Sie sind auf Seite 1von 86

6

Matrix Analysis
Exercises 6.3.3
1(a) Yes, as the three vectors are linearly independent and span three-
dimensional space.
1(b) No, since they are linearly dependent
_
_
3
2
5
_
_
2
_
_
1
0
1
_
_
=
_
_
1
2
3
_
_
1(c) No, do not span three-dimensional space. Note they are also linearly
dependent.
2 Transformation matrix is
A =
1

2
_
_
1 1 0
1 1 0
0 0

2
_
_
_
_
1 0 0
0 1 0
0 0 1
_
_
=
_
_
1

2
1

2
0
1

2

1

2
0
0 0 1
_
_
Rotates the (e
1
, e
2
) plane through /4 radians about the e
3
axis.
3 By checking axioms (a)-(h) on p. 427 it is readily shown that all cubics
ax
3
+ bx
2
+ cx + d form a vector space. Note that the space is four dimensional.
3(a) All cubics can be written in the form
ax
3
+ bx
2
+ cx + d
and 1, x, x
2
, x
3
are a linearly independent set spanning four-dimensional space.
Thus it is an appropriate basis.
c _
Pearson Education Limited 2004
Glyn James: Advanced Modern Engineering Mathematics, Third edition 289
3(b) No, does not span the required four-dimensional space. Thus a general
cubic cannot be written as a linear combination of
(1 x), (1 + x), (1 x
3
), (1 + x
3
)
as no term in x
2
is present.
3(c) Yes as linearly independent set spanning the four-dimensional space
a(1 x) + b(1 + x) + c(x
2
x
3
) + d(x
2
+ x
3
)
= (a + b) + (b a)x + (c + a)x
2
+ (d c)x
3
+ x + x
2
+ x
3
3(d) Yes as a linear independent set spanning the four-dimensional space
a(x x
2
) + b(x + x
2
) + c(1 x
3
) + d(1 + x
3
)
= (a + b) + (b a)x + (c + d)x
2
+ (d c)x
3
+ x + x
2
+ x
3
3(e) No not linearly independent set as
(4x
3
+ 1) = (3x
2
+ 4x
3
) (3x
2
+ 2x) + (1 + 2x)
4 x + 2x
3
, 2x 3x
5
, x + x
3
form a linearly independent set and form a basis
for all polynomials of the form + x
3
+ x
5
. Thus S is the space of all odd
quadratic polynomials. It has dimension 3.
c _
Pearson Education Limited 2004
290 Glyn James: Advanced Modern Engineering Mathematics, Third edition
Exercises 6.4.3
5(a) Characteristic polynomial is
3
p
1

2
p
2
p
3
with
p
1
= trace A = 12
B
1
= A12I =
_
_
9 2 1
4 7 1
2 3 8
_
_
A
2
= A B
1
=
_
_
17 5 7
18 30 7
2 5 33
_
_
p
2
=
1
2
trace A
2
= 40
B
2
= A
2
+ 40I =
_
_
23 5 7
18 10 7
2 5 7
_
_
A
3
= A B
2
=
_
_
35 0 0
0 35 0
0 0 35
_
_
p
3
=
1
3
trace A
3
= 35
Thus characteristic polynomial is

3
12
2
+ 40 35
Note that B
3
= A
3
35I = 0 conrming check.
5(b) Characteristic polynomial is
4
p
1

3
p
2

2
p
3
p
4
with
p
1
= trace A = 4
B
1
= A4I =
_

_
2 1 1 2
0 3 1 0
1 1 3 1
1 1 1 4
_

_
A
2
= A B
1
=
_

_
3 4 0 3
1 2 2 1
2 0 2 5
3 3 1 3
_

_
p
2
=
1
2
trace A
2
= 2
c _
Pearson Education Limited 2004
Glyn James: Advanced Modern Engineering Mathematics, Third edition 291
B
2
= A
2
+ 2I =
_

_
1 4 0 3
1 0 2 1
2 0 0 5
3 3 1 5
_

_
A
3
= A B
2
=
_

_
5 2 0 2
1 0 2 4
1 7 3 4
0 4 2 7
_

_
p
3
=
1
3
trace A
3
= 5
B
3
= A
3
+ 5I =
_

_
0 0 0 2
1 5 2 4
1 8 2 4
0 4 2 2
_

_
A
4
= A B
3
=
_

_
2 0 0 0
0 2 0 0
0 0 2 0
0 0 0 2
_

_
p
4
=
1
4
trace A
4
= 2
Thus characteristic polynomial is
4
4
3
+ 2
2
+ 5 + 2
Note that B
4
= A
4
+ 2I = 0 as required by check.
6(a) Eigenvalues given by

1
1
1
1

=
2
2 = ( 2) = 0
so eigenvectors are
1
= 2,
2
= 0
Eigenvectors given by corresponding solutions of
(1
i
)e
i1
+ e
i2
= 0
e
i1
+ (1
i
)e
i2
= 0
Taking i = 1, 2 gives the eigenvectors as
e
1
= [1 1]
T
, e
2
= [1 1]
T
(1)
6(b) Eigenvalues given by

1
3
2
2

=
2
3 4 = ( + 1)( 4) = 0
so eigenvectors are
1
= 4,
2
= 1
Eigenvectors given by corresponding solutions of
(l
i
)e
i1
+ 2e
i2
= 0
3e
i1
+ (2
i
)e
i2
= 0
c _
Pearson Education Limited 2004
292 Glyn James: Advanced Modern Engineering Mathematics, Third edition
Taking i = 1, 2 gives the eigenvectors as
e
1
= [2 3]
T
, e
2
= [1 1]
T
6(c) Eigenvalues given by

1 0 4
0 5 4
4 4 3

=
3
+ 9
2
+ 9 81 = ( 9)( 3)( + 3) = 0
So the eigenvalues are
1
= 9,
2
= 3,
3
= 3.
The eigenvectors are given by the corresponding solutions of
(1
i
)e
i1
+ 0e
i2
4e
i3
= 0
0e
i1
+ (5
i
)e
i2
+ 4e
i3
= 0
4e
i1
+ 4e
i2
+ (3
i
)e
i3
= 0
Taking i = 1,
i
= 9 solution is
e
11
8
=
e
12
16
=
e
13
16
=
1
e
1
= [1 2 2]
T
Taking i = 2,
i
= 3 solution is
e
21
16
=
e
22
16
=
e
23
8
=
2
e
2
= [2 2 1]
T
Taking i = 3,
i
= 3 solution is
e
31
32
=
e
32
16
=
e
33
32
=
3
e
3
= [2 1 2]
T
6(d) Eigenvalues given by

1 1 2
0 2 2
1 1 3

= 0
Adding column 1 to column 2 gives

1 2 2
0 2 2
1 0 3

= (2 )

1 1 2
0 1 2
1 0 3

R
1
R
2
(2 )

1 0 0
0 1 2
1 0 3

= (2 )(1 )(3 )
c _
Pearson Education Limited 2004
Glyn James: Advanced Modern Engineering Mathematics, Third edition 293
so the eigenvalues are
1
= 3,
2
= 2,
3
= 1.
Eigenvectors are the corresponding solutions of (A
i
I)e
i
= 0
When =
1
= 3 we have
_
_
2 1 2
0 1 2
1 1 0
_
_
_
_
e
11
e
12
e
13
_
_
= 0
leading to the solution
e
11
2
=
e
12
2
=
e
13
1
=
1
so the eigenvector corresponding to
2
= 3 is e
1
=
1
[2 2 1]
T
,
1
constant.
When =
2
= 2 we have
_
_
1 1 2
0 0 2
1 1 1
_
_
_
_
e
21
e
22
e
23
_
_
= 0
leading to the solution
e
21
2
=
e
22
2
=
e
23
0
=
3
so the eigenvector corresponding to
2
= 2 is e
2
=
2
[1 1 0]
T
,
2
constant.
When =
3
= 1 we have
_
_
0 1 2
0 1 2
1 1 2
_
_
_
_
e
31
e
32
e
33
_
_
= 0
leading to the solution
e
31
0
=
e
32
2
=
e
33
1
=
1
so the eigenvector corresponding to
3
= 1 is e
3
=
3
[0 2 1]
T
,
3
constant.
6(e) Eigenvalues given by

5 0 6
0 11 6
6 6 2

=
3
14
2
23 686 = ( 14)( 7)( + 7) = 0
so eigenvalues are
1
= 14,
2
= 7,
3
= 7
c _
Pearson Education Limited 2004
294 Glyn James: Advanced Modern Engineering Mathematics, Third edition
Eigenvectors are given by the corresponding solutions of
(5
i
)e
i1
+ 0e
i2
+ 6e
i3
= 0
0e
i1
+ (11
i
)e
i2
+ 6e
i3
= 0
6e
i1
+ 6e
i2
+ (2
i
)e
i3
= 0
When i = 1,
1
= 14 solution is
e
11
12
=
e
12
36
=
e
13
18
=
1
e
1
= [2 6 3]
T
When i = 2,
2
= 7 solution is
e
21
72
=
e
22
36
=
e
23
24
=
2
e
2
= [6 3 2]
T
When i = 3,
3
= 7 solution is
e
31
54
=
e
32
36
=
e
33
108
=
3
e
3
= [3 2 6]
T
6(f ) Eigenvalues given by

1 1 0
1 2 1
2 1 1

R
1
+R
2

1 0 1
1 2 1
2 1 1

= (1 + )

1 0 0
1 2 0
2 1 1

= 0, i.e. (1 + )(2 )(1 ) = 0


so eigenvalues are
1
= 2,
2
= 1,
3
= 1
Eigenvectors are given by the corresponding solutions of
(1
i
)e
i1
e
i2
+ 0e
i3
= 0
e
i1
+ (2
i
)e
i2
+ e
i3
= 0
2e
i1
+ e
i2
(1 +
i
)e
i3
= 0
Taking i = 1, 2, 3 gives the eigenvectors as
e
1
= [1 1 1]
T
, e
2
= [1 0 1]
T
, e
3
= [1 2 7]
T
c _
Pearson Education Limited 2004
Glyn James: Advanced Modern Engineering Mathematics, Third edition 295
6(g) Eigenvalues given by

4 1 1
2 5 4
1 1

R
1
+ (R
2
+ R
3
)

5 5 5
2 5 4
1 1

= (5 )

1 0 0
2 3 2
1 0 1

= (5 )(3 )(1 ) = 0
so eigenvalues are
1
= 5,
2
= 3,
3
= 1
Eigenvectors are given by the corresponding solutions of
(4
i
)e
i1
+ e
i2
+ e
i3
= 0
2e
i1
+ (5
i
)e
i2
+ 4e
i3
= 0
e
i1
e
i2

i
e
i3
= 0
Taking i = 1, 2, 3 and solving gives the eigenvectors as
e
1
= [2 3 1]
T
, e
2
= [1 1 0]
T
, e
3
= [0 1 1]
T
6(h) Eigenvalues given by

1 4 2
0 3 1
1 2 4

R
1
+2R
2

1 2 2 0
0 3 1
1 2 4

= (1 )

1 0 0
0 3 1
1 0 4

= (1 )(3 )(4 ) = 0
so eigenvalues are
1
= 4,
2
= 3,
3
= 1
Eigenvectors are given by the corresponding solutions of
(1
i
)e
i1
4e
i2
2e
i3
= 0
2e
i1
+ (3
i
)e
i2
+ e
i3
= 0
e
i1
+ 2e
i2
+ (4
i
)e
i3
= 0
c _
Pearson Education Limited 2004
296 Glyn James: Advanced Modern Engineering Mathematics, Third edition
Taking i = 1, 2, 3 and solving gives the eigenvectors as
e
1
= [2 1 1]
T
, e
2
= [2 1 0]
T
, e
3
= [4 1 2]
T
Exercises 6.4.5
7(a) Eigenvalues given by

2 2 1
1 3 1
1 2 2

R
1
R
2

1 1 + 0
0 3 1
1 2 2

= (1 )

1 0 0
1 4 1
1 3 2

= (1 )[
2
6 + 5] = (1 )( 1)( 5) = 0
so eigenvalues are
1
= 5,
2
=
3
= 1
The eigenvectors are the corresponding solutions of
(2
i
)e
i1
+ 2e
i2
+ e
i3
= 0
e
i1
+ (3
i
)e
i2
+ e
i3
= 0
e
i1
+ 2e
i2
+ (2
i
)e
i3
= 0
When i = 1,
1
= 5 and solution is
e
11
4
=
e
12
4
=
e
13
4
=
1
e
1
= [1 1 1]
T
When
2
=
3
= 1 solution is given by the single equation
e
21
+ 2e
22
+ e
23
= 0
Following the procedure of Example 6.5 we can obtain two linearly independent
solutions. A possible pair are
e
2
= [0 1 2]
T
, e
3
= [1 0 1]
T
c _
Pearson Education Limited 2004
Glyn James: Advanced Modern Engineering Mathematics, Third edition 297
7(b) Eigenvalues given by

2 2
1 1 2
1 1 2

=
3
+ 3
2
4 = ( + 1)( 2)
2
= 0
so eigenvalues are
1
=
2
= 2,
3
= 1
The eigenvectors are the corresponding solutions of

i
e
i1
2e
i2
2e
i3
= 0
e
i1
+ (1
i
)e
i2
+ 2e
i3
= 0
e
i1
e
i2
+ (2
i
)e
i3
= 0
When i = 3,
3
= 1 corresponding solution is
e
31
8
=
e
32
1
=
e
33
3
=
3
e
3
= [8 1 3]
T
When
1
=
2
= 2 solution is given by
2e
21
2e
22
2e
23
= 0 (1)
e
21
e
22
+ 2e
23
= 0 (2)
e
21
e
22
= 0 (3)
From (1) and (2) e
23
= 0 and it follows from (3) that e
21
= e
22
. We deduce that
there is only one linearly independent eigenvector corresponding to the repeated
eigenvalues = 2. A possible eigenvector is
e
2
= [1 1 0]
T
7(c) Eigenvalues given by

4 6 6
1 3 2
1 5 2

R
1
3R
3

1 3 + 3 0
1 3 2
1 5 2

= (1 )

1 3 0
1 3 2
1 5 2

= (1 )

1 0 0
1 6 2
1 8 2

= (1 )(
2
+ + 4) = (1 )( 2)
2
= 0
c _
Pearson Education Limited 2004
298 Glyn James: Advanced Modern Engineering Mathematics, Third edition
so eigenvalues are
1
=
2
= 2,
3
= 1.
The eigenvectors are the corresponding solutions of
(4
i
)e
i1
+ 6e
i2
+ 6e
i3
= 0
e
i1
+ (3
i
)e
i2
+ 2e
i3
= 0
e
i1
5e
i2
(2 +
i
)e
i3
= 0
When i = 3,
3
= 1 corresponding solution is
e
31
4
=
e
32
1
=
e
33
3
=
3
e
3
= [4 1 3]
T
When
1
=
2
= 2 solution is given by
2e
21
+ 6e
22
+ 6e
23
= 0
e
21
+ e
22
+ 2e
23
= 0
e
21
5e
22
4e
23
= 0
so that
e
21
6
=
e
22
2
=
e
23
4
=
2
leading to only one linearly eigenvector corresponding to the eigenvector = 2. A
possible eigenvector is
e
2
= [3 1 2]
T
7(d) Eigenvalues given by

7 2 4
3 2
6 2 3

R
1
2R
2

1 2 + 2 0
3 2
6 2 3

= (1 )

1 2 0
3 2
6 2 3

= (1 )

1 0 0
3 6 2
6 10 3

= (1 )( 2)( 1) = 0
so eigenvalues are
1
= 2,
2
=
3
= 1.
The eigenvectors are the corresponding solutions of
(7
i
)e
i1
2e
i2
4e
i3
= 0
3e
i1

i
e
i2
2e
i3
= 0
6e
i1
2e
i2
(3 +
i
)e
i3
= 0
c _
Pearson Education Limited 2004
Glyn James: Advanced Modern Engineering Mathematics, Third edition 299
When i = 1,
2
= 2 and solution is
e
11
6
=
e
12
3
=
e
13
6
=
1
e
1
= [2 1 2]
T
When
2
=
3
= 1 the solution is given by the single equation
3e
21
e
22
2e
23
= 0
Following the procedures of Example 6.5 we can obtain two linearly independent
are solutions. A possible pair are
e
2
= [0 2 1]
T
, e
3
= [2 0 3]
T
8
(AI) =
_
_
4 7 5
2 3 3
1 2 1
_
_
Performing a series of row and column operators this may be reduced to the form
_
_
0 0 0
0 0 1
1 0 0
_
_
indicating that (A I) is of rank 2. Thus the nullity q = 3 2 = 1
conrming that there is only one linearly independent eigenvector associated with
the eigenvalue = 1. The eigenvector is given by the solution of
4e
11
7e
12
5e
13
= 0
2e
11
+ 3e
12
+ 3e
13
= 0
e
11
+ 2e
12
+ e
13
= 0
giving
e
11
3
=
e
12
1
=
e
13
1
=
1
e
1
= [3 1 1]
T
9
(AI) =
_
_
1 1 1
1 1 1
1 1 1
_
_
c _
Pearson Education Limited 2004
300 Glyn James: Advanced Modern Engineering Mathematics, Third edition
Performing a series of row and column operators this may be reduced to the form
_
_
1 0 0
0 0 0
0 0 0
_
_
indicating that (AI) is of rank 1. Then the nullity of q = 31 = 2
conrming that there are two linearly independent eigenvectors associated with the
eigenvalue = 1. The eigenvectors are given by the single equation
e
11
+ e
12
e
13
= 0
and two possible linearly independent eigenvectors are
e
1
= [1 0 1]
T
and e
2
= [0 1 1]
T
Exercises 6.4.8
10 These are standard results.
11(a) (i) Trace A = 4 + 5 + 0 = 9 = sum eigenvalues;
(ii) det A = 15 = 5 3 1 = product eigenvalues;
(iii) A
1
=
1
15
_
_
4 1 1
4 1 14
3 3 18
_
_
. Eigenvalues given by

