu
s
i
n
g120-124
-phasc125-130
t
wo
e
dAnn
e
a
l
i
n
g(SA)

350

358-366
S
i
m
u
l
a
1t
e
m
a
t
i
v
ea
l
g
o
r
i
t
h
m

359-365
1
g
o
r
i
t
h
mof

358-359
b
a
s
i
ca
n5Ui
n

h
l
d
J
i
n
g364
365-366
i
nd
i
s
c
r
e
t
ep
r
o
b
l
e
m
s
S
l
a
c
kv
a
r
i
a
b
l
e
s
95

s
e
eS
e
q
u
e
n
l
i
a
lL
in
e
a
rProgramming
SLP
SOC

s
l
!
eSccond-orderconditions
So
w
a
r
e

24
S
o
l
u
t
i
o
n
s
16-24
a
eptab1eI7

1
7
i
n
f
i
n
i
t
e
2
1
i
n
i
l
i
a
ld
e
s
i
g
n
o
p
t
i
m
a
l1
7

s
e
eS
e
q
u
e
n
t
i
a
lQ
u
u
d
r
n
t
i
cP
r
o
g
r
a
r
n
m
i
n
g
SQP
9-10
S
t
a
n
d
a
r
df
o
n
n
n
t
stO

ntmethod
241-244
S
t
SUMT

s
e
e
S
u
e
n
t
i
a
lU
n
c
o
n
s
t
m
i
n
e
d
M
l
n
I
m
i
z
a
t
i
o
nT
e
c
h
n
i
q
u
e
s
l
icMathT

lbox(MATlAB)
Sym
1
5
9
1
6
1
6
9
1
7
1
T
a
y
l
o
rs
c
r
i
e
s1
T
o
1
e
r
u
n
c
e
209
TPBVP

s
Tw
o

p
o
i
n
tboundyv
a
1
u
e
p
r
o
b
l
e
m
i
o
np
r
o
b
1
e
mc
x
p
1
e
1
2
4-1
3
0
Transpon
Twc
p
o
i
n
tb
o
u
n
d
n
r
yv
a
l
u
ep
r
o
b
l
e
m
'
P
BVP)
220-223

n5Uinedoptimizationproblems
Un
227-262
388-389

f
i
t
t
i
n
g
8
e
z
i
c
rp
n
r
a
m
e
t
r
i
cc
u
258-262
S
hanno
8
r
o
y
d
o
n
F
l
e
t
c
h
e
r
G
o
l
d
f
a
r
t
249-251
methodf
o
r5
0
l
v
i
n
g
C
o
n
j
u
g
a
t
eG
r
a
d
i
e
n
tmelhodf
o
r8
0
1
v
i
n
g
244246
O
a
v
i
d
o
n
F
1
c
l
c
h
e
r
P
o
w
cl
Imelhodf
o
r
s
o
l
v
i
n
g
24249
l
D
It
i
o
no
fp
r
o
b
l
e
m

227-230
def

229-230
g
e
n
e
r
i
ca
1
g
o
r
i
t
h
mf
o
r

398

INDEX

U
n
c
o
r
u
l
l
r
a
i
n
e
do
pl
i
m
i
z
.
i
o
np
r
o
b
l
e
m
s
(
c
o
n
e

gd
i
e
n
t
b
a
s
e
dl
e
t
h
o
d
sf
o
rs
o
l
v
i
n
g

241-251

m
o
d
i
f
i
e
dNe
WlO
nm
e
t
h
o
df
o
rs
o
l
v
i
n
g
252-253

n
e
c
e
s
r
yI
lDds
u
f
f
i
c
i
e
n
tc
o
n
d
i
t
i
o
n
sf
o
r

228-229

n
o
n
g
r
a
d
i
e
n
tm
e
l
h
o
d
sf
o
rs
o
l
v
i
n
g
230

240

P
a
t
t
e
mS
hm
e
t
h
o
df
o
rs
o
l
v
i
n
g
23
4-238
P
o
w
c
U
'
sm
e
l
h
o
df
o
rs
o
l
v
i
n
g
238-240
randomwa
1
ka
1
g
o
r
i
l
h
mf
o
r

23
02
3
3
R
o
s
e
n
b
r
o
c
kp
r
o
b
l
e
m
253-255
r
o
t
a
t
i
n
gd
i
s
k
t
h
r
c
e
d
i
m
e
n
s
i
o
n
a
lsow

255S

s
e
c
o
n
d
o
r
d
e
rm
e
l
h
o
d
sf
o
rs
o
l
v
i
n
g

2
5
13

S
l
e
e
p
c
s
tD
c
s
c
e
n
tm
c
l
h
o
df
o
rs
o
l
v
i
n
g
241-244

U
n
i
q
u
eo
p
t
i
r
n
a
1s
o
l
u
t
i
o
n
s
.1
7
Unitv
o
r
s
1
06-1
0
7
246
V
a
r
i
l
l
b
l
eM
e
t
r
i
cM
e
t
h
o
d
s(VMM)
V
a
r
i
n
b
l
e
s
:
n
r
t
i
f
i
c
i

l
H 98
c
o
n
t
i
n
u
o
u
s
20
d
e
s
i
45

d
i
s
c
r
e
l

e20
i
n
t
e
g
2
0
s
l
a
c
k

95
V
e
c

r
(
s
)
:
d
e
s
i
g
n

5
u
n
i

tl
O

1
0
7
VMM(
V
a
r
i
a
b
l
eM
e
t
r
i
cM
e
l
h
o
d
s
)
246
Z
e
r
o
-Oneln
Pro
g
r
a
m
r
n
i
n
g(
Z
I
P
)
343-348
Z
e
r
o
o
r
d
e
rm
e
t
h
o
d
s
230
Z
I
P

s
e
eZero-Onel
n
l
e
g
e
rP
r
o
gramming

Das könnte Ihnen auch gefallen