Beruflich Dokumente
Kultur Dokumente
FUZZY SETS
B. Sc. MATHEMATICS
SIXTH SEMESTER
UNIVERSITY OF CALICUT
SCHOOL OF DISTANCE EDUCATION
CALICUT UNIVERSITY P.O., THENJIPALAM, MALAPPURAM-679 635
UNIVERSITY OF CALICUT
SCHOOL OF DISTANCE EDUCATION
B.Sc. MATHEMATICS
STUDY MATERIAL
SIXTH SEMESTER
ADDITIONAL COURSE IN LIEU OF PROJECT
Santhosh
Assistant Professor
Assistant Professor
P. G.Department of Mathematics
P. G.Department of Mathematics
Govt.BrennenCollege,Thalassery.
Email: vinodunical@gmail.com
Email:
P.K.
Scrutinised by ;
Dr. Anilkumar. V.
Associate Professor,
Department of Mathematics,
University of Calicut.
notes
abridgedversion
versionofofthe
the topics
topics given
given in
in the
the
This This
studystudy
notes
are are
onlyonly
an anabridged
syllabus.
further
details
proofs
of theorems,
illustrations
students
syllabus.
For For
further
details
like like
proofs
of theorems,
illustrations
etc.,etc.,
students
are
are to
advised
to go the
through
the prescribed
Text. Books .
advised
go through
prescribed
Text Books
Reserved
Module-I
THEORY OF EQUATIONS
Prepared by:
Vinod Kumar P.
Asst. Professor
Dept. of Mathematics
T. M. Govt. College, Tirur.
THEORY OF EQUATIONS
1.0 Introduction
In this module, we will study about polynomial functions and various
methods to find out the roots of polynomial equations. Solving equations was an
important problem from the beginning of study of Mathematics itself. The notion of
complex numbers was first introduced because equations like x2 + 1 = 0 has no
solution in the set of real numbers. The fundamental theorem of algebra which
states that every polynomial of degree >1 has at least one zero was first proved by
the famous German Mathematician Karl Fredrich Gauss. We shall look at
polynomials in detail and will discuss various methods for solving polynomial
equations.
a o , a1 ,...., a n
Remark:
Fundamental theorem of algebra says that, if f(x) = a 0 x n + a1 x n1 + ..... + a n ,
1.1.2
Theorem
Every polynomial of degree n has n and only n zeroes.
Proof:
Let f ( x) = a 0 x n + a1 x n 1 + ..... + a n , where a o 0, be a polynomial of degree n > 1.
By fundamental theorem of algebra, f(x) has at least one zero, let 1 be that zero.
Then ( x 1 ) is a factor of f(x).
Therefore, we can write:
f(x) = ( x 1 ) Q1 ( x) , where Q1(x) is a polynomial function of degree n - 1.
If n -1 > 1, again by Fundamental Theorem of Algebra, Q1 (x) has at least one zero,
say 2 .
Therefore, f ( x) = ( x 1 )( x 2 )Q2 ( x) where Q2 ( x) is a polynomial function of
degree n - 2.
Repeating the above arguments, we get
f ( x) = ( x 1 )( x 2 ).....( x n )Qn ( x), where Qn (x) is a polynomial function
Therefore, f ( x) = a o ( x 1 )( x 2 )....( x n ) .
If is any number other than 1 , 2 ,.... n , then f (x) 0 is not a zero of f(x).
Hence f(x) has n and only n zeros, namely 1 , 2 ...., n .
Note:
Let
Then,
Theorem
If the equation a o x n + a1 x n 1 + .... + a n = 0 , where a o , a1 ,....a n are real numbers
(ao 0) , has a complex root + i , then it also has a complex root i . (i.e.,
complex roots occur in conjugate pairs for a polynomial equation with real
coefficients).
Proof:
Let
f(x) = a o x n + a1 x n1 + .... + a n , a o 0
f ( x) = ( x ) 2 + 2 Q( x) + Ax + B
= [( x ( + i ))( x ( i ))]Q( x) + Ax + B
f ( + i) = 0 + A ( + i) + B = A ( + i) + B = (A + B) + iA
But f ( + i) = 0 .