4 15 1 1
4 1 15 14
3 3 18 15

C
3
C
2

4 15 1 0
4 1 15 15 + 15
3 3 15 15

= (15 15)

4 15 1 0
4 1 15 1
3 3 1

= (15 15)(15 5)(15 3) = 0


conrming eigenvalues as 1,
1
3
,
1
5
.
c _
Pearson Education Limited 2004
Glyn James: Advanced Modern Engineering Mathematics, Third edition 301
(iv) A
T
=
_
_
4 2 1
1 5 1
1 4 0
_
_
having eigenvalues given by

4 2 1
1 5 1
1 4

= ( 5)( 3)( 1) = 0
i.e. eigenvalue as for A.
11(b) (i) 2A =
_
_
8 2 2
4 10 8
2 2 0
_
_
having eigenvalues given by

8 2 2
4 10 8
2 2

C
1
C
2

6 2 2
6 + 10 8
0 2

= (6 )

1 2 2
1 10 8
0 2

= (6 )

1 2 2
0 12 10
0 2

= (6 )( 10)( 2) = 0
Thus eigenvalues are 2 times those of A; namely 6, 10 and 2.
(ii) A+ 2I =
_
_
6 1 1
2 7 4
1 1 2
_
_
having eigenvalues given by

6 1 1
2 7 4
1 1 2

=
3
+ 15
2
71 + 105 = ( 7)( 5)( 3) = 0
conrming the eigenvalues as 5 + 2, 3 + 2, 1 + 2.
Likewise for A2I
c _
Pearson Education Limited 2004
302 Glyn James: Advanced Modern Engineering Mathematics, Third edition
(iii) A
2
=
_
_
17 8 8
14 23 22
6 6 5
_
_
having eigenvalues given by

17 8 8
14 23 22
6 6 5

R
1
+ (R
2
) + R
3
)

25 25 25
14 23 22
6 6 5

= (25 )

1 0 0
14 9 8
6 0 1

= (25 )(9 )(1 ) = 0


i.e. eigenvalues A
2
are 25, 9, 1 which are those of A squared.
12 Eigenvalues of A given by

3 3 3
3 1 1
3 1 1

R
3
+R
2

3 3 3
3 1 1
0 2 + 2

= ( 2)

3 3 3
3 1 1
0 1 1

C
3
+C
2
( 2)

3 3 6
3 (1 )
0 1 0

= ( 2)( + 6)( 3) = 0
so eigenvalues are
1
= 6,
2
= 3,
3
= 2
Eigenvectors are given by corresponding solutions of
(3
i
)e
i1
3e
i2
3e
i3
= 0
3e
i1
+ (1
i
)e
i2
e
i3
= 0
3e
i1
e
i2
+ (1
i
)e
i3
= 0
Taking i = 1, 2, 3 gives the eigenvectors as
e
1
= [2 1 1]
T
, e
2
= [1 1 1]
T
, e
3
= [0 1 1]
T
It is readily shown that
e
T
1
e
2
= e
T
1
e
3
= e
T
2
e
3
= 0
so that the eigenvectors are mutually orthogonal.
c _
Pearson Education Limited 2004
Glyn James: Advanced Modern Engineering Mathematics, Third edition 303
13 Let the eigenvector be e = [a b c]
T
then since the three vectors are mutually
orthogonal
a + b 2c = 0
a + b c = 0
giving c = 0 and a = b so an eigenvector corresponding to = 2 is e = [1 1 0]
T
.
Exercises 6.5.3
14 Taking x
(0)
= [1 1 1]
T
iterations may then be tabulated as follows:
Iteration k 0 1 2 3 4
1 .9 .874 .869 .868
x
(k)
1 1 1 1 1
1 .5 .494 .493 .492
9 7.6 7.484 7.461 7.457
A x
(k)
10 8.7 8.61 8.592 8.589
5 4.3 4.242 4.231 4.228
10 8.7 8.61 8.592 8.589
Thus estimate of dominant eigenvalue is 8.59 and corresponding eigenvector
x [7.46 8.59 4.23]
T
or x [0.62 0.71 0.35]
T
in normalised form.
15(a) Taking x
(0)
= [1 1 1]
T
iterations may then be tabulated as follows:
Iteration k 0 1 2 3 4 5 6
1 .75 .667 0.636 0.625 0.620 0.619
x
(k)
1 1 1 1 1 1 1
1 1 1 1 1 1 1
3 2.5 2.334 2.272 2.250 2.240
A x
(k)
4 3.75 3.667 3.636 3.625 3.620
4 3.75 3.667 3.636 3.625 3.620
4 3.75 3.667 3.636 3.625 3.620
Thus correct to two decimal places dominant eigenvalue is 3.62 having
corresponding eigenvectors [0.62 1 1]
T
.
c _
Pearson Education Limited 2004
304 Glyn James: Advanced Modern Engineering Mathematics, Third edition
15(b) Taking x
(0)
= [1 1 1]
T
iterations may be tabulated as follows:
Iteration k 0 1 2 3 4 5
1 0.364 0.277 0.257 0.252 0.251
x
(k)
1 0.545 0.506 0.501 0.493 0.499
1 1 1 1 1 1
4 2.092 1.831 1.771 1.756
A x
(k)
6 3.818 3.566 3.561 3.49
11 7.546 7.12 7.03 6.994
11 7.546 7.12 7.03 6.994
Thus correct to two decimal places dominant eigenvalue is 7 having corresponding
eigenvector [0.25 0.5 1]
T
.
15(c) Taking x
(0)
= [1 1 1 1]
T
iterations may then be tabulated as follows:
Iteration k 0 1 2 3 4 5 6
1 1 1 1 1 1 1
x
(k)
1 0 -0.5 -0.6 -0.615 -0.618 - 0.618
1 1 -0.5 -0.6 -0.615 -0.618 -0.618
1 1 1 1 1 1 1
1 2 2.5 2.6 2.615 2.618
A x
(k)
0 -1 -1.5 -1.6 - 1.615 -1.618
0 -1 -1.5 -1.6 -1.615 -1.618
1 2 2.5 2.6 2.615 2.618
1 2 2.5 2.6 2.615 2.618
Thus correct to two decimal places dominant eigenvalue is 2.62 having
corresponding eigenvector [1 0.62 0.62 1]
T
.
16 The eigenvalue
1
corresponding to the dominant eigenvector e
1
= [1 1 2]
T
is such that A e
1
=
1
e
1
so
_
_
3 1 1
1 3 1
1 1 5
_
_
_
_
1
1
2
_
_
=
1
_
_
1
1
2
_
_
so
1
= 6.
c _
Pearson Education Limited 2004
Glyn James: Advanced Modern Engineering Mathematics, Third edition 305
Then
A
1
= A6 e
1
e
T
1
where e
1
=
_
1

6
1

6
2

T
so
A
1
=
_
_
3 1 1
1 3 1
1 1 5
_
_

_
_
1 1 2
1 1 2
2 2 4
_
_
=
_
_
2 0 1
0 2 1
1 1 1
_
_
Applying the power method with x
(0)
= [1 1 1]
T
y
(1)
= A
1
x
(0)
=
_
_
1
1
1
_
_
= x
(1)
y
(2)
= A
1
x
(1)
=
_
_
3
3
3
_
_
= 3
_
_
1
1
1
_
_
Clearly
2
= 3 and e
2
=
1

3
[1 1 1]
T
.
Repeating the process
A
2
= A
1

2
e
2
e
T
2
=
_
_
2 0 1
0 2 1
1 1 1
_
_

_
_
1 1 1
1 1 1
1 1 1
_
_
=
_
_
1 1 0
1 1 0
0 0 0
_
_
Taking x
(0)
= [1 1 0]
T
the power method applied to A
2
gives
y
(1)
= A
2
x
(0)
=
_
_
2
2
0
_
_
= 2
_
_
1
1
0
_
_
and clearly
3
= 2 with e
3
=
1

2
[1 1 0]
T
.
17 The three Gerschgorin circles are
[ 5 [= 2, [ [= 2, [ + 5 [= 2
c _
Pearson Education Limited 2004
306 Glyn James: Advanced Modern Engineering Mathematics, Third edition
which are three non-intersecting circles. Since the given matrix A is symmetric its
three eigenvalues are real and it follows from Theorem 6.2 that
3 <
1
< 7 , 2 <
2
< 2 , 7 <
3
< 7
(that is, an eigenvalue lies within each of the three circles).
18 The characteristic equation of the matrix A is

10 1 0
1 2 2
0 2 3

= 0
i.e. (10 )[(2 )(3 ) 4] (3 ) = 0
or f() =
3
15
2
+ 51 17 = 0
Taking
0
= 10 as the starting value the NewtonRaphson iterative process
produces the following table:
i
i
f(
i
) f

(
i
)
f(
i
)
f

(
i
)
0 10 7 -51.00 0.13725
1 10.13725 -0.28490 -55.1740 -0.00516
2 10.13209 -0.00041 -55.0149 -0.000007
Thus to three decimal places the largest eigenvalue is = 10.132
Using Properties 6.1 and 6.2 of section 6.4.6 we have
3

i=1

i
= trace A = 15 and
3

i=1

i
=[ A [= 17
c _
Pearson Education Limited 2004
Glyn James: Advanced Modern Engineering Mathematics, Third edition 307
Thus

2
+
3
= 15 10.132 = 4.868

3
= 1.67785
so
2
(4.868
2
) = 1.67785

2
2
4.868
2
+ 1.67785 = 0

2
= 2.434 2.0607
i.e.
2
= 4.491 and
3
= 0.373
19(a) If e
1
, e
2
, . . . , e
n
are the corresponding eigenvectors to
1
,
2
, . . . ,
n
then
(KI A)e
i
= (K
i
)e
i
so that A and (KI A) have the same eigenvectors
and eigenvalues dier by K.
Taking x
(o)
=
n

i=1

r
e
i
then
x
(p)
= (KI A)x
(p1)
= (KI A)
2
x
(p2)
= . . . =
n

r=1

r
(K
r
)
p
e
r
Now K
n
> K
n1
> . . . > K
1
and
x
(p)
=
n
(K
n
)
p
e
n
+
n

r=1

r
(K
r
)
p
e
r
= (K
n
)
p
[
n
e
n
+
n1

r=1

r
_
K
r
K
n

p
e
r
]
(K
n
)
p

n
e
n
= Ke
n
as p
Also
x
(p+1)
i
x
(p)
i

(K
n
)
p+1
(K
n
)
p

n
e
n

n
e
n
= K
n
Hence we can nd
n
19(b) Since A is a symmetric matrix its eigenvalues are real. By Gerschgorins
theorem the eigenvalues lie in the union of the intervals
[ 2 [ 1, [ 2 [ 2, [ 2 [ 1
i.e. [ 2 [ 2 or 0 4.
c _
Pearson Education Limited 2004
308 Glyn James: Advanced Modern Engineering Mathematics, Third edition
Taking K = 4 in (a)
KI A = 4I A =
_
_
2 1 0
1 2 1
0 1 2
_
_
Taking x
(0)
= [1 1 1]
T
iterations using the power method are tabulated as follows:
Iteration k 0 1 2 3 4 5
1 0.75 0.714 0.708 0.707 0.707
x
(k)
1 1 1 1 1 1
1 0.75 0.714 0.708 0.707 0.707
3 2.5 2.428 2.416 2.414
A x
(k)
4 3.5 3.428 3.416 3.414
3 2.5 2.428 2.416 2.414
4 3.5 3.428 3.416 3.414
Thus
3
= 4 3.41 = 0.59 correct to two decimal places.
Exercises 6.6.3
20 Eigenvalues given by
=

1 6 12
0 13 30
0 9 20

= 0
Now = (1 )

13 30
9 20

= (1 )(
2
7 + 10)
= (1 )( 5)( 2) so = 0 gives
1
= 5,
2
= 2,
3
= 1
Corresponding eigenvectors are given by the solutions of
(A
i
I)e
i
= 0
When =
1
= 5 we have
_
_
6 6 12
0 18 30
0 9 15
_
_
_
_
e
11
e
12
e
13
_
_
= 0
c _
Pearson Education Limited 2004
Glyn James: Advanced Modern Engineering Mathematics, Third edition 309
leading to the solution
e
11
36
=
e
12
180
=
e
13
108
=
1
so the eigenvector corresponding to
1
= 5 is e
1
=
1
[1 5 3]
T
When =
2
= 2 we have
_
_
3 6 12
0 15 30
0 9 18
_
_
_
_
e
21
e
22
e
23
_
_
= 0
leading to the solution
e
21
0
=
e
22
90
=
e
23
45
=
2
so the eigenvector corresponding to
2
= 2 is e
2
=
2
[0 2 1]
T
When =
3
= 1 we have
_
_
0 6 12
0 12 30
0 9 21
_
_
_
_
e
31
e
32
e
33
_
_
= 0
leading to the solution
e
31
18
=
e
32
0
=
e
33
0
=
3
so the eigenvector corresponding to
3
= 1 is e
3
=
3
[1 0 0]
T
A modal matrix M and spectral matrix are
M =
_
_
1 0 1
5 2 0
3 1 0
_
_
=
_
_
5 0 0
0 2 0
0 0 1
_
_
M
1
=
_
_
0 1 2
0 3 5
1 1 2
_
_
and matrix multiplication conrms M
1
A M =
21 From Example 6.8 the eigenvalues and corresponding normalised eigenvectors
of A are

1
= 6,
2
= 3,
3
= 1
e
1
=
1

5
[1 2 0]
T
, e
2
= [0 0 1]
T
, e
3
=
1

5
[2 1 0]
T
,

M =
1

5
_
_
1 0 2
2 0 1
0

5 0
_
_
c _
Pearson Education Limited 2004
310 Glyn James: Advanced Modern Engineering Mathematics, Third edition

M
T
A M =
1
5
_
_
1 2 0
0 0

5
2 1 0
_
_
_
_
2 2 0
2 5 0
0 0 3
_
_
_
_
1 0 2
2 0 1
0

5 0
_
_
=
1
5
_
_
6 12 0
0 0 3

5
2 1 0
_
_
_
_
1 0 2
2 0 1
0

5 0
_
_
=
1
5
_
_
30 0 0
0 15 0
0 0 5
_
_
=
_
_
6 0 0
0 3 0
0 0 1
_
_
=
22 The eigenvalues of A are given by

5 10 8
10 2 2
8 2 11

= (
3
18
2
81+1458) = (9)(+9)(18) = 0
so eigenvalues are
1
= 18,
2
= 9,
3
= 9
The eigenvectors are given by the corresponding solutions of
(5
i
)e
i1
+ 10e
i2
+ 8e
i3
= 0
10e
i1
+ (2
i
)e
i2
2e
i3
= 0
8e
i1
2e
i2
+ (11
i
)e
i3
= 0
Taking i = 1, 2, 3 and solving gives the eigenvectors as
e
1
= [2 1 2]
T
, e
2
= [1 2 2]
T
, e
3
= [2 2 1]
T
Corresponding normalised eigenvectors are
e
1
=
1
3
[2 1 2]
T
, e
2
=
1
3
[1 2 2]
T
, e
3
=
1
3
[2 2 1]
T

M =
1
3
_
_
2 1 2
1 2 2
2 2 1
_
_
,

M
T
=
1
3
_
_
2 1 2
1 2 2
2 2 1
_
_
c _
Pearson Education Limited 2004
Glyn James: Advanced Modern Engineering Mathematics, Third edition 311

M
T
A M =
1
9
_
_
2 1 2
1 2 2
2 2 1
_
_
_
_
5 10 8
10 2 2
8 2 11
_
_
_
_
2 1 2
1 2 2
2 2 1
_
_
=
1
9
_
_
36 18 36
9 18 18
18 18 9
_
_
_
_
2 1 2
1 2 2
2 2 1
_
_
=
_
_
4 2 4
1 2 2
2 2 1
_
_
_
_
2 1 2
1 2 2
2 2 1
_
_
=
_
_
18 0 0
0 9 0
0 0 9
_
_
=
23
A =
_
_
1 1 2
1 2 1
0 1 1
_
_
Eigenvalues given by
0 =

1 1 2
1 2 1
0 1 1

= (
3
2
2
+2) = (1)(2)(+1) = 0
so eigenvalues are
1
= 2,
2
= 1,
3
= 1.
The eigenvectors are given by the corresponding solutions of
(1
i
)e
i1
+ e
i2
2e
i3
= 0
e
i1
+ (2
i
)e
i2
+ e
i3
= 0
0e
i1
+ e
i2
(1 +
i
)e
i3
= 0
Taking i = 1, 2, 3 and solving gives the eigenvectors as
e
1
= [1 3 1]
T
, e
2
= [3 2 1]
T
, e
3
= [1 0 1]
T
M =
_
_
1 3 1
3 2 0
1 1 1
_
_
, =
_
_
2 0 0
0 1 0
0 0 1
_
_
M
1
=
1
6
_
_
2 2 2
3 0 3
1 2 7
_
_
c _
Pearson Education Limited 2004
312 Glyn James: Advanced Modern Engineering Mathematics, Third edition
Matrix multiplication then conrms
M
1
A M = and A = M M
1
24 Eigenvalues given by

3 2 4
2 2 6
4 6 1

=
3
+ 63 162 = ( + 9)( 6)( 3) = 0
so the eigenvalues are
1
= 9,
2
= 6,
3
= 3. The eigenvectors are the
corresponding solutions of
(3
i
)e
i1
2e
i2
+ 44e
i3
= 0
2e
i1
(2 +
i
)e
i2
+ 6e
i3
= 0
4e
i1
+ 6e
i2
(1 +
i
)e
i3
= 0
Taking i = 1, 2, 3 and solving gives the eigenvectors as
e
1
= [1 2 2]
T
, e
2
= [2 1 2]
T
, e
3
= [2 2 1]
T
Since e
T
1
e
2
= e
T
1
e
3
= e
T
2
e
3
= 0 the eigenvectors are orthogonal
L = [ e
1
e
2
e
3
] =
1
3
_
_
1 2 2
2 1 2
2 2 1
_
_

L A L =
1
9
_
_
1 2 2
2 1 2
2 2 1
_
_
_
_
3 2 4
2 2 6
4 6 1
_
_
_
_
1 2 2
2 1 2
2 2 1
_
_
=
1
9
_
_
9 18 18
12 6 12
6 6 3
_
_
_
_
1 2 2
2 1 2
2 2 1
_
_
=
1
9
_
_
81 0 0
0 54 0
0 0 27
_
_
=
_
_
9 0 0
0 6 0
0 0 3
_
_
=
25 Since the matrix A is symmetric the eigenvectors
e
1
= [1 2 0]
T
, e
2
= [2 1 0]
T
, e
3
= [e
31
e
32
e
33
]
T
c _
Pearson Education Limited 2004
Glyn James: Advanced Modern Engineering Mathematics, Third edition 313
are orthogonal. Hence
e
T
1
e
3
= e
31
+ 2e
32
= 0 and e
T
2
e
3
= 2e
31
+ e
32
= 0
Thus e
31
= e
32
= 0 and e
33
arbitrary so a possible eigenvector is e
3
= [0 0 1]
T
.
Using A =

M

M
T
where =
_
_
6 0 0
0 1 0
0 0 3
_
_
gives
A =
_
_
1

5

2

5
0
2

5
1

5
0
0 0 1
_
_
_
_
6 0 0
0 1 0
0 0 3
_
_
_
_
1

5
2

5
0

5
1

5
0
0 0 1
_
_
=
_
_
2 2 0
2 5 0
0 0 3
_
_
26 AI =
_
_
4 7 5
2 3 3
1 2 1
_
_

_
_
0 0 0
0 1 0
1 0 0
_
_
is of rank 2
Nullity (AI) = 3 2 = 1 so there is only one linearly independent vector
corresponding to the eigenvalue 1. The corresponding eigenvector e
1
is given by
the solution of (AI)e
1
= 0 or
4e
11
7e
12
5e
13
= 0
2e
11
+ 3e
12
+ 3e
13
= 0
e
11
+ 2e
12
+ 2
12
= 0
that is, e
1
= [3 1 1]
T
. To obtain the generalised eigenvector e

1
we solve
(AI)e

1
= e
1
or
_
_
4 7 5
2 3 3
1 2 1
_
_
_
_
e

11
e

12
e

13
_
_
=
_
_
3
1
1
_
_
giving e

1
= [1 1 0]
T
. To obtain the second generalised eigenvector e

1
we solve
(AI)e

1
= e

1
or
_
_
4 7 5
2 3 3
1 2 1
_
_
_
_
e

11
e

12
e

13
_
_
=
_
_
1
1
0
_
_
c _
Pearson Education Limited 2004
314 Glyn James: Advanced Modern Engineering Mathematics, Third edition
giving e

1
= [2 1 0]
T
.
M = [e
1
e

1
e

1
] =
_
_
3 1 2
1 1 1
1 0 0
_
_
det M = 1 and M
1
=
_
_
0 0 1
1 2 1
1 1 2
_
_
=
_
_
0 0 1
1 2 1
1 1 2
_
_
Matrix multiplication then conrms
M
1
A M =
_
_
1 1 0
0 1 1
0 0 1
_
_
27 Eigenvalues are given by
[ AI [= 0
that is,
4
4
3
12
2
+ 32 + 64 = ( + 2)
2
( 4)
2
= 0 so the eigenvalues are
-2, -2, 4 and 4 as required.
Corresponding to the repeated eigenvalue
1
,
2
= 2
(A+ 2I) =
_

_
3 0 0 3
0 3 3 0
0.5 3 3 0.5
3 0 0 3
_

_

_

_
1 0 0 0
0 1 0 0
0 0 0 0
0 0 0 0
_

_
is of rank 2
Thus nullity (A+2I) is 42 = 2 so there are two linearly independent eigenvectors
corresponding to = 2.
Corresponding to the repeated eigenvalues
3
,
4
= 4
(A4I) =
_

_
3 0 0 3
0 3 3 0
0.5 3 3 0.5
3 0 0 3
_

_

_

_
1 0 0 0
0 1 0 0
0 0 0 0
0 0 0 1
_

_
is of rank 3
Thus nullity (A 4I) is 4 3 = 1 so there is only one linearly independent
eigenvector corresponding to = 4.
When =
1
=
2
= 2 the eigenvalues are given by the solution of (A+2I)e = 0
giving e
1
= [0 1 1 0]
T
, e
2
= [1 0 0 1]
T
as two linearly independent solutions.
c _
Pearson Education Limited 2004
Glyn James: Advanced Modern Engineering Mathematics, Third edition 315
When =
3
=
4
= 8 the eigenvectors are given by the solution of
(A4I)e = 0
giving the unique solution e
3
= [0 1 1 0]
T
. The generalised eigenvector e

3
is
obtained by solving
(A4I)e

3
= e
3
giving e

3
= (6 1 0 6]
T
. The Jordan canonical form is
J =
_

_
2 0 0 0
0 2 0 0
0 0 4 1
0 0 0 4
_

_
Exercises 6.6.5
28(a) The matrix of the quadratic form is A =
_
_
1 1 2
1 2 1
2 1 7
_
_
and its leading
principal minors are
1,

1 1
1 2

= 1, det A = 2
Thus by Sylvesters condition (a) the quadratic form is positive denite.
28(b) Matrix A =
_
_
1 1 2
1 2 1
2 1 5
_
_
and its leading principal minors are
1,

1 1
1 2

= 1, det A = 0
Thus by Sylvesters condition (c) the quadratic form is positive semidenite.
c _
Pearson Education Limited 2004
316 Glyn James: Advanced Modern Engineering Mathematics, Third edition
28(c) Matrix A =
_
_
1 1 2
1 2 1
2 1 4
_
_
and its leading principal minors are
1,

1 1
1 2

= 1, det A = 1.
Thus none of Sylvesters conditions are satised and the quadratic form is indenite.
29(a) The matrix of the quadratic form is A =
_
a b
b c
_
and its leading
principal minors are a and ac b
2
. By Sylvesters condition (a) in the text the
quadratic form is positive denite if and only if
a > 0 and ac b
2
> 0
i.e. a > 0 and ac > b
2
29(b) The matrix of the quadratic form is A =
_
_
2 1 0
1 a b
0 b 3
_
_
having principal
minors 2, 2a 1 and det A = 6a 2b
2
3. Thus by Sylvesters condition (a) in
the text the quadratic form is positive denite if and only if
2a 1 > 0 and 6a 2b
2
3 > 0
or 2a > 1 and 2b
2
< 6a 3
30 The eigenvalues of the matrix A are given by
0 =

2 1 1
1 2 1
1 1 2

R
1
+R
3

3 3 0
1 2 1
1 1 2

= (3 )

1 1 0
1 2 1
1 1 2

= (3 )

1 0 0
1 1 1
1 2 2

= (3 )(
2
3)
c _
Pearson Education Limited 2004
Glyn James: Advanced Modern Engineering Mathematics, Third edition 317
so the eigenvalues are 3, 3, 0 indicating that the matrix is positive semidenite.
The principal minors of A are
2,

2 1
1 2

= 3, det A = 0
conrming, by Sylvesters condition (a), that the matrix is positive semidenite.
31 The matrix of the quadratic form is A =
_
_
K 1 1
1 K 1
1 1 1
_
_
having principal
minors
K,

K 1
1 K

= K
2
1 and det A = K
2
K 3
Thus by Sylvesters condition (a) the quadratic form is positive denite if and only
if
K
2
1 = (K 1)(K + 1) > 0 and K
2
K 3 = (K 2)(K + 1) > 0
i.e. K > 2 or K < 1.
If K = 2 then det A = 0 and the quadratic form is positive semidenite.
32 Principal minors of the matrix are
3 + a,

3 + a 1
1 a

= a
2
+ 3a 1, det A = a
3
+ 3a
2
6a 8
Thus by Sylvesters condition (a) the quadratic form is positive denite if and only
if
3 + a > 0, a
2
+ 3a 1 > 0 and a
3
+ 3a
2
6a 8 > 0
or (a + 1)(a + 4)(a 2) > 0
3 + a > 0 a > 3
a
2
+ 3a 1 > 0 a < 3.3 or a > 0.3
(a + 1)(a + 4)(a 2) > 0 a > 2 or 4 < a < 1
Thus minimum value of a for which the quadratic form is positive denite is a = 2.
33 A =
_
_
1 2 2
2 3
2 3
_
_
c _
Pearson Education Limited 2004
318 Glyn James: Advanced Modern Engineering Mathematics, Third edition
Principal minors are
1,

1 2
2

= 4, det A =
2
8 + 15 = 0
Thus by Sylvesters condition (a) the quadratic form is positive denite if and only
if
4 > 0 > 4
and ( 5)( 3) > 0 < 3 or > 5
Thus it is positive denite if and only if > 5.
Exercises 6.7.1
34 The characteristic equation of A is

5 6
2 3

=
2
8 + 3 = 0
Now A
2
=
_
5 6
2 3
_ _
5 6
2 3
_
=
_
27 48
16 21
_
so
A
2
8A+ 3I =
_
37 48
16 21
_

_
40 48
16 24
_
+
_
3 0
0 3
_
=
_
0 0
0 0
_
so that A satises its own characteristic equation.
35 The characteristic equation of A is

1 2
1 1

=
2
2 1 = 0
c _
Pearson Education Limited 2004
Glyn James: Advanced Modern Engineering Mathematics, Third edition 319
By CayleyHamilton theorem
A
2
2AI = 0
35(a) follows that A
2
= 2A+I =
_
2 4
2 2
_
+
_
1 0
0 1
_
=
_
3 4
2 3
_
35(b) A
3
= 2A
2
+A =
_
6 8
4 6
_
+
_
1 2
1 1
_
=
_
7 10
5 7
_
35(c) A
4
= 2A
3
+A
2
=
_
14 20
10 14
_
+
_
3 4
2 3
_
=
_
17 24
12 14
_
36(a) The characteristic equation of A is

2 1
1 2

= 0
i.e.
2
4 + 3 = 0
Thus by the CayleyHamilton theorem
A
2
4A+ 3I = 0
I =
1
3
[4AA
2
]
so that A
1
=
1
3
[4I A]
=
1
3
__
4 0
0 4
_

_
2 1
1 2
__
=
1
3
_
2 1
1 2
_
36(b) The characteristic equation of A is

1 1 2
3 1 1
2 3 1

= 0
i.e.
3
3
2
7 11 = 0
c _
Pearson Education Limited 2004
320 Glyn James: Advanced Modern Engineering Mathematics, Third edition
A
2
=
_
_
1 1 2
3 1 1
2 3 1
_
_
_
_
1 1 2
3 1 1
2 3 1
_
_
=
_
_
8 8 5
8 7 8
13 8 8
_
_
Using (6.44)
A
1
=
1
11
(A
2
3A7I)
=
1
11
_
_
2 5 1
1 3 5
7 1 2
_
_
37 A
2
=
_
_
2 3 1
3 1 2
1 2 3
_
_
_
_
2 3 1
3 1 2
1 2 3
_
_
=
_
_
14 11 11
11 14 11
11 11 14
_
_
The characteristic equation of A is

2
6
2
3 + 18 = 0
so by the CayleyHamilton theorem
A
3
= 6A
2
+ 3A18I
giving
A
4
= 6(6A
2
+ 3A18I) + 3A
2
18A = 39A
2
108I
A
5
= 39(6A
2
+ 3A18I) + 108A = 234A
2
+ 9A702I
A
6
= 234(6A
2
+ 3A18I) + 9A
2
702A = 1413A
2
4212I
A
7
= 1413(6A
2
+ 3A18I) + 4212A = 8478A
2
+ 27A25434I
Thus
A
7
3A
6
+A
4
+ 3A
3
2A
2
+ 3I = 4294A
2
+ 36A12957I
=
_
_
47231 47342 47270
47342 47195 47306
47270 47306 47267
_
_
c _
Pearson Education Limited 2004
Glyn James: Advanced Modern Engineering Mathematics, Third edition 321
38(a) Eigenvalues A are = 1 (repeated). Thus
e
At
=
0
I +
1
A with
e
t
=
0
+
1
te
t
=
1
_

1
= te
t
,
0
= (1 t)e
t
so e
At
= (1 t)e
t
I + te
t
A =
_
e
t
0
te
t
e
t
_
38(b) Eigenvalues A are = 1 and = 2. Thus
e
At
=
0
I +
1
A with
e
t
=
0
+
1
e
2t
=
0
+ 2
1
_

0
= 2e
t
e
2t
,
1
= e
2t
e
t
so e
At
= (2e
t
e
2t
)I + (e
2t
e
t
)A =
_
e
t
0
e
2t
e
t
e
2t
_
39 Eigenvalues of A are
1
= ,
2
=

2
,
3
=

2
.
Thus
sin A =
0
A+
1
A+
2
A
2
with
sin = 0 =
0
+
1
+
2

2
sin

2
= 1 =
0
+
1

2
+
2

2
4
cos

2
= 0 =
1
+
2
Solving gives
0
= 0,
1
=
4

,
2
=
4

2
so that
sin A =
4

A
4

2
A
2
=
_
_
0 0 0
0 1 0
0 0 1
_
_
40(a)
dA
dt
=
_
d
dt
(t
2
+ 1)
d
dt
(2t 3)
d
dt
(5 t)
d
dt
(t
2
t + 3)
_
=
_
2t 2
1 2t 1
_
c _
Pearson Education Limited 2004
322 Glyn James: Advanced Modern Engineering Mathematics, Third edition
40(b)
_
2
1
Adt =
_ _
2
1
(t
2
+ 1)dt
_
2
1
(2t 3)dt
_
2
1
(5 t)dt
_
2
1
(t
2
t + 3)dt
_
=
_
_
10
3
0
7
2
23
6
_
_
41
A
2
=
_
t
2
+ 1 t 1
5 0
_ _
t
2
+ 1 t 1
5 0
_
=
_
t
4
+ 2t
2
+ 5t 4 t
3
t
2
+ t 1
5t
2
+ 1 5t 5
_
d
dt
(A
2
) =
_
4t
3
+ 4t + 5 3t
2
2t + 1
10t 5
_
2A
dA
dt
=
_
4t
3
+ 4t 2t
2
+ 1
20t 0
_
Thus
d
dt
(A
2
) ,= 2A
dA
dt
.
Exercises 6.8.3
42(a) Taking x
1
= y
x
1
= x
2
=
dy
dt
x
2
= x
3
=
d
2
y
dt
2
x
3
=
d
3
y
dt
3
= u(t) 4x
1
5x
2
4x
3
Thus state space form is
x =
_
_
x
1
x
2
x
3
_
_
=
_
_
0 1 0
0 0 1
4 5 4
_
_
_
_
x
1
x
2
x
3
_
_
+
_
_
0
0
1
_
_
u(t)
y = x
1
= [1 0 0] [x
1
x
2
x
3
]
T
c _
Pearson Education Limited 2004
Glyn James: Advanced Modern Engineering Mathematics, Third edition 323
42(b)
x
1
= y
x
2
= x
1
=
dy
dt
x
3
= x
2
=
d
2
y
dt
2
x
4
= x
3
=
d
3
y
dt
3
x
4
=
d
4
y
dt
4
= 4x
2
2x
3
+ 5u(t)
Thus state space form is
x =
_