1.1.4. Theorem
In an equation with rational coefficients, the roots which are quadratic surds
occur in conjugate pairs.
Proof:
Let f ( x) = a o x n + a1 x n 1 + ... + a n , a o 0 , be an nth degree polynomial with
rational coefficients.
Let + is a root of f(x) = 0.
1.1.5. Theorem
If the rational number p
Since
a o p + a1 p
q
q
n 1
+ ... + a n 1 p + a n = 0
q
Multiplying by qn , we get
a o p n + a1 p n1 q + .... + a n 1 pq n 1 + a n q n = 0
------------ (1)
Dividing by p, we have
ao p
n 1
+ a1 p
n2
q + ... + a n 1q
an q n
=
p
n 1
Now, the left side of the above equation is an integer and therefore
an q n
is also
p
a1 p
n 1
+ .... + a n 1 pq
n 2
+ an q
n 1
ao p n
=
q
Since the left side is an integer and since q does not divide p, q must be a
divisor of a 0 . This completes the proof.
Multiple Roots
If a root of f(x) = 0 repeats r times, then is called an r-multiple root.
A 2- multiple root is usually called a double root.
For example, consider f(x) = (x - 2)3 (x - 5)2 (x + 1).
Here 2 is a 3 - multiple root, 5 is a double root, and -1 is a single root of the
equation f(x) = 0.
1.1.6. Theorem
If is an r - multiple root of f(x) = 0 then is an (r 1 ) multiple root of
f1 (x) = 0, where f1(x) is the derivative of f(x).
Proof:
Given that is an r - multiple root of f(x) = 0.
Then
f ( x) = ( x ) r ( x) where ( ) 0.
Remark:
If is an (r 1) -multiple root of f1(x) = 0, similarly as above, we can see that
will be an (r 2 ) multiple root of f11 (x) = 0 ; (r 3) - multiple root of f111 (x) = 0,
and so on.
polynomial.
Dividing the given polynomial by this factor, we obtain the other factor as
x2 + 4x + 5 .
4 16 20
= 2 i .
2
Dividing the given polynomial by this factor, we obtain the other factor as x2 3x + 2.
Also, x2 3x + 2 = (x 2) (x 1)
Thus the roots of the given polynomial equation are
1 + 5 ,1 5 ,1, 2 .
Solution:
Since quadratic surds occur in pairs as roots, 3 is also a root.
Since complex roots occur in conjugate pairs, 1 + 2i is also a root of the required
polynomial equation. Therefore the desired equation is given by
( x 3 ) ( x + 3 ) ( x (1 2i) (x (1 + 2i) ) = 0
i.e., x 4 2 x 3 + 2 x 2 + 6 x 15 = 0
4. Solve 4x5 + x3 + x2 3x + 1 = 0, given that it has rational roots.
Solution:
Let
f(x) = 4x5 + x3 + x2 3x + 1.
6
f(x)
= (x )2 (x + 1) (4x2 + 4)
= 4 (x )2 (x + 1) (x2 + 1)
f(x) = x3 x2 8x + 12
Differentiating, we obtain:
f1(x) = 3x2 2x 8.
Since the multiple roots of f(x) = 0 are also the roots of f1(x) = 0, the product of the
factors corresponding to these roots will be the g.c.d of f(x) and f1(x). Let us find the
g.c.d of f(x) and f1(x).
3x
3x2 2x 8
x3 x2 8x + 12
3x2 6x
4
4x 8
4x 8
3x3 2x2 8x
- x2 16x + 36
3
- 3x2 48x + 108 - 1
- 3x2 + 2x + 8
-50
50x + 100
x2
Therefore, g.c.d = (x 2)
f(x) has a factor (x 2)2.
7
f(x) = x3 + qx + r
-------------- (1)
f ( x) = a o ( x 1 )( x 2 )....( x n )
a
a
x n + 1 x n1 + .... + n
ao
ao
= ( x 1 ) ( x 2 )....( x n )
= x n S1 x n 1 + S 2 x n 2 .... + (1) n S n
where Sr stands for the sum of the products of the roots 1 , ...., n taken r at a
time.
Comparing the coefficients on both sides , we see that
S1 =
a
a1
a
, S 2 = 2 ,.... S n = (1) n n .
ao
ao
ao
b
c
and =
a
a
b
,
a
c
d
and =
.
a
a
-----------------------------------------------------------------------------------------------------------------------------------------------------------------------
Illustrative Examples:
x3 + px2 + qx + r = 0 are in arithmetic progression,
p
3
On simplification, we obtain
2. Solve 27x3 + 42x2 28x 8 = 0, given that its roots are in geometric progression.
Solution:
Let the roots be
a
, a, ar
r
a
8
2
.a.ar = a 3 =
a=
r
27
3
Then ,
Since a =
2
2
60 60 2 4 27 12 2
=
2 27
9
or 2
2
2
, 2, .
9
3
progression.
2ac
]
a+c
+ + =
9
15
.. (1)
1
15
. (2)
2
+
+ = 2
.9
9
3 =
15
15
Substitute in (1), 2 + =
=
3
.
15
3
1
=
15
15
1
is a root of the given polynomial.
3
1
1
, 1, .
3
5
4. Show that the roots of the equation ax3 + bx2 + cx + d = 0 are in geometric
progression, then c3a = b3d.
Solution:
Suppose the roots are
Then
k
, k , kr
r
k
d
.k .kr =
r
a
i.e., k 3 =
d
a
10
bk 2 = ck
d2
d
= c 3
2
a
a
b3
b3d
= c 3 b3d = c 3a .
a
5. Solve the equation x3 - 9x2 + 14x + 24 = 0, given that two of whose roots are in
the ratio 3: 2.
Solution:
Let the roots be 3 , 2 ,
Then,
3 + 2 + = 5 + = 9
. (1)
3 .2 + 2 . + 3 . = 14
i.e., 6 2 + 5 = 14 (2)
3 . 2 . = 6 2 = 24
and
2 = 4
(3)
= 2 or
7
.
19
7
136
, from (1), we get =
. But these values do not satisfy (3).
19
19
1.3.
1 , 2 ,....., n taken r at a time, are symmetric functions. These are called elementary
symmetric functions.
Now we discuss some results about the sums of powers of the roots of a given
polynomial equation.
1.3.1. Theorem
The sum of the rth powers of the roots of the equation f(x) = 0 is the
coefficient of xr in the expansion of
xf ( x)
in descending powers of x.
f ( x)
Proof:
Let f(x) = 0 be the given nth degree equation and let its roots be
1
1 + .... + 1 n
x
x
1 1 2
n n2
1 +
1 +
+
+
....
+
........
+
+
+
....
2
2
x
x
x
x
1
2
= n + ( i )x + i x + ...... + ......
Therefore
r
i
is the coefficient of
descending powers of x.
12
xr in the expansion of
xf ( x)
in
f ( x)
if r n.
1
y
P
P
1
1
1
1
+ n11 + n22 + .... + Pn = ( 1 )( 2 ).....( n ) ,
n
y
y
y
y
y
y
=
1 (1 1 y) 1 2 (1 2 y) 1 ...... n (1 n y) 1
=
1 (1 + 1 y + 2 y 2 + ....) 2 (1 + 2 y + 22 y 2 + ....)
...... n (1 + n y + 2n y 2 + ....)
= S1 S 2 y S 3 y 2 ..... S r +1 y r ...
Cross multiplying, we get
P1 + 2P2 y + 3P3 y 2 + ...... + nPn y n 1 = (1 + P1 y + P2 y 2 + .... + Pn y n)
[S1 + S 2 y + ..... + S r +1 y r + .....]
P1 = S1
S1 + 1.P1 = 0
2P2 = S 2 S1P1
S 2 + S1 P1 + 2P2 = 0
3p 3 = S 3 S 2 P1 S1P2
S 3 + S 2 P1 + S1 P2 + 3P3 = 0 , and so on .
rPr = S r S r 1 P1 S r 2 P2 ...... S1 Pr 1
13
i.e.,
S r + S r 1 P1 + S r 2 P2 + ...... + S r n Pn = 0
Remark:
To find the sum of the negative powers of the roots of f(x) = 0, put x =
1
y
and find the sums of the corresponding positive powers of the roots of the new
equation.
Illustrative Examples
1. If , , are the roots of the equation x3 + px2 + qx + r = 0, find the value of the
following in terms of the coefficients.
(i)
1
1
(ii) (iii ) 2
Solution:
Here + + = p, + + = q, = r
1
+ +
(i)
(ii)
(iii)
2 = 2 + 2 + 2 + 2 + 2 + 2
=
=
+
1
+ +
=
=
p
p
=
r
r
q
q
=
r
r
= ( + + )( + + ) 3 = (q . p) 3 (- r ) = 3r pq .
b = 2 + 5 c = 3 + 4
14
i.e.,
a + b + c = + 2 + 3 + 4 + 5 + 6 =
( 6 1) 7 1
=
=
= 1
1
1
1
= ( + + ) 2 + 2 + 2 ( + + ) + 3
= ( + + ) [( + + ) 2 2 ( + + )] ( + + ) + 3
= 3[(9 4) 2] 3
= 9 3 = 12
Also,
2 2 + 2 2 + 2 2
=
2 2 2
( + + ) 2 2 2
2 2 2
.. (1)
We have:
2 = ( + + ) = 3 .1 = 3
2
(1) =
4 2 .3
= 2
1
4. Find the sum of the 4th powers of the roots of the equation x4 5x3 + x 1 = 0.
Solution:
Let
f(x) = x4 5x3 + x 1 = 0
xf 1 ( x)
can be evaluated as follows :
f ( x)
Therefore,
xf | ( x )
5 25 122 609
= 4 + + 2 + 3 + 4 + ......
f (x )
x x
x
x
Sum of the fourth powers of the roots = coefficient of x4.
= 609.
5. If + + = 1, 2 + 2 + 2 = 2, 3 + 3 + 3 = 3. Find 4 + 4 + 4 .
Solution:
Let x3 + P1x2 + P2x + P3 = 0 be the equation whose roots are , , , then
+ + = P1 P1 = 1
By Newtons theorem,
S2 + S1P1 + 2P2 = 0
i.e.,
2 + 1. ( 1) + 2 P2 = 0 P2 =
1/2
S4 + S3P1 + S2P2 + S1P3 = 0 (By Newtons theorem for the case r < n)
25/6
25
6
P1 = P2 = 0, P3 = 2 and P4 = 3
S3 + 0 + 0 + 3 . 2 = 0
S3 = 6
----------- (1)
f = a o 1 2 ...... n
k
k
k
k
f ( y / k ) = 0, are k 1 ,......, k n
y
y
y
f = a o + a1
k
k
k
i.e.,
n 1
+ ...... + a n = 0
a o y n + ka 1 y n 1 + k 2 a 2 y n 2 + ...... + k n a n = 0
Thus; to obtain the equation whose roots are k times the roots of a given equation,
we have to multiply the coefficients of x n , x n 1 ,......, x and the constant term by 1, k,
k2,kn-1 and kn respectively.
17
.. (1)
In (1), put y =
.. (2)
1
1
i.e., x =
x
y
1
1
1
Then f = a o 1 2 ........ n
y
y
y
1
1
1
f = a o + a 1
y
y
y
1 2
,......
n 1
+ ...... + a n = 0
a o + a 1 y + a o y 2 + ...... + a n y n = 0
. (1)
f ( y + h ) = a o (y + h 1 )(y + h 2 ).......(y + h n )
= a o (y (1 h ) )(y ( 2 h ) )......(y ( n h ) )
The roots of f (y + h) = 0 are 1 h,....., n h .
By (1), we obtain,
a o (y + h ) + a 1 ( y + h ) n 1 + ...... + a n = 0
n
18
. (3)
Replacing y = x h, we get
b o (x h ) + b1 (x h )
n
n 1
+ ........ + b n = 0
. (4)
a o x n + a 1 x n 1 + ...... + a n = 0
(1)
Suppose it is required to remove the second term of the equation (1). Diminish the
roots of the given equation by h.