_
x
1
x
2
x
3
x
4
_

_
=
_

_
0 1 0 0
0 0 1 0
0 0 0 1
0 4 2 0
_

_
_

_
x
1
x
2
x
3
x
4
_

_
+
_

_
0
0
0
5
_

_
u(t)
y = x
1
= [1 0 0 0] [x
1
x
2
x
3
x
4
]
T
43(a) Taking A to be the companion matrix
A =
_
_
0 1 0
0 0 1
7 5 6
_
_
then b = [0 0 1]
T
c = [5 3 1]
T
Then from equation (6.49) in the text the state space form of the dynamic model
is
x = A x +bu
y = c
T
x
43(b) Taking A to be the companion matrix
A =
_
_
0 1 0
0 0 1
0 3 4
_
_
then b = [0 0 1]
T
c = [2 3 1]
T
c _
Pearson Education Limited 2004
324 Glyn James: Advanced Modern Engineering Mathematics, Third edition
And state space model is
x = A x +bu, y = c
T
x
44 Applying Kirchhos second law to the individual loops gives
e = R
1
(i
1
+ i
2
) + v
c
+ L
1
di
1
dt
, v
c
=
1
C
(i
1
+ i
2
)
e = R
1
(i
1
+ i
2
) + v
c
+ L
2
di
2
dt
+ R
2
i
2
so that
di
1
dt
=
R
1
L
1
i
1

R
1
L
1
i
2

v
c
L
1
+
e
L
1
di
2
dt
=
R
1
L
2
i
1

(R
1
+ R
2
)
L
2
i
2

v
c
L
2
+
e
L
2
dv
c
dt
=
1
C
(i
1
+ i
2
)
Taking x
1
= i
1
, x
2
= i
2
, x
3
= v
c
, u = e(t) gives the state equation as
_
_
x
1
x
2
x
3
_
_
=
_
_

R
1
L
1

R
2
L
1

1
L
1

R
1
L
2

(R
1
+R
2
)
L
2

1
L
2
1
C
1
C
0
_
_
_
_
x
1
x
2
x
3
_
_
+
_
_
1
L
1
1
L
2
0
_
_
u(t) (1)
The output y = voltage drop across R
2
= R
2
i
2
= R
2
x
2
so that
y = [0 R
2
0] [x
1
x
2
x
3
]
T
(2)
Equations (1) and (2) are then in the required form
x = A x +bu , y = c
T
x
45 The equations of motion, using Newtons second law, may be written down
for the body mass and axle/wheel mass from which a state space model can be
deduced. Alternatively a block diagram for the system, which is more informative
for modelling purposes, may be drawn up as follows
c _
Pearson Education Limited 2004
Glyn James: Advanced Modern Engineering Mathematics, Third edition 325
where s denotes the Laplace s and upper case variables X, Y, Y
1
denote the
corresponding Laplace transforms of the corresponding lower case time domain
variables x(t), y(t), y
1
(t); y
1
(t) is the vertical displacement of the axle/wheel mass.
Using basic block diagram rules this block diagram may be reduced to the
input/output transfer function model
X

K
1
(K + Bs)
(M
1
s
2
+ K
1
)(Ms
2
+ Bs + K) + Ms
2
(K + Bs)
Y

or the time domain dierential equation model


M
1
M
d
4
y
dt
4
+ B(M
1
+ M)
d
3
y
dt
3
+ (K
1
M + KM
1
+ KM)
d
2
y
dt
2
+ K
1
B
dy
dt
+ K
1
Ky = K
1
K
2
x + K
1
B
dx
dt
A possible state space model is
_

_
z
1
z
2
z
3
z
4
_

_
=
_

_
B(M
1
+ M) 1 0 0
(K
1
M+KM
1
+KM)
MM
1
0 1 0
K
1
B
M
1
M
0 0 1
K
1
K
M
1
M
0 0 0
_

_
_

_
z
1
z
2
z
3
z
4
_

_
+
_

_
0
0
K
1
B
M
1
M
K
1
K
2
MM
1
_

_
x(t)
y = [1 0 0 0]z(t), z = [z
1
z
2
z
3
z
4
]
T
.
Clearly alternative forms may be written down, such as, for example, the
companion form of equation (6.52) in the text. Disadvantage is that its output
y is not one of the state variables (see Exercise 43).
c _
Pearson Education Limited 2004
326 Glyn James: Advanced Modern Engineering Mathematics, Third edition
46 Applying Kirchhos second law to the rst loop gives
x
1
+ R
3
(i i
1
) + R
1
i = u
i.e. (R
1
+ R
3
)i R
3
i
1
+ x
1
= u
Applying it to the outer loop gives
x
2
+ (R
4
+ R
2
)i
1
+ R
1
i = u
Taking = R
1
R
3
+ (R
1
+ R
3
)(R
4
+ R
2
) then gives
i = (R
2
+ R
3
+ R
4
)u (R
2
+ R
4
)x
1
R
3
x
2
and i
1
= R
3
u + R
1
x
1
(R
1
+ R
3
)x
2
Thus
(i i
1
) = (R
4
+ R
2
)u (R
1
+ R
2
+ R
4
)x
1
+ R
1
x
2
Voltage drop across C
1
: x
1
=
1
C
1
(i i
1
)
=
1
C
1
[(R
1
+ R
2
+ R
4
)x
1
+ R
1
x
2
+ (R
4
+ R
2
)u](1)
Voltage drop across C
2
: x
2
=
1
C
2
i
1
=
1
C
2
[R
1
x
1
(R
1
+ R
3
)x
2
+ R
3
u] (2)
y
1
= i
1
=
R
1

x
(R
1
+ R
3
)

x
2
+
R
3

u (3)
y
2
= R
2
(i i
1
) =
R
3

(R
1
+ R
2
+ R
4
)x
1
+
R
3
R
1

x
2
+ R
3
(R
4
+ R
2
)

u (4)
Equations (1) - (4) give the required state space model.
Substituting the given values for R
1
, R
2
, R
3
, R
4
, C
1
and C
2
gives the state matrix
A as
A =
_
_
9
35.10
3
1
35.10
3
1
35.10
3
4
35.10
3
_
_
=
10
3
35
_
9 1
1 4
_
c _
Pearson Education Limited 2004
Glyn James: Advanced Modern Engineering Mathematics, Third edition 327
Let =
10
3
35
then eigenvalues are solutions of

9
4

=
2
+ 13 + 35
2
= 0
giving
=
13

29
2
2.6 10
2
or 1.1 10
2
47 We are required to express the transfer function in the state space form
x = A x +bu
y = c
T
x
where A is the companion matrix A =
_
_
0 1 0
0 0 1
6 11 6
_
_
and y = [1 0 0]x. To
determine b we divide the denominator into the numerator as follows
s
3
+ 6s
2
+ 11s + 6
5s
1
29s
2
+ 120s
3
5s
2
+ s + 1
5s
2
+ 30s + 55
29s 54
[
[
neglect these terms
29s 174
120
[
[
giving b = [5 29 120]
T
. Thus state space form is
x(t) =
_
_
x
1
x
2
x
3
_
_
=
_
_
0 1 0
0 0 1
6 11 6
_
_
x(t) +
_
_
5
29
120
_
_
u(t)
y = [1 0 0]x(t)
It is readily checked that this is a true representation of the given transfer function.
c _
Pearson Education Limited 2004
328 Glyn James: Advanced Modern Engineering Mathematics, Third edition
Exercises 6.9.4
48 (t) = e
At
where A =
_
1 0
1 1
_
Eigenvalues of A are = 1, = 1 so
e
At
=
0
(t)I +
1
(t)A
where
0
,
1
satisfy
e
t
=
0
+
1
, = 1
te
t
=
1
giving
1
= te
t
,
0
= e
t
te
t
Thus
(t) = e
At
=
_
e
t
te
t
0
0 e
t
te
t
_
+
_
te
t
0
te
t
te
t
_
=
_
e
t
0
te
t
e
t
_
48(a) (0) =
_
1 0
0 1
_
= I
48(b)
(t
2
t
1
) (t
1
) =
_
e
t
2
e
t
1
0
(t
2
t
1
)e
t
2
e
t
1
e
t
2
e
t
1
_ _
e
t
1
0
t
1
e
t
1
e
t
1
_
=
_
e
t
2
0
(t
2
t
1
)e
t
2
+ t
1
e
t
2
e
t
2
_
=
_
e
t
2
0
t
2
e
t
2
e
t
2
_
= (t
2
)
48(c)
1
=
1
e
2t
_
e
t
0
te
t
e
t
_
=
_
e
t
0
te
t
e
t
_
= (t)
49 Take x
1
= y, x
2
= x
1
=
dy
dt
, x
2
=
d
2
y
dt
2
= x
1
2x
2
so in vector-matrix
form the dierential equation is
x =
_
0 1
1 2
_
x, y = [1 0]A
c _
Pearson Education Limited 2004
Glyn James: Advanced Modern Engineering Mathematics, Third edition 329
Taking A =
_
0 1
1 2
_
its eigenvalues are = 1, = 1
e
At
=
0
I +
1
A where
0
,
1
satisfy
e
t
=
0
+
1
, = 1
te
t
=
1
giving
0
= e
t
+ te
t
,
1
= te
t
. Thus
e
At
=
_
e
t
+ te
t
te
t
te
t
e
t
te
t
_
Thus solution of dierential equation is
x(t) = e
At
x(0), x(0) = [1 1]
T
=
_
e
t
+ 2te
t
e
t
2te
t
_
giving y(t) = x
1
(t) = e
t
+ 2te
t
The dierential equation may be solved directly using the techniques of Chapter 10
of the companion text Modern Engineering Mathematics or using Laplace
transforms. Both approaches conrm the solution
y = (1 + 2t)e
2t
50 Taking A =
_
1 0
1 1
_
then from Exercise 44
e
At
=
_
e
t
0
te
t
e
t
_
and the required solution is
x(t) = e
At
x(0) =
_
e
t
0
te
t
e
t
_ _
1
1
_
=
_
e
t
(1 + t)e
t
_
c _
Pearson Education Limited 2004
330 Glyn James: Advanced Modern Engineering Mathematics, Third edition
51 Taking A =
_
0 1
6 5
_
its eigenvalues are
1
= 3,
2
= 2.
Thus e
At
=
0
I +
1
A where
0
,
1
satisfy
e
3t
=
0
3
1
, e
2t
=
0
2
1

0
= 3e
2t
2e
3t
,
1
= e
2t
e
3t
so
e
At
=
_
3e
2t
2e
3t
e
2t
e
3t
6e
3t
te
2t
3e
3t
2e
2t
_
Thus the rst term in (6.73) becomes
e
At
x(0) = e
At
[1 1]
T
=
_
2e
2t
e
3t
3e
3t
4e
2t
_
and the second term is
_
t
0
e
A(t)
bu()d =
_
t
0
2
_
6e
2(t)
6e
3(t)
18e
3(t)
12e
2(t)
_
d
= 2
_
3e
2(t)
2e
3(t)
6e
3(t)
6e
2(t)
_
t
0
= 2
_
1 3e
2t
+ 2e
3t
6e
2t
6e
3t
_
Thus required solution is
x(t) =
_
2e
2t
e
3t
+ 2 6e
3t
+ 4e
3t
3e
3t
4e
2t
+ 12e
2t
12e
3t
_
=
_
2 4e
2t
+ 3e
3t
8e
2t
9e
3t
_
i.e. x
1
= 2 4e
2t
+ 3e
3t
, x
2
= 8e
2t
9e
3t
52 In state space form
x =
_
0 1
2 3
_
x +
_
2
0
_
u(t), u(t) = e
t
, x(0) = [0 1]
T
Taking A =
_
0 1
2 3
_
its eigenvalues are
1
= 2,
2
= 1 so
e
At
=
0
I +
1
A where
0
,
1
satisfy
e
2t
=
0
2
1
, e
t
=
0

1

0
= 2e
t
e
2t
,
1
= e
t
e
2t
c _
Pearson Education Limited 2004
Glyn James: Advanced Modern Engineering Mathematics, Third edition 331
Thus
e
At
=
_
2e
t
e
2t
e
t
e
2t
2e
t
+ 2e
2t
e
t
+ 2e
2t
_
and e
At
x(0) =
_
e
t
e
2t
e
t
+ 2e
2t
_
_
t
0
A
(t)
bu()d =
_
t
0
_
4e
(t)
2e
2(t)
4e
(t)
+ 4e
2(t)
_
e

d
=
_
t
0
_
4e
t
2e
2t
e

4e
t
+ 4e
2t
e

_
d
=
_
4e
t
2e
2t
e

4e
t
+ 4e
2t
e

_
t
0
=
_
4te
t
2e
t
+ 2e
2t
4te
t
+ 4e
t
4e
2t
_
We therefore have the solution
x(t) = e
At
x(0) +
_
t
0
e
A(t)
bu()d
=
_
4te
t
+ e
2t
e
t
4te
t
+ 3e
t
2e
2t
_
that is
x
1
= 4te
t
+ e
2t
e
t
, x
2
= 4te
t
+ 3e
t
2e
2t
53 Taking A =
_
3 4
2 1
_
its eigenvalues are
1
= 5,
2
= 1.
e
At
=
0
I +
1
A where
0
,
1
satisfy
e
5t
=
0
+ 5
1
, e
t
=
0

1

0
=
1
6
e
5t
+
5
6
e
t
,
1
=
1
6
e
5t
+
1
6
e
t
Thus transition matrix is
e
At
=
_
1
3
e
t
+
2
3
e
5t 2
3
e
5t

2
3
e
t
1
3
e
5t

1
3
e
t 1
3
e
5t
+
2
3
e
t
_
and e
At
x(0) = e
At
[1 2]
T
=
_
2e
5t
e
t
e
5t
+ e
t
_
c _
Pearson Education Limited 2004
332 Glyn James: Advanced Modern Engineering Mathematics, Third edition
_
t
0
e
A(t)
Bu()d =
_
t
0
e
A(t)
_
0 1
1 1
_ _
4
3
_
d
=
_
t
0
A
t
_
3
7
_
d
=
_
t
0
_
20
3
e
5(t)

11
3
e
(t)
10
3
e
5(t)
+
11
3
e
(t)
_
d
=
_

4
3
e
5(t)

11
3
e
(t)

2
3
e
5(t)
+
11
3
e
(t)
_
t
0
=
_
5 +
11
3
e
t
+
4
3
e
5t
3
11
3
e
t
+
2
3
e
5t
_
Thus solution is
x(t) = e
At
x(0) +
_
t
0
e
A(t)
Bu(t)d
=
_
5 +
8
3
e
t
+
10
3
e
5t
3
8
3
e
t
+
5
3
e
5t
_
Exercises 6.9.6
54 [sI A] =
_
s 3 4
2 s 1
_
, det[sI A] = (s 5)(s + 1)
Thus
[sI A]
1
=
1
(s 5)(s + 1)
_
s 1 4
2 s 3
_
=
_
2/3
s5
+
1/3
s+1
2/3
s5

2/3
s+1

1/3
s5

1/3
s+1
1/3
s5
+
2/3
s+1
_
/
1
[sI A]
1
=
_
2
3
e
5t
+
1
3
e
t 2
3
e
5t

2
3
e
t
1
3
e
5t

1
3
e
t 1
3
e
5t
+
2
3
e
t
_
=
1
[sI A]
1
B U(s) =
1
(s 5)(s + 1)
_
s 1 4
2 5 3
_ _
0 1
1 1
_ _
4
s
7
s
_
=
1
s(s 5)(s + 1)
_
3s + 25
7s 15
_
=
_

5
s
+
11/3
s+1
+
4/3
s5
3
s

11/3
s+1
+
2/3
s5
_
/
1
[sI A]
1
B U(s) =
_
5 +
11
3
e
t
+
4
3
e
5t
3
11
3
e
t
+
2
3
e
5t
_
=
2
c _
Pearson Education Limited 2004
Glyn James: Advanced Modern Engineering Mathematics, Third edition 333
Thus solution is
x(t) =
1
x(0) +
2
=
_
5 +
8
3
e
t
+
10
3
e
5t
3
8
3
e
t
+
5
3
e
5t
_
which conrms the answer obtained in Exercise 53.
55 A =
_
1 3
2 4
_
, sI A =
_
s 1 3
2 s + 4
_
[ sI A [ = = (s 1)(s + 4) + 6 = (s + 2)(s + 1)
[sI A]
1
=
1

_
s + 4 3
2 s 1
_
=
_
3
s+1

2
s+2

3
s+1
+
3
s+2
2
s+1

2
s+2

2
s+1
+
3
s+2
_
giving /
1
[sI A]
1
=
_
3e
t
2e
2t
3e
t
+ 3e
2t
2e
t
2e
2t
2e
t
+ 3e
2t
_
so /
1
[sI A]
1
x(0) =
_
e
2t
e
2t
_
[sI A]
1
bU(s) =
1

_
s + 4 3
2 s 1
_ _
1
1
_
1
s + 3
=
_
1
(s+2)(s+3)
1
(s+2)(s+3)
_
=
_
1
s+2

1
s+3
1
s+2

1
s+3
_
so /
1
[sI A]
1
bU(s) =
_
e
2t
e
3t
e
2t
e
3t
_
Thus
x(t) =
_
e
2t
e
2t
_
+
_
e
2t
e
3t
e
2t
e
3t
_
giving
x
1
(t) = x
2
(t) = 2e
2t
e
3t
56 A =
_
0 2
1 3
_
, B =
_
1 1
1 1
_
, u =
_
1
t
_
, x
0
=
_
0
1
_
The solution x(t) is given by equation (6.75) in the text.
(sI A) =
_
s 2
1 s + 3
_
giving (sI A)
1
=
1
(s + 1)(s + 2)
_
s + 3 2
1 s
_
c _
Pearson Education Limited 2004
334 Glyn James: Advanced Modern Engineering Mathematics, Third edition
so that
(sI A)
1
x
0
=
1
(s + 1)(s + 2)
_
2
s
_
=
_
2
s+2

2
s+1
2
s+2

1
s+1
_
(sI A)
1
B U(s) =
1
(s + 1)(s + 2)
_
s + 3 2
1 s
_ _
1 1
1 1
_
_
_
1
s
1
s
2
_
_
=
_
1
s(s+2)

(s+5)
s
2
(s+1)(s+2)
1
s(s+2)
+
s1
s
2
(s+1)(s+2)
_
Thus
X
1
(s) = /x
1
(t) =
2
s + 2

2
s + 1
+
1
s(s + 2)

(s + 5)
s
2
(s + 1)(s + 2)
=
9/4
s + 2

6
s + 1
+
15/4
s

5/2
s
2
giving x
1
(t) =
9
4
e
2t
6e
t
+
15
4

5
2
t
57 A =
_
0 1
2 3
_
, b =
_
2
0
_
, u(t) = e
t
H(t), x
0
= [1 0]
T
with X(s) and the solution x(t) given by equations (6.74) and (6.75) in the text
(sI A) =
_
s 1
2 s + 3
_
giving (sI A)
1
=
1
(s + 1)(s + 2)
_
s + 3 1
2 5
_
so that
(sI A)
1
x
0
=
1
(s + 1)(s + 2)
_
s + 3
2
_
=
_
2
s+1

1
s+2

2
s+1
+
2
s+2
_
and
(sI A)
1
bU(s) =
1
(s + 1)(s + 2)
_
2(s + 3)
4
_
1
s + 1
=
_
2
s+2
+
4
(s+1)
2

2
s+1
4
s+1

4
(s+1)
2

4
s+2
_
Thus X(s) =
_
4
(s+1)
2
+
1
(s+2)
2
(s+1)

2
(s+2)

4
(s+1)
2
_
giving x(t) =
_
4te
t
+ e
2t
2e
t
2e
2t
4te
t
_
c _
Pearson Education Limited 2004
Glyn James: Advanced Modern Engineering Mathematics, Third edition 335
Exercises 6.9.9
58 Eigenvalues of matrix A =
_

3
2
3
4
1
5
2
_
are given by
[ AI [=
2
+ 4 + 3 = ( + 3)( + 1) = 0
that is
1
= 1,
2
= 3
having corresponding eigenvectors e
1
= [3 2]
T
, e
2
= [1 2]
T
.
Denoting the reciprocal basis vectors by
r
1
= [r
11
r
12
]
T
, r
2
= [r
21
r
22
]
T
and using the relationships r
T
i
e
j
=
ij
(i, j = 1, 2) we have
3r
11
+ 2r
12
= 1
r
11
2r
12
= 0
_
r
1
= [
1
4
1
8
]
T
3r
21
+ 2r
22
= 0
r
21
2r
22
= 1
_
r
2
= [
1
4

3
8
]
T
Thus
r
T
1
x(0) =
1
2
+
1
2
= 1, r
T
2
x(0) =
1
2

3
2
= 1
so the spectral form of solution is
x(t) = e
t
e
1
e
3t
e
2
The trajectory is readily drawn showing that it approaches the origin along the
eigenvector e
1
since e
3t
0 faster than e
t
. See Figure 6.11 in the text.
59 Taking A =
_
2 2
2 5
_
eigenvalues are
1
= 6,
2
= 1 having
corresponding eigenvectors e
1
= [1 2]
T
, e
2
= [2 1]
T
.
Denoting the reciprocal basis vectors by
r
1
= [r
11
r
12
]
T
, r
2
= [r
21
r
22
]
T
c _
Pearson Education Limited 2004
336 Glyn James: Advanced Modern Engineering Mathematics, Third edition
and using the relationships r
T
i
e
j
=
ij
(i, j = 1, 2) we have
r
11
2r
12
= 1
2r
11
+ r
12
= 0
_
r
11
=
1
5
, r
12
=
2
5
r
1
=
1
5
[1 2]
T
r
21
2r
22
= 0
2r
21
+ r
22
= 1
_
r
21
=
2
5
, r
22
=
1
5
r
2
=
1
5
[2 1]
T
Thus
r
T
1
x(0) =
1
5
[1 2]
_
2
3
_
=
4
5
r
T
2
x(0) =
1
5
[2 1]
_
2
3
_
=
7
5
then response is
x(t) =
2

i=1
r
T
i
x(0)e
it
e
i
=
4
5
e
6t
_
1
2
_
+
7
5
e
t
_
2
1
_
=
1
5
_
4e
6t
+ 14e
t
8e
6t
+ 7e
t
_
Again, following Figure 6.11 in the text, the trajectory is readily drawn and showing
that it approaches the origin along the eigenvector e
2
since e
6t
0 faster than
e
t
.
60 Taking A =
_
0 4
2 4
_
eigenvalues are
1
= 2 + j2,
2
= 2 j2 having
corresponding eigenvectors e
1
= [2 1 j]
T
, e
2
= [2 1 + j]
T
.
Let r
1
= r

1
+ jr

1
be reciprocal base vector to e
1
then
r
T
1
e
1
= 1 = [r

+ jr

1
]
T
[e

1
+ je

1
]
T
where e
1
= e

1
+ je

1
r
T
1
e
2
= 0 = [r

1
+ jr

1
]
T
[e

1
je

1
]
T
since e
2
= conjugate e
1
Thus
[(r

1
)
T
e

1
(r

1
)
T
e

1
] + j[(r

1
)
T
e

1
+ (r

1
)
T
e

1
] = 1
and
[(r

1
)
T
e

1
(r

1
)
T
e

1
] + j[(r

1
)
T
e
1
1
(r

1
)
T
e

1
] = 0
c _
Pearson Education Limited 2004
Glyn James: Advanced Modern Engineering Mathematics, Third edition 337
giving
(r

1
)
T
e

1
=
1
2
, (r

1
)
T
e

1
=
1
2
, (r

1
)
T
e

1
= (r

1
)
T
e

1
= 0
Now e

1
= [2 1]
T
, e

1
= [0 1]
T
Let r

1
= [a b]
T
and r

1
= [c d]
T
then from above
2a + b =
1
2
, b = 0 and d =
1
2
, 2c + d = 0
giving a =
1
4
, b = 0, c =
1
4
, d =
1
2
so that
r
1
= r

1
+ jr

1
=
1
4
[1 j 2j]
T
Since r
2
is the complex conjugate of r
1
r
2
=
1
4
[1 + j 2j]
T
so the solution is given by
x(t) = r
T
1
x(0)e

1
t
e
1
+r
T
2
x(0)e

2
t
e
2
and since r
T
1
x(0) =
1
2
(1 + j), r
T
2
x(0) =
1
2
(1 j)
x(t) = e
2t
_
1
2
(1 + j)e
2jt
_
2
1 j
_
+
1
2
(1 j)e
2jt
_
2
1 + j
__
= e
2t
_
(cos 2t sin 2t)
_
2
1
_
(cos 2t + sin 2t)
_
0
1
__
= e
2t
_
(cos 2t sin 2t)e

1
(cos 2t + sin 2t)e

1
_
where e
1
= e

1
+ je

1
To plot the trajectory rst plot e

1
, e

1
in the plane and then using these as a frame
of reference plot the trajectory. A sketch is as follows
c _
Pearson Education Limited 2004
338 Glyn James: Advanced Modern Engineering Mathematics, Third edition
61 Following section 6.9.8 if the equations are representative of
x = A x +bu , y = c
T
x
then making the substitution x = M , where M is the modal matrix of A,
reduces the system to the canonical form

= + (M
1
b)u , y = (c
T
M)
where is the spectral matrix of A.
Eigenvalues of A are given by

1 1 2
1 2 1
0 1 1

=
3
2
2
+ 2 = ( 1)( + 2)( + 1) = 0
so the eigenvalues are
1
= 2,
2
= 1,
3
= 1. The corresponding eigenvectors
are readily determined as
e
1
= [1 3 1]
T
, e
2
= [3 2 1]
T
, e
3
= [1 0 1]
T
Thus M =
_
_
1 3 1
3 2 0
1 1 1
_
_
and =
_
_
2 0 0
0 1 0
0 0 1
_
_
c _
Pearson Education Limited 2004
Glyn James: Advanced Modern Engineering Mathematics, Third edition 339
M
1
=
1
det M
adj M =
1
6
_
_
2 2 2
3 0 3
1 2 7
_
_
so required canonical form is
_
_

3
_
_
=
_
_
2 0 0
0 1 0
0 0 1
_
_
_
_

3
_
_
+
_
_
1
3
0

4
3
_
_
u
y = [1 4 2] [
1

2

3
]
T
62 Let r
1
= [r
11
r
12
r
13
]
T
, r
2
= [r
21
r
22
r
23
]
T
, r
3
= [r
31
r
32
r
33
]
T
be the
reciprocal base vectors to e
1
= [1 1 0]
T
, e
2
= [0 1 1]
T
, e
3
= [1 2 3]
T
.
r
T
1
e
1
= r
11
+ r
12
= 1
r
T
1
e
2
= r
11
+ r
13
= 0
r
T
1
e
3
= r
11
+ 2r
12
+ 3r
13
= 0
_
_
_
r
1
=
1
2
[1 1 1]
T
r
T
2
e
1
= r
21
+ r
22
= 0
r
T
2
e
2
= r
22
+ r
23
= 1
r
T
2
e
3
= r
21
+ 2r
22
+ 3r
23
= 0
_
_
_
r
2
=
1
2
[3 3 1]
T
r
T
3
e
1
= r
31
+ r
32
= 0
r
T
3
e
2
= r
32
+ r
33
= 0
r
T
3
e
3
= r
31
+ 2r
32
+ 3r
33
= 1
_
_
_
r
3
=
1
2
[1 1 1]
T
Then using the fact that x(0) = [1 1 1]
T

0
= r
T
1
x(0) =
1
2
,
1
= r
T
2
x(0) =
1
2
,
3
= r
T
3
x(0) =
1
2
63 The eigenvectors of A are given by

5 4
1 2

= ( 6)( 1) = 0
so the eigenvalues are
1
= 6,
2
= 1. The corresponding eigenvectors are readily
determined as e
1
= [4 1]
T
, e
2
= [1 1]
T
.
Taking M to be the modal matrix M =
_
4 1
1 1
_
then substituting x = M
into x = Ax(t) reduces it to the canonical form

=
c _
Pearson Education Limited 2004
340 Glyn James: Advanced Modern Engineering Mathematics, Third edition
where =
_
6 0
0 1
_
. Thus the decoupled canonical form is
_

2
_
=
_
6 0
0 1
_ _

2
_
or

1
= 6
1
and

2
=
2
which may be individually solved to give

1
= e
6t
and
1
= e
t
Now (0) = M
1
x(0) =
1
5
_
1 1
1 4
_ _
1
4
_
=
_
1
3
_
so
1
(0) = 1 = and
2
(0) = 3 =
giving the solution of the uncoupled system as
=
_
e
6t
3e
t
_
The solution for x(t) as
x = M =
_
4 1
1 1
_ _
e
6t
3e
t
_
=
_
4e
6t
3e
t
e
6t
+ 3e
t
_
64 Taking A =
_
3 4
2 1
_
its eigenvalues are
1
= 5,
2
= 1 having
corresponding eigenvectors e
1
= [2 1]
T
, e
2
= [1 1]
T
.
Let M =
_
2 1
1 1
_
be the modal matrix of A, then x = M reduces the
equation to

(t) =
_
5 0
0 1
_
+M
1
_
0 1
1 1
_
u(t)
Since M
1
=
1
det M
adj M =
1
3
_
1 1
1 2
_
we have

(t) =
_
5 0
0 1
_
+
1
3
_
1 2
2 1
_
u(t)
With u(t) = [4 3]
T
the decoupled equations are

1
= 5
1
+
10
3

2
=
2

11
3
c _
Pearson Education Limited 2004
Glyn James: Advanced Modern Engineering Mathematics, Third edition 341
which can be solved independently to give

1
= e
5t

2
3
,
2
= e
t

11
3
We have that (0) = MM M
1
x(0) =
1
3
_
1 1
1 2
_ _
1
2
_
=
_
1
1
_
so
1 =
2
3
=
5
3
1 =
11
3
=
8
3
giving
=
_
5
3
e
5t

2
3
8
3
e
t

11
3
_
and x = MM M =
_
2 1
1 1
_ _
5
3
e
5t

2
3
8
3
e
t

11
3
_
=
_
5 +
8
3
e
t
+
10
3
e
5t
3
8
3
e
t
+
5
3
e
5t
_
which conrms Exercises 49 and 50.
65 Taking A as the companion matrix following the procedure of Example 6.33
we have
A =
_
_
0 1 0
0 0 1
6 11 6
_
_
, b = [0 0 1]
T
, c = [1 2 3]
T
Eigenvalue of A given by

1 0
0 1
6 11 6

= (
3
+ 6
2
+ 11 + 6) = ( + 1)( + 2)( + 3) = 0
so the eigenvalues are
1
= 3,
2
= 2,
3
= 1. The eigenvectors are given by
the corresponding solutions of

i
e
i1
+ e
i2
+ 0e
i3
= 0
0e
i1

i
e
i2
+ e
i3
= 0
6e
i1
11e
i2
(6 +
i
)e
i3
= 0
Taking i = 1, 2, 3 and solving gives the eigenvectors as
e
1
= [1 3 9]
T
, e
2
= [1 2 4]
T
, e
3
= [1 1 1]
T
c _
Pearson Education Limited 2004
342 Glyn James: Advanced Modern Engineering Mathematics, Third edition
Taking M to be the modal matrix M =
_
_
1 1 1
3 2 1
9 4 1
_
_
then the transformation
X = M will reduce the system to the canonical form

= +M
1
bu , y = c
T
M
M
1
=
1
2
_
_
2 3 1
6 8 2
6 5 1
_
_
, M
1
b =
1
2
_
_
1
2
1
_
_
, c
T
M = [22 9 2]
Thus canonical form is
_
_

3
_
_
=
_
_
3 0 0
0 2 0
0 0 1
_
_
_
_

3
_
_
+
_
_
1
2
1
1
2
_
_
u
y = [22 9 2] [
1

2

3
]
T
Since the eigenvalues of A are negative the system is stable. Since vector M
1
b
has no zero elements the system is controllable and since c
T
M has no zero elements
the system is also observable.
b = [0 0 1]
T
, A b = [0 1 6]
T
, A
2
b = [1 6 25]
T
[b A b A
2
b] =
_
_
0 0 1
0 1 6
1 6 25
_
_

_
_
0 0 1
0 1 0
1 0 0
_
_
which is of full rank 3
so the system is controllable.
c = [1 2 3]
T
, A
T
c = [18 32 16], (A
T
)
2
c = [96 158 63]
T
[c A
T
c (A
T
)
2
c] =
_
_
1 18 96
2 32 158
3 16 63
_
_

_
_
1 0 0
0 1 0
0 0 1
_
_
which is of full rank 3
so the system is observable.
c _
Pearson Education Limited 2004
Glyn James: Advanced Modern Engineering Mathematics, Third edition 343
66 A =
_
_
0 1 0
0 0 1
0 5 6
_
_
, b = [0 0 1]
T
, c = [5 3 1]
T
Eigenvalues of A given by
0 =

1 0
0 1
0 5 6

= ( + 5)( + 1) = 0
so eigenvalues are
1
= 5,
2
= 1,
3
= 0. The corresponding eigenvectors are
determined as
e
1
= [1 5 25]
T
, e
2
= [1 1 1] , e
3
= [1 0 0]
T
Take M to be the modal matrix M =
_
_
1 1 1
5 1 0
25 1 0
_
_
then the transformation
x = M will reduce the system to the canonical form

= +M
1
bu , y = c
T
M
M
1
=
1
20
_
_
0 1 1
0 25 5
20 24 4
_
_
, M
1
b =
1
20
_
_
1
5
4
_
_
, c
T
M = [15 3 5]
Thus canonical form is
_
_

3
_
_
=
_
_
5 0 0
0 1 0
0 0 0
_
_
_
_

3
_
_
+
_
_
1
20

1
4
1
5
_
_
u
y = [15 3 5] [
1

2

3
]
T
Since A has zero eigenvalues the system is marginally stable, since M
1
b has no
zero elements the system is controllable and since c
T
M has no zero elements the
system is observable. Again as in Exercise 61 these results can be conrmed by
using the Kalman matrices.
c _
Pearson Education Limited 2004
344 Glyn James: Advanced Modern Engineering Mathematics, Third edition
Exercises 6.10.3
67(a)
zI A =
_
z 1
4 z
_
(zI A)
1
=
1
(z 2)(z + 2)
_
z 1
4 z
_
=
_
1/2
z2
+
1/2
z+2
1/4
z2

1/4
z+2
1
z2

1
z+2
1/2
z2
+
1/2
z+2
_
A
k
= Z
1
z(zI A)
1
= Z
1
_
1
2
1/2
z2
+
1
2
1/2
z+2
1
4
z
z2

1
4
z
z+2
z
z2

z
z+2
1
2
z
z2
+
z
z+2
_
=
1
4
_
2
k
_
2 1
4 2
_
+ (2)
k
_
2 1
4 2
__
67(b)
zI A =
_
z + 1 3
3 z + 1
_
(zI A)
1
=
1
(z + 4)(z 2)
_
z + 1 3
3 z + 1
_
=
_
1/2
z+4
+
1/2
z2
1/2
z+4
+
1/2
z2

1/2
z+4
+
1/2
z2
1/2
z2
+
1/2
z+2
_
A
k
= Z
1
z(zI A)
1
= Z
1
_
1
2
z
z+4
+
1
2
z
z2

1
2
z
z+4
+
1
2
z
z2

1
2
z
z+4
+
1
2
z
z2
1
2
z
z2
+
1
2
z
z2
_
=
1
2
_
(4)
k
_
1 1
1 1
_
+ (2)
k
_
1 1
1 1
__
67(c)
zI A =
_
z + 1 1
0 z + 1
_
(zI A)
1
=
1
(z + 1)
2
_
z + 1 1
0 z + 1
_
=
_
1
z+1
1
(z+1)
2
0
1
z+1
_
A
k
= Z
1
z(zI A)
1
= Z
1
_
z
z+1
z
(z+1)
2
0
z
z+1
_
= (1)
k
_
1 k
0 1
_
c _
Pearson Education Limited 2004
Glyn James: Advanced Modern Engineering Mathematics, Third edition 345
68 Taking x
1
= x and x
2
= y we can express equations in the form
x(k + 1) =
_
7 4
8 1
_
x(k) with x(0) = [1 2]
The solution is given by
x(k + 1) = A
k
x(0), A =
_
7 4
8 1
_
where A
k
=
1
I +
1
A. That is, the solution is
x(k) =
_

0
7
1
4
1
8
1

0
+
1
_ _
1
2
_
=
_

0
+
1
2
0
6
1
_
The eigenvalues of A are given by

2
6 + 25 = 0 so = 3 j4
or, in polar form,
1
= 5e
j
,
2
= 5e
j
where = cos
1
(
3
5
).
Thus
0
and
1
are given by
5
k
e
jk
=
0
+
1
5e
j
, 5
k
e
jk
=
0
+
1
5e
j
which are readily solved to give

0
=
(5)
k
sin(k 1)
sin
,
1
=
1
5
(5)
k
sin k
sin
Then
0
+
1
=
(5)
k
sin
_
1
5
sin k sin(k 1)
_
=
5
4
(5)
k
_
1
5
sin k +
3
5
sin k +
4
5
cos k
_
= (5)
k
[sin k + cos k]
2
0
6
1
=
(5)
k
sin
_
2(sin k cos cos k sin )
6
5
sin k
_
= (5)
k
[2 cos(k)]
c _
Pearson Education Limited 2004
346 Glyn James: Advanced Modern Engineering Mathematics, Third edition
Thus solution is
x(k) = (5)
k
[sin k + cos k]
y(k) = (5)
k
[2 cos(k)]
We have
x(1) = 1 , y(1) = 6 , x(2) = 31 , y(2) = 14
69 A =
_
0 1
0.16 1
_
zI A =
_
z 1
0.16 z + 1
_
(zI A)
1
=
_
z+1
(z+0.2)(z+0.8)
1
(z+0.2)(z+0.8)
0.16
(z+0.2)(z+0.8)
z
(z+0.2)(z+0.8)
_
=
_
4
3

1
z+0.2

1
3

1
z+0.8
5
3

1
z+0.2

5
3

1
z+0.8
0.8
3
1
z+0.2
+
0.8
3
1
z+0.8

1
3
1
z+0.2
+
4
3

1
z+0.8
_
A
k
= Z
1
z(zI A)
1
=
_
4
3
(0.2)
k

1
3
(0.8)
k 5
3
(0.2)
k

5
3
(0.8)
k
0.8
3
(0.2)
k
+
0.8
3
(0.8)
k

1
3
(0.2)
k
+
4
3
(0.8)
k
_
A
k
x(0) = A
k
[1 1]
T
=
_

1
3
(0.2)
k
+
4
3
(0.8)
k
0.2
3
(0.2)
k

3.2
3
(0.8)
k
_
U(z) = Zu(k) = z/(z 1)
(zI A)
1
bU(z) =
1
(z + 0.2)(z + 0.8)
_
z + 1 1
0.16 z
_ _
1
1
_
z
z 1
=
z
(z + 0.2)(z + 0.8)(z 1)
_
z + 2
z 0.16
_
=
_

5
2
z
z+0.2
+
10
9
z
z+0.8
+
25
18
z
z1
1
2
z
z+0.2

8
9
z
z+0.8
+
7
18
z
z1
_
Z
1
(zI A)
1
bU(z) =
_

5
2
(0.2)
k
+
10
9
(0.8)
k
+
25
18
1
2
(0.2)
k

8
9
(0.8)
k
+
7
18
_
Thus solution is
x(k) = A
k
x(0) + Z
1
(zI A)
1
bU(t)
=
_

17
6
(0.2)
k
+
22
9
(0.8)
k
+
25
18
3.4
6
(0.2)
k

17.6
9
(0.8)
k
+
7
18
_
c _
Pearson Education Limited 2004
Glyn James: Advanced Modern Engineering Mathematics, Third edition 347
70 Let x
1
(k) = y(k), x
2
(k) = x
1
(k +1) = y(k +1) then the dierence equation
may be written
x(k + 1) =
_
x
1
(k + 1)
x
2
(k + 1)
_
=
_
0 1
1 1
_ _
x
1
(k)
x
2
(k)
_
, x(0) = [0 1]
T
Taking A =
_
0 1
1 1
_
its eigenvalues are
1
=
1 +

5
2
,
2
=
1

5
2
A
k
=
1
I +
1
A where
0
and
1
satisfy
_
1 +

5
2
_
k
=
0
+
_
1 +

5
2
_

1
,
_
1

5
2
_
k
=
0
+
_
1

5
2
_

1
giving

1
=
1

5
_
_
1 +

5
2
_
k

_
1

5
2
_
k_

0
=
1

5
_
_
1 +

5
2
_
k
_

5 1
2
_
+
_
1

5
2
_
k
_
1 +

5
2
_
_
Solution to the dierence equation is
x(k) =
_
y(k)
y(k + 1)
_
=
_

0

1

1

0
+
1
_ _
0
1
_
=
_

1

0
+
1
_
so y(k) =
1
=
1

5
_
_
1 +

5
2
_
k

_
1

5
2
_
k_
[Note that y(k + 1) =
0
+
1
=
1

5
_
_
1 +

5
2
_
k

_
1

5
2
_
k+1_
using above
values.]
As k ,
_
1

5
2
_
k
0 and y(k + 1)/y(k)
(
1+

5
2
)
k+1
(
1+

5
2
)
k
=
1
2
(

5 + 1)
Exercises 6.12.3
71 Choose x
1
(t) = y(t), x
2
(t) = x
1
(t) =
dy
dt
then
x(t) =
_
0 1
1
2
1
2
_
x(t) +
_
0
1
_
u(t), y(t) = [1 0]x(t)
c _
Pearson Education Limited 2004
348 Glyn James: Advanced Modern Engineering Mathematics, Third edition
Taking u(t) = K
1
x
1
(t) + K
2
x
2
(t) + u
ext
(t)
x(t) =
_
0 1
K
1
+
1
2
K
2
+
1
2
_
x(t) +
_
0
1
_
u
ext
The eigenvalues of the matrix are given by

0 1
K
1
+
1
2
K
2
+
1
2

= 0
or
2
(K
2
+
1
2
) (K
1
+
1
2
) = 0
If the poles are to be at = 4 then we require the characteristic equation to be

2
+ 8 + 16 = 0
By comparison we have K
2

1
2
= 8 and K
1

1
2
= 18 giving K
1
=
33
2
,
K
2
=
17
2
so u(t) =
_

33
2

17
2

x + u
ext
72
x(t) =
_
0 1

1
4

5
4
_
x(t) +
_
0
1
_
u(t) , y(t) = [0 2]x(t)
Setting u = K
T
x + u
ext
, K = [K
1
K
2
]
T
gives the system matrix
A =
_
0 1
K
1

1
4
K
2

5
4
_
whose eigenvalues are given by
2
(K
2

5
4
) (K
1

5
4
) = 0. Comparing with
the desired characteristic equation
( + 5)
2
=
2
+ 10 + 25 = 0
gives K
1
=
99
4
, K
2
=
35
4
. Thus
u(t) =
_

99
4

35
4

x(t) + u
ext
73 With u
1
= [K
1
K
2
]x(t) the system matrix A becomes
A =
_
K
1
1 + K
2
6 + K
1
1 + K
2
_
c _
Pearson Education Limited 2004
Glyn James: Advanced Modern Engineering Mathematics, Third edition 349
having characteristic equation

2
(1 + K
1
+ K
2
) 6(1 K
2
) = 0
which on comparing with

2
+ 10 + 25 = 0
gives K
1
=
35
6
, K
2
=
31
6
so that u
1
(t) =
_

35
6

31
6

x(t)
Using u
2
(t) the matrix A becomes
_
0 1
6 + K
1
1 + K
2
_
where K
T
= [31 11]
74 See p. 524 in the text.
75 For the matrix of Exercise 74 the Kalman matrix is
M =
_
2 2
1 1
_

_
1 0
0 0
_
which is of rank 1. Thus the system is uncontrollable.
For the matrix of Exercise 67 the Kalman matrix is
M =
_
0 1
1
1
2
_

_
0 1
1 0
_
which is of full rank 2. Thus the system is controllable.
76
M = (b A b) =
_
1 0
4 1
_
, M
1
=
_
1 0
4 1
_
, v
T
= [4 1]
v
T
A = [4 1]
_
8 2
35 9
_
= [3 1] so T =
_
4 1
3 1
_
, T
1
=
_
1 1
3 4
_
Taking z(t) = Tx or x = T
1
z(t) then equation reduces to
T
1
z(t) =
_
8 2
35 9
_
T
1
z(t) +
_
1
4
_
u
or z(t) = I
_
8 2
35 9
_
T
1
z(t) +T
_
1
4
_
u
=
_
4 1
3 1
_ _
8 2
35 9
_ _
1 1
3 4
_
z(t) +
_
4 1
3 1
_ _
1
4
_
u
=
_
0 1
2 1
_
z(t) +
_
0
1
_
u
c _
Pearson Education Limited 2004
350 Glyn James: Advanced Modern Engineering Mathematics, Third edition
Clearly both system matrices have the same eigenvalues = 2, = 1. This will
always be so since we have carried out a singularity transformation.
Review Exercises 6.13
1(a) Eigenvalues given by

1 6 12
0 13 30
0 9 20

= (1 + )[(13 )(20 ) + 270] = 0


i.e. (1 + )( 5)( 2) = 0
so eigenvalues are
1
= 5,
2
= 2,
3
= 1
Eigenvectors are given by corresponding solutions of
_
_
1
i
6 12
0 13
i
30
0 9 20
i
_
_
_
_
e
i1
e
i2
e
i3
_
_
= 0
When i = 1,
i
= 5 and solution given by
e
11
198
=
e
12
90
=
e
13
54
=
1
so e
1
= [11 5 3]
T
When i = 2,
i
= 2 and solution given by
e
21
216
=
e
22
54
=
e
23
27
=
2
so e
2
= [8 2 1]
T
When i = 3,
i
= 1 and solution given by
e
31
1
=
e
32
0
=
e
33
0
=
3
so e
3
= [1 0 0]
T
1(b) Eigenvalues given by

2 0 1
1 4 1
1 2 0

4 1
2

1 4
1 2

= 0
c _
Pearson Education Limited 2004
Glyn James: Advanced Modern Engineering Mathematics, Third edition 351
i.e. 0 = (2 )[(4 )() + 2] + [2 + (4 )]
= (2 )(
2
4 + 3) = (2 )( 3)( 1) = 0
so eigenvalues are

1
= 3,
2
= 2,
3
= 1
Eigenvectors are given by the corresponding solutions of
(2
i
)e
i1
+ 0e
i2
+ e
i3
= 0
e
i1
+ (4
i
)e
i2
e
i3
= 0
e
i1
+ 2e
i2

i
e
i3
= 0
Taking i = 1, 2, 3 gives the eigenvectors as
e
1
= [1 2 1]
T
, e
2
= [2 1 0]
T
, e
3
= [1 0 1]
T
1(c) Eigenvalues given by

1 1 0
1 2 1
0 1 1

R
1
+ (R
2
+ R
3
)


1 2 1
0 1 1

= 0
i.e.

1 1 1
1 2 1
0 1 1

1 0 0
1 3 0
0 1 1

= (3 )(1 ) = 0
so eigenvalues are
1
= 3,
2
= 1,
3
= 0
Eigenvalues are given by the corresponding solutions of
(1
i
)e
i1
e
i2
0e
i3
= 0
e
i1
+ (2
i
)e
i2
e
i3
= 0
0e
i1
e
i2
+ (1
i
)e
i3
= 0
Taking i = 1, 2, 3 gives the eigenvectors as
e
1
= [1 2 1]
T
, e = [1 0 1]
T
, e
3
= [1 1 1]
T
c _
Pearson Education Limited 2004
352 Glyn James: Advanced Modern Engineering Mathematics, Third edition
2 Principal stress values (eigenvalues) given by

3 2 1
2 3 1
1 1 4

R
1
+ (R
2
+ R
3
)

6 6 6
2 3 1
1 1 4

= (6 )

1 1 1
2 3 1
1 1 4

= 0
i.e. (6 )

1 0 0
2 1 1
1 0 3

= (6 )(1 )(3 ) = 0
so the principal stress values are
1
= 6,
2
= 3,
3
= 1.
Corresponding principal stress direction e
1
, e
2
and e
3
are given by the solutions
of
(3
i
)e
i1
+ 2e
i2
+ e
i3
= 0
2e
i1
+ (3
i
)e
i2
+ e
i3
= 0
e
i1
+ e
i2
+ (4
i
)e
i3
= 0
Taking i = 1, 2, 3 gives the principal stress direction as
e
1
= [1 1 1]
T
, e
2
= [1 1 2]
T
, e
3
= [1 1 0]
T
It is readily shown that e
T
1
e
2
= e
T
1
e
3
= e
T
2
e
3
= 0 so that the principal stress
directions are mutually orthogonal.
3 Since [1 0 1]
T
is an eigenvector of A
_
_
2 1 0
1 3 b
0 b c
_
_
_
_
1
0
1
_
_
=
_
_
1
0
1
_
_
so 2 = , 1 + b = 0, c =
giving b = 1 and c = 2.
Taking these values A has eigenvalues given by

2 1 0
1 3 1
0 1 2

= (2 )

3 1
1 2

(2 )
= (2 )( 1)( 4) = 0
c _
Pearson Education Limited 2004
Glyn James: Advanced Modern Engineering Mathematics, Third edition 353
i.e. eigenvalues are
1
= 4,
2
= 2,
3
= 1
Corresponding eigenvalues are given by the solutions of
(2
i
)e
i1
e
i2
+ 0e
i3
= 0
e
i1
+ (3
i
)e
i2
+ e
i3
= 0
0e
i1
+ e
i2
+ (2
i
)e
i3
= 0
Taking i = 1, 2, 3 gives the eigenvectors as
e
1
= [1 2 1]
T
, e
2
= [1 0 1]
T
, e
3
= [1 1 1]
T
4 The three Gerschgorin circles are
[ 4 [ =[ 1 [ + [ 0 [= 1
[ 4 [ =[ 1 [ + [ 1 [= 2
[ 4 [ = 1
Thus [ 4 [ 1 and [ 4 [ 2 so [ 4 [ 2 or 2 6.
Taking x
(o)
= [1 1 1]
T
iterations using the power method may be tabulated
as follows
Iteration k 0 1 2 3 4 5 6
-1 -0.833 -0.765 -0.734 -0.720 -0.713 -0.710
x
(k)
1 1 1 1 1 1 1
-1 -0.833 -0.765 -0.734 -0.720 -0.713 -0.710
-5 -4.332 -4.060 -3.936 -3.88 -3.852
A x
(k)
6 5.666 5.530 5.468 5.44 5.426
-5 -4.332 -4.060 -3.936 3.88 -3.852
6 5.666 5.530 5.468 5.44 5.426
Thus correct to one decimal place the dominant eigenvalue is = 5.4
c _
Pearson Education Limited 2004
354 Glyn James: Advanced Modern Engineering Mathematics, Third edition
5(a) Taking xx x
(o)
= [1 1 1]
7
iterations may be tabulated as follows
Iteration k 0 1 2 3 4 5 6 7
1 0.800 0.745 0.728 0.722 0.720 0.719 0.719
x
(k)
1 0.900 0.862 0.847 0.841 0.838 0.837 0.837
1 1 1 1 1 1 1 1
4 3.500 3.352 3.303 3.285 3.278 3.275
A x
(k)
4.5 4.050 3.900 3.846 3.825 3.815 3.812
5 4.700 4.607 4.575 4.563 4.558 4.556
5 4.700 4.607 4.575 4.563 4.558 4.556
Thus estimate of dominant eigenvalues is 4.56 with associated eigenvector
x = [0.72 0.84 1]
T
5(b)

3
i=1

i
= trace A 7.5 = 4.56 + 1.19 +
3

3
= 1.75
5(c) (i) det A =
3

i=1

i
= 9.50 so A
1
exists and has eigenvalues
1
1.19
,
1
1.75
,
1
4.56
so power method will generate the eigenvalue 1.19 corresponding to A.
(ii) A3I has eigenvalues
1.19 3, 1.75 3, 4.56 3
i.e. 1.91, 1.25, 1.56
so applying the power method on A3I generates the eigenvalues corresponding
to 1.75 of A.
6 x = e
t
, y = e
t
, z = e
t
so the dierential equations become
e
t
= 4e
t
+ e
t
+ e
t
e
t
= 2e
t
+ 5e
t
+ 4e
t
e
t
= e
t
e
t
c _
Pearson Education Limited 2004
Glyn James: Advanced Modern Engineering Mathematics, Third edition 355
Provided e
t
,= 0 (i.e. non-trivial solution) we have the eigenvalue problem
_
_
4 1 1
2 5 4
1 1 0
_
_
_
_

_
_
=
_
_

_
_
Eigenvalues given by

4 1 1
2 5 4
1 1 0

C
2
C
3

4 0 1
2 1 4
1 1

= ( 1)

4 0 1
2 1 4
1 1

= ( 1)( 5)( 3)
so its eigenvalues are 5, 3 and 1.
When = 1 the corresponding eigenvector is given by
3e
11
+ e
12
+ e
13
= 0
2e
11
+ 4e
12
+ 4e
13
= 0
e
11
e
12
e
13
= 0
having solution
e
11
0
=
e
12
2
=
e
13
2
=
1
Thus corresponding eigenvector is [0 1 1]
T
7 Eigenvalues are given by
[ AI [ =

8 8 2
4 3 2
3 4 1

= 0
Row 1 - (Row 2 + Row 3) gives
[ AI [ =

1 1 + 1 +
4 3 2
3 4 1

= (1 )

1 1 1
4 3 2
3 4 1

= (1 )

1 0 0
4 1 2
3 1 4

= (1 )[(1 )(4 ) + 2]
= (1 )( 2)( 3)
Thus eigenvalues are
1
= 3,
2
= 2,
3
= 1.
c _
Pearson Education Limited 2004
356 Glyn James: Advanced Modern Engineering Mathematics, Third edition
Corresponding eigenvectors are given by
(8 )e
i1
8e
i2
2e
i3
= 0
4e
i1
(3 + )e
i2
2e
i3
= 0
3e
i1
4e
i2
+ (1 )e
i3
= 0
When i = 1,
i
=
1
= 3 and solution given by
e
11
4
=
e
12
2
=
e
13
2
=
1
so a corresponding eigenvector is e
1
= [2 1 1]
T
.
When i = 2,
i
=
2
= 2 and solution given by
e
21
3
=
e
22
2
=
e
23
1
=
2
so a corresponding eigenvector is e
2
= [3 2 1]
T
.
When i = 3,
i
=
3
= 1 and solution given by
e
31
8
=
e
32
6
=
e
33
4
=
3
so a corresponding eigenvector is e
3
= [4 3 2]
T
.
Corresponding modal and spectral matrices are
M =
_
_
2 3 4
1 2 3
1 1 2
_
_
and =
_
_
3 0 0
0 2 0
1 0 1
_
_
M
1
=
_
_
1 2 1
1 0 2
1 1 1
_
_
and matrix multiplication conrms M
1
A M =
8 Eigenvectors of A are given by

1 0 4
0 5 4
4 4 3

= 0
i.e.
3
9
2
9 + 81 = ( 9)( 3)( + 3) = 0
c _
Pearson Education Limited 2004
Glyn James: Advanced Modern Engineering Mathematics, Third edition 357
so the eigenvalues are
1
= 9,
2
= 3 and
3
= 3.
The eigenvectors are given by the corresponding solutions of
(1
i
)e
i1
+ 0e
i2
4e
i3
= 0
0e
i1
+ (5
i
)e
i2
+ 4e
i3
= 0
4e
i1
+ 4e
i2
+ (3
i
)e
i3
= 0
Taking i = 1, 2, 3 the normalised eigenvectors are given by
e
1
= [
1
3
2
3
2
3
]
T
, e
2
= [
2
3
2
3
1
3
]
T
, e
3
= [
2
3
1
3
2
3
]
T
The normalised modal matrix

M =
1
3
_
_
1 2 2
2 2 1
2 1 2
_
_
so

M
T
A

M =
1
9
_
_
1 2 2
2 2 1
2 1 2
_
_
_
_
1 0 4
0 5 4
4 4 3
_
_
_
_
1 2 2
2 2 1
2 1 2
_
_
=
_
_
9 0 0
0 3 0
0 0 3
_
_
=
9

N =
_

_
6 0 0 0
6 4 0 0
0 4 2 0
0 0 2 0
_

_
N, N = [N
1
N
2
N
3
N
4
]
T
Since the matrix A is a triangular matrix its eigenvalues are the diagonal elements.
Thus, the eigenvalues are

1
= 6,
2
= 4,
3
= 2,
4
= 0
The eigenvectors are the corresponding solutions of
(6
i
)e
i1
+ 0e
i2
+ 0e
i3
+ 0e
i4
= 0
6e
i1
+ (4
i
)e
i2
+ 0e
i3
+ 0e
i4
= 0
0e
i1
+ 4e
i2
+ (2
i
)e
i3
+ 0e
i4
= 0
0e
i1
+ 0e
i2
+ 2e
i3

i
e
i4
= 0
c _
Pearson Education Limited 2004
358 Glyn James: Advanced Modern Engineering Mathematics, Third edition
Taking i = 1, 2, 3, 4 and solving gives the eigenvectors as
e
1
= [1 33 1]
T
, e
2
= [0 1 2 1]
T
e
3
= [0 0 1 1]
T
, e
4
= [0 0 0 1]
T
Thus spectral form of solution to the equation is
N = e
6t
e
1
+ e
4t
e
2
+ e
2t
e
3
+ e
4
Using the given initial conditions at t = 0 we have
_

_
C
0
0
0
_

_
=
_

_
1
3
3
1
_

_
+
_

_
0
1
2
1
_

_
+
_

_
0
0
1
1
_

_
+
_

_
0
0
0
1
_

_
so C = , 0 = 3 + , 0 = 3 2 + , 0 = + +
which may be solved for , , and to give
= C, = 3C, = 3C, = C
Hence
N
4
= e
6t
+ e
4t
e
2t
+
= Ce
6t
+ 3Ce
4t
3Ce
2t
+ C
10(a)
(i) Characteristic equation of A is
2
3 + 2 = 0 so by the CayleyHamilton
theorem
A
2
= 3A2I =
_
4 0
3 1
_
A
3
= 3(3A2I) 2A = 7A6I =
_
8 0
7 1
_
A
4
= 7(3A2I) 6A = 15A14I =
_
16 0
15 1
_
A
5
= 15(3A2I[) 14A = 31A30I =
_
32 0
31 1
_
A
6
= 31(3A2I) 30A = 63A62I =
_
64 0
63 1
_
A
7
= 63(3A2I) 62A = 127A126I =
_
128 0
127 1
_
c _
Pearson Education Limited 2004
Glyn James: Advanced Modern Engineering Mathematics, Third edition 359
Thus A
7
3A
6
+A
4
+ 3A
3
2A
2
+ 3I =
_
29 0
32 3
_
(ii) Eigenvalues of A are
1
= 2,
2
= 1. Thus
A
k
=
0
I +
1
A where
0
and
1
satisfy
2
k
=
0
+ 2
1
, 1 =
0
+
1

1
= 2
k
1,
0
= 2 2
k
Thus A
k
=
_

0
+ 2
1
0

1

0
+
1
_
=
_
2
k
0
2
k
1 1
_
10(b) Eigenvalues of A are
1
= 2,
2
= 0. Thus
e
At
=
0
I +
1
A where
0
and
1
satisfy
e
2t
=
0
2
1
, 1 =
0

0
= 1,
1
=
1
2
(1 e
2t
)
Thus e
At
=
_

0

1
0
0
2
1
_
=
_
1
1
2
(1 e
2t
)
0 e
2t
_
11 The matrix A =
_
_
1 2 3
0 1 4
0 0 1
_
_
has the single eigenvalue = 1 (multiplicity 3)
(AI) =
_
_
0 2 3
0 0 4
0 0 0
_
_

_
_
0 1 0
0 0 1
0 0 0
_
_
is of rank 2 so has nullity 3-2=1 indicating
that there is only one eigenvector corresponding to = 1.
This is readily determined as
e
1
= [1 0 0]
T
The corresponding Jordan canonical form comprises a single block so
J =
_
_
1 1 0
0 1 1
0 0 1
_
_
c _
Pearson Education Limited 2004
360 Glyn James: Advanced Modern Engineering Mathematics, Third edition
Taking T = A I the triad of vectors (including generalised eigenvectors) has
the form T
2
, T , with T
2
= e
1
. Since T
2
=
_
_
0 0 8
0 0 0
0 0 0
_
_
we may take
= [0 0
1
8
]
T
. Then T = [
2
8
1
8
0]
T
. Thus the triad of vectors is
e
1
= [1 0 0]
T
, e

1
= [
3
8
1
2
0]
T
, e

1
= [0 0
1
8
]
T
The corresponding modal matrix is
M =
_
_
1
3
8
0
0
1
2
0
0 0
1
8
_
_
M
1
= 16
_
_
1
16

3
64
0
0
1
8
0
0 0
1
2
_
_
and by matrix multiplication
M
1
A M = 16
_
_
1
16

3
64
0
0
1
8
0
0 0
1
2
_
_
_
_
1 2 3
0 1 4
0 0 1
_
_
_
_
1
3
8
0
0
1
2
0
0 0
1
8
_
_
=
_
_
1 1 0
0 1 1
0 0 1
_
_
= J
12 Substituting x = X cos t, y = Y cos t, z = Z cos t gives

2
X = 2X + Y

2
Y = X 2Y + Z

2
Z = Y 2Z
or taking =
2
( 2)X + Y = 0
X + ( 2)Y + Z = 0
Y + ( 2)Z = 0
For non-trivial solution

2 1 0
1 2 1
0 1 2

= 0
c _
Pearson Education Limited 2004
Glyn James: Advanced Modern Engineering Mathematics, Third edition 361
i.e. ( 2)[( 2)
2
1] ( 2) = 0
( 2)(
2
4 + 2) = 0
so = 2 or = 2

2
When = 2 , Y = 0 and X = Z so X : Y : Z = 1 : 0 : 1
When = 2 +

2 , X = Z and Y =

2X so X : Y : Z = 1 :

2 : 1
When = 2

2 , X = Z and Y =

2X so X : Y : Z = 1 :

2 : 1
13 In each section A denotes the matrix of the quadratic form.
13(a) A =
_
_
2 1 0
1 1 1
0 1 2
_
_
has principal minors of 2,

2 1
1 1

= 1 and
det A = 0
so by Sylvesters condition (c) the quadratic form is positive semidenite.
13(b) A =
_
_
3 2 2
2 7 0
2 0 2
_
_
has principal minors of 3,

3 2
2 7

= 17 and
det A = 6
so by Sylvesters condition (a) the quadratic form is positive denite.
13(c) A =
_
_
16 16 16
16 36 8
16 8 17
_
_
has principal minors of 16,

16 16
16 36

= 320 and
det A = 704
so none of Sylvesters conditions are satised and the quadratic form is indenite.
13(d) A =
_
_
21 15 6
15 11 4
6 4 2
_
_
has principal minors of -21,

21 15
15 11

= 6
and det A = 0
so by Sylvesters condition (d) the quadratic form is negative semidenite.
13(e) A =
_
_
1 1 1
1 3 1
1 1 5
_
_
has principal minors of -1,

1 1
1 3

= 2 and
det A = 4 so by Sylvesters condition (b) the quadratic form is negative denite.
c _
Pearson Education Limited 2004
362 Glyn James: Advanced Modern Engineering Mathematics, Third edition
14 A e
1
=
_
_
7
2

1
2

1
2
4 1 0

3
2
3
2
1
2
_
_
_
_
1
2
3
_
_
=
_
_
1
2
3
_
_
Hence e
1
= [1 2 3]
T
is an eigenvector with
1
= 1 the corresponding eigenvalue.
Eigenvalues are given by
0 =

7
2

1
2

1
2
4 1 0

3
2
3
2
1
2

=
3
+ 3
2
+ 3
= ( 1)(
2
+ 2 + 3)
= ( 1)( 3)( + 1)
so the other two eigenvalues are
2
= 3,
3
= 1.
Corresponding eigenvectors are the solutions of
(
7
2

i
)e
i1

1
2
e
i2

1
2
e
i3
= 0
4e
i1
(1 +
i
)e
i2
+ 0e
i3
= 0

3
2
e
i1
+
3
2
e
i2
+ (
1
2

i
)e
i3
= 0
Taking i = 2, 3 gives the eigenvectors as
e
2
= [1 1 0]
T
, e
3
= [0 1 1]
T
The dierential equations can be written in the vector-matrix form
x = A x , x = [x y z]
T
so, in special form, the general solution is
x = e

1
t
e
1
+ e

2
t
e
2
+ e

3
t
e
3
= e
t
_
_
1
2
3
_
_
+ e
3t
_
_
1
1
0
_
_
+ e
t
_
_
0
1
1
_
_
With x(0) = 2, y(0) = 4, z(0) = 6 we have
= 2, = 0, = 0
c _
Pearson Education Limited 2004
Glyn James: Advanced Modern Engineering Mathematics, Third edition 363
so
x = 2e
t
_
_
1
2
3
_
_
i.e. x = 2e
t
, y = 4e
t
, z = 6e
t
.
15(a)
15(b)
/e
At
= [sI A]
1
=
1
(s + 2)(s + 3)
_
s + 5 6
1 8
_
=
_
3
s+2

2
s+3
6
s+2

6
s+3
1
s+3

1
s+2
2
s+2

1
s+3
_
Taking inverse transforms gives
e
At
=
_
3e
2t
2e
3t
6e
2t
6e
3t
e
3t
e
2t
3e
3t
2e
2t
_
15(c) Taking Laplace transforms
[sI A]X(s) = x(0) +bU(s) ; Y (s) = c
T
X(s)
With x(0) = 0 and U(s) = 1 the transform X

(s) of the impulse response is


X

(s) = [sI A]
1
b , Y

(s) = c
T
[sI A]
1
b
c _
Pearson Education Limited 2004
364 Glyn James: Advanced Modern Engineering Mathematics, Third edition
Inverting then gives the impulse response as
y

(t) = [1 1]
_
6e
2t
6e
3t
3e
3t
2e
2t
_
= 4e
2t
3e
3t
, t 0
With x(0) = [1 0]
T
and X(s) =
1
s
Y (s) = [1 1]
_
[sI A]
1
_
1
0
_
+ [sI A]
1
_
0
1
_
1
s
_
= [1 1]
_
3
s+2

2
s+3
+ 6
_
1/6
s

1/2
s+2
+
1/3
s+3
_
1
s+3

1
s+2
+
_
1
s+2

1
s+3
_
_
so y(t) = [1 1]
_
3e
2t
2e
3t
+ 1 3e
2t
+ 2e
3t
e
3t
e
2t
+ e
2t
e
3t
_
i.e. y(t) = 1, t 0
16 Eigenvalues are given by
0 =

1 1 2
1 2 1
0 1 1

= (1 )(
2
3) + (1 )
= (1 )( 2)( + 1)
so eigenvalues are
1
= 2,
2
= 1,
3
= 1
Corresponding eigenvectors are the solutions of
(1
i
)e
i1
+ e
i2
2e
i3
= 0
e
i1
+ (2
i
)e
i2
+ e
i3
= 0
0e
i1
+ e
i2
(1 +
i
)e
i3
= 0
Taking i = 1, 2, 3 the eigenvectors are
e
1
= [1 3 1]
T
, e
2
= [3 2 1]
T
, e
3
= [1 0 1]
T
c _
Pearson Education Limited 2004
Glyn James: Advanced Modern Engineering Mathematics, Third edition 365
The modal matrix M and spectral matrix are
M =
_
_
1 3 1
3 2 0
1 1 1
_
_
, =
_
_
2 0 0
0 1 0
0 0 1
_
_
M =
_
_
2 3 1
6 2 0
2 1 1
_
_
, A M =
_
_
1 1 2
1 2 1
0 1 1
_
_
_
_
1 3 1
3 2 0
1 1 1
_
_
=
_
_
2 3 1
6 2 0
2 1 1
_
_
= M
Substituting x = M y in x(k + 1) = A x(k) gives
M y(k + 1) = A M y(k)
or y(k + 1) = M
1
A M y
(k)
= y(k)
so y(1) = y(0), y(2) = y(1) =
2
y(0)
y(k) =
k
y(0)
Thus
x(k) =
_
_
1 3 1
3 2 0
1 1 1
_
_
_
_
2k 0 0
0 1 0
0 0 (1)
k
_
_
_
_

_
_
say
=
_
_
1 3 1
3 2 0
1 1 1
_
_
_
_
2
k

(1)
k
_
_
=
_
_
2
k
+ 3 + (1)
k
32
k
+ 2
2
k
+ + (1)
k
_
_
When k = 0
_
_
1
0
0
_
_
=
_
_
+ 3 +
3 + 2
+ +
_
_
which gives =
1
3
, =
1
2
, =
1
6
so that
x(k) =
_
_

1
3
(2)
k
+
3
2
+
1
6
(1)
k
(2)
k
+ 1

1
3
(2)
k
+
1
2
+
1
6
(1)
k
_
_
c _
Pearson Education Limited 2004
366 Glyn James: Advanced Modern Engineering Mathematics, Third edition
17 x = A x +bu , y = c
T
x
Let
i
, e
i
, i = 1, 2, . . . , n, be the eigenvalues and corresponding eigenvectors of A.
Let M = [e
1
, e
2
, . . . , e
n
] then since
i
s are distinct the e
i
s are linearly
independent and M
1
exists. Substituting x = M gives
M

= A M +bu
Premultiplying by M
1
gives

= M
1
A M +M
1
bu = +b
1
u
where = M
1
A M = (
i

ij
), i, j = 1, 2, . . . , n, and b
1
= M
1
b
Also y = c
T
x y = c
T
M = c
T
1
, c
T
1
= c
T
M. Thus we have the desired
canonical form.
If the vector b
1
contains a zero element then the corresponding mode is
uncontrollable and consequently (A
1
b
1
c) is uncontrollable. If the matrix c
T
has a zero element then the system is unobservable.
The eigenvalues of A are
1
= 2,
2
= 1,
3
= 1 having corresponding
eigenvectors e
1
= [1 3 1]
T
, e
2
= [3 2 1]
T
and e
3
= [1 0 1]
T
.
The modal matrix
M = [e
1
e
2
e
3
] =
_
_
1 3 1
3 2 0
1 1 1
_
_
with M
1
=
1
6
_
_
2 2 2
3 0 3
1 2 7
_
_
so canonical form is
_
_

3
_
_
=
_
_
2 0 0
0 1 0
0 0 1
_
_
_
_

3
_
_
+
_
_
1
3
0

4
3
_
_
u
y = [1 4 2][
1

2

3
]
T
We observe that the system is uncontrollable but observable. Since the system
matrix A has positive eigenvalues the system is unstable. Using Kelman matrices
(i) A
2
=
_
_
0 1 1
3 4 3
1 1 2
_
_
, A b =
_
_
2
2
2
_
_
, A
2
b =
_
_
0
4
0
_
_
c _
Pearson Education Limited 2004
Glyn James: Advanced Modern Engineering Mathematics, Third edition 367
Thus [b A b A
2
b] =
_
_
1 2 0
1 2 4
1 2 0
_
_

_
_
1 0 0
0 1 0
0 0 0
_
_
and is of rank 2
so the system is uncontrollable.
(ii) [c A
T
c (A
T
)
2
c] =
_
_
2 3 3
1 0 2
0 5 1
_
_

_
_
0 0 1
1 0 0
0 1 0
_
_
and is of full rank 3
so the system is observable.
18 /e
At
= [sI A]
1
=
_
s + 2 1
2 s
_
1
=
_
s
(s+1)
2
+1
1
(s+1)
2
+1
2
(s+1)
2
+1
s+2
(s+1)
2
+1
_
Thus e
At
=
_
e
t
(cos t sin t) e
t
sin t
2e
t
sin t e
t
(cos t + sin t)
_
and e
At
x(0) = 0 since x(0) = 0
With U(s) = /u(t) =
1
s
we have
[sI A]
1
bU(s) =
_
s
s
2
+2s+2
1
s
2
+2s+2
2
s
2
+2s+2
s+2
s
2
+2s+2
_ _
1
0
_
1
s
=
_
1
s
2
+2s+2
2
s(s
2
+2s+2)
_
=
_
1
(s+1)
2
+1
1
s

s+2
(s+1)
2
+1
_
so /
1
(sI A)
1
bU(s) =
_
e
t
sin t
1 e
t
(cos t + sin t)
_
Thus x(t) = e
At
x(0) +/
1
(sI A)
1
bU(s)
=
_
e
t
sin t
1 e
t
(cos t + sin t)
_
For the transfer function we have
Y(s) = c X(s) and when x(0) = 0
Y(s) = c[s I A]
1
bU(s) = HU(s)
where H = c[s I A]
1
b
For this system we have H(s) = [1 1]
_
s
s
2
+2s+1
2
s
2
+2s+1
_
=
s + 2
(s + 1)
2
+ 1
c _
Pearson Education Limited 2004
368 Glyn James: Advanced Modern Engineering Mathematics, Third edition
When u(t) = (t), U(s) = 1 and so the impulse response y

(t) is given by
/y

(t) = Y

(s) =
s + 2
(s + 1)
2
+ 1
=
s + 1
(s + 1)
2
+ 1
+
1
(s + 1)
2
+ 1
y

(t) = e
t
(cos t + sin t)
19
x
1
(k + 1) = u(k) 3x
1
(k) 4x
2
(k)
x
2
(k + 1) = 2x
1
(k) x
2
(k)
y(k) = x
1
(k) + x
2
(k)
or in vector-matrix form
x(k + 1) =
_
3 4
2 1
_
x(k) +
_
1
0
_
u(k), y(k) = [1 1]x(k) (1)
D(x) = c[zI A]
1
b =
1
z
2
+ 4z 5
[1 1]
_
z + 1 4
2 z + 3
_ _
1
0
_
=
z + 3
z
2
+ 4z 5
(i) M
c
=
_
1 3
0 2
_
(ii) det M
c
= 2 ,= 0 so M
c
is of rank 2
(iii) M
1
c
=
1
2
_
2 3
0 1
_
=
_
1
3
2
0
1
2
_
so v
T
= [0
1
2
]
(iv) T =
_
0
1
2
1
1
2
_
(v) det T ,= 0 and T
1
= 2
_
1
2
1
2
1 0
_
=
_
1 1
2 0
_
c _
Pearson Education Limited 2004
Glyn James: Advanced Modern Engineering Mathematics, Third edition 369
Substituting z(k) = T x(k) in (1) gives
T
1
z(k + 1) = A T
1
z(k) +bu(k)
or z(k + 1) = T A T
1
z(k) +T bu(k)
=
_
0
1
2
1
1
2
_ _
3 4
2 1
_ _
1 1
2 0
_
z(k) +
_
0
1
2
1
1
2
_ _
1
0
_
u(t)
i.e. z(k + 1) =
_
0 1
5 4
_
z(k) +
_
0
1
_
u(k)
Thus C and b
c
are of the required form with = 5, = 4 which are coecients
in the characteristic polynomial of D(z).
20(a) Eigenvalues of the matrix given by
0 =

5 2 1
3 6 9
1 1 1

C
1
C
2

3 2 1
3 + 6 9
0 1 1

= (3 )

1 2 1
0 8 10
0 1 1

= (3 )(
2
9 + 18) = (3 )( 3)( 6)
so the eigenvalues are
1
= 6,
2
=
3
= 3
When = 3, A3I =
_
_
2 2 1
3 3 9
1 1 2
_
_

_
_
0 0 1
1 0 0
0 0 0
_
_
is of rank 2
so there is only 3-2=1 corresponding eigenvectors.
The eigenvector corresponding to
1
= 6 is readily determined as e
1
= [3 2 1]
T
.
Likewise the single eigenvector corresponding to
2
= 6 is determined as
e
2
= [1 1 0]
T
The generalised eigenvector e

2
determined by
(A2I)e

2
= e
2
or 3e

21
+ 2e

22
e

23
= 1
3e

21
+ 3e

22
9e

23
= 1
e

21
+ e

22
2e

23
= 0
c _
Pearson Education Limited 2004
370 Glyn James: Advanced Modern Engineering Mathematics, Third edition
giving e

2
= [
1
3
1
3
1
3
]
T
.
For convenience we can take the two eigenvectors corresponding to = 3 as
e
2
= [3 3 0]
T
, e

2
= [1 1 1]
T
The corresponding Jordan canonical form being J =
_
_
6 0 0
0 3 1
0 0 3
_
_
20(b) The generalised modal matrix is then
M =
_
_
3 3 1
2 3 1
1 0 1
_
_
A M =
_
_
5 2 1
3 6 9
1 1 1
_
_
_
_
3 3 1
2 3 1
1 0 1
_
_
=
_
_
18 9 6
12 9 0
6 0 3
_
_
M J =
_
_
3 3 1
2 3 1
1 0 1
_
_
_
_
6 0 0
0 3 1
0 0 3
_
_
=
_
_
13 9 6
12 9 0
6 0 3
_
_
so A M = M J
20(c) M
1
=
1
9
_
_
3 3 6
1 2 1
3 3 15
_
_
, e
Jt
=
_
_
e
6t
0 0
0 e
3t
te
3t
0 0 e
3t
_
_
so
x(t) =
1
9
_
_
3 3 1
2 3 1
1 0 1
_
_
_
_
e
6t
0 0
0 e
3t
te
3t
0 0 e
3t
_
_
_
_
3 3 6
1 2 1
3 3 15
_
_
_
_
0
1
0
_
_
=
1
9
_
_
9e
6t
9(1 + t)e
3t
6e
6t
+ (3 + 9t)e
3t
3e
6t
3e
3t
_
_
c _
Pearson Education Limited 2004
Glyn James: Advanced Modern Engineering Mathematics, Third edition 371
21 The controllability question can be answered by either reducing to canonical
form as in section 6.7.8 of the text or by using the Kalman matrix criterion given
in Exercise 61 of the text. Adopting the Kalman matrix approach
A =
_
_
1 2 0
0 1 0
3 3 2
_
_
and
b =
_
_
0
1
0
_
_
, A b =
_
_
2
1
3
_
_
, A
2
b =
_
_
0
1
9
_
_
so the controllability Kalman matrix is
[b A b A
2
b] =
_
_
0 2 0
1 1 1
0 3 9
_
_
= C
Since det C ,= 0, rank C = 3 so the system is controllable.
The eigenvalues of A are given by

1 2 0
0 1 0
3 3 2

= (1 )

(1 + ) 0
3 (2 + )

= (1 )(1 )(2 + ) = 0
so that the eigenvalues are
1
= 2,
2
= 1,
3
= 1. The system is therefore
unstable with
3
= 1 corresponding to the unstable mode. The corresponding
eigenvectors of A are given by
(A
i
I)e
i
= 0
and are readily determined as
e
1
= [0 0 1]
T
e
2
= [1 1 0]
T
e
3
= [1 0 1]
T
To determine the control law to relocate
3
= 1 at -5 we need to determine the
eigenvector v
3
of A
T
corresponding to
3
= 1. This is readily obtained as
v
3
= [1 1 0]
T
c _
Pearson Education Limited 2004
372 Glyn James: Advanced Modern Engineering Mathematics, Third edition
Thus the required control law is
u(t) = Kv
T
3
x(t) = K[1 1 0]
T
x(t)
where K =
p
3

3
v
T
3
b
=
(5) 1
[1 1 0]
_
_
0
1
0
_
_
=
6
1
= 6
So u(t) = 6(x
1
(t) + x
2
(t))
22 Substituting x = e
t
u, where u is a constant vector, in x = A x gives

2
u = A u or (A
2
I)u = 0 (1)
so that there is a non-trivial solution provided
[ A
2
I [= 0 (2)
If
2
1
,
2
2
, . . . ,
2
n
are the solutions of (2) and u
1
, u
2
, . . . , u
n
the corresponding
solutions of (1) dene
M = [u
1
u
2
. . . u
n
] and S = diag (
2
1

2
2
. . .
2
n
)
Applying the transformation x = M q , q = [q
1
q
2
. . . q
n
] gives
M q = A M q
giving q = M
1
A M q provided u
1
, u
2
, . . . , u
n
are linearly independent
so that q = S q since M
1
A M = S
This represents n dierential equations of the form
q
i
=
2
i
q
i
, i = 1, 2, . . . , n
When
2
i
< 0 this has the solution of the form
q
i
= C
i
sin(
i
t +
i
)
where C
i
and
i
are arbitrary constants and
i
= j
i
c _
Pearson Education Limited 2004
Glyn James: Advanced Modern Engineering Mathematics, Third edition 373
The given dierential equations may be written in the vector-matrix form
x =
_
x
1
x
2
_
=
_
3 2
1 2
_ _
x
1
x
2
_
which is of the above form
x = A x
0 =[ A
2
I [ gives (
2
)
2
+ 5(
2
) + 4 = 0 or
2
1
= 1,
2
2
= 4.
Solving the corresponding equation
(A
2
i
I) u
i
= 0
we have that u
1
= [1 1]
T
and u
2
= [2 1]
T
. Thus we take
M =
_
1 2
1 1
_
and S =
_
1 0
0 4
_
The normal modes of the system are given by
_
q
1
q
2
_
=
_
1 0
0 4
_ _
q
1
q
2
_
giving
q
1
(t) = C
1
sin(t +
1
)
1
sin t +
1
cos t
q
2
(t) = C
2
sin(2t +
2
)
2
sin 2t +
2
cos 2t
Since x = M q we have that q(0) = M
1
x(0) =
1
3
_
1 2
1 1
_ _
1
2
_
=
_
5
3

1
3
_
also q(0) = M
1
x(0) so that q
1
(0) = 2 and q
2
(0) = 0
Using these initial conditions we can determine
1
,
1
,
2
and
2
to give
q
1
(t) =
5
3
cos t + 2 sin t
q
2
(t) =
1
3
cos 2t
The general displacements x
1
(t) and x
2
(t) are then given by x = M q so
x
1
= q
1
+ 2q
2
=
5
3
cos t + 2 sin t
2
3
cos 2t
x
2
= q
1
q
2
=
5
3
cos t + 2 sin t
1
3
cos 2t
c _
Pearson Education Limited 2004

Das könnte Ihnen auch gefallen