For this, put y = x h i.e., x = y + h in (1), we obtain the new equation as
a o (y + h ) + a 1 (y + h )
n
n 1
+ ...... + a n = 0
ie a o y n + (na o h + a 1 )y n 1 + ...... + a n = 0
Now to remove the second term of the equation (1), we must have na o h + a 1 = 0
i.e., we must have
h=
a1
na o
Thus to remove the second term of the equation (1), we have to diminish its roots by
h=
a1
na o
Remarks:
If
f(x) = 0.
Solution:
To obtain the required equation, we have to multiply the coefficients of x3, x2,
x, and 1 by 1, 3, 32, and 33 respectively.
Thus x 3 3x 2 + 9 x + 27 = 0 is the desired equation.
2. Form an equation whose roots are the negatives of the roots of the equation
Solution:
By multiplying the coefficients successively by 1, -1, 1, -1 we obtain the required
equation as x 3 + 6 x 2 + 8x + 9 = 0 .
3. Form an equation whose roots are the reciprocals of the roots of
x 4 5x 3 + 7 x 2 4 x + 5 = 0 .
Solution:
We obtain the required equation, by replacing the coefficients in the reverse
order, as 5x 4 4 x 3 + 7 x 2 5x + 1 = 0
4. Find the equation whose roots are less by 2, than the roots of the equation
x 5 3x 4 2 x 3 + 15x 2 + 20 x + 15 = 0 .
Solution:
To find the desired equation, divide the given equation successively by x 2.
2
-3
-2
-2
-8
+14 68
-1
-4
+7
+34 83
+2
-4
+6
+1 -2
+3
+40
+2 +6
+8
+3 +4
+11
2
1
+10
+5 +14
+2
+7
1
20
h=
a1
8
=
= 2.
na o 4.1
On solving, we get x = - 4 , 4, - 3 , 3.
Thus the roots of the original equation are 2, 6, -1 and 5.
6. If , , are the roots of the equation x 3 + ax 2 + bx + c = 0. Form the equation
whose roots are , , .
Solution:
Note that =
Put
y=
c
x
x=
c
y
c
c
c
+ a
+ b
+ c = 0
y
y
y
i.e., y 3 by 2 + acy c 2 = 0 , which is the required equation.
7. If , , are the roots of x 3 x + 1 = 0 , show that
1+ 1+ 1+
+
+
=1
1 1 1
Solution:
We have to form the equation whose roots are
For this, put y =
1+ x
1 x
i.e., x =
1+ 1+ 1+
,
,
.
1 1 1
y 1
y +1
3
y 1 y 1
+ 1 = 0
Therefore the required equation is
y +1 y +1
On simplifying, we obtain y 3 y 2 + 7 y + 1 = 0
21
1+ 1+ 1+
+
+
=1
1 1 1
1 1
1
,
,......,
are also roots of the same equation, then such equations are
1 2
n
roots 1, 2 , ......, n .
Then
1 1
1
,
,......,
are also roots of the same equation. The equation with roots
1 2
n
1 1
1
,
, ......,
is :
1 2
n
a n x n + a n 1 x n 1 + ...... + a o = 0
(2)
Since (1) and (2) represents the same equation, we must have
ao
a
a
= 1 = n =k
a n a n 1 a o
Taking the first and last terms in the above equality, we obtain k2 = 1 i.e., k = + 1
when k = 1, we have ao = an, a1 = an-1 . . . .
Such equations are called reciprocal equations of first type.
When k = 1, we have ao = an, a1 = an-1, These type of equations are called
reciprocal equations of second type.
A reciprocal equation of first type and even degree is called a standard
reciprocal equation.
Note:
1. If f(x) = 0 is a reciprocal equation of first type and odd degree, the x = 1 is
always a root. If we remove the factor x + 1 corresponding to this root, we
obtain a standard reciprocal equation.
2. If f(x) = 0 is a reciprocal equation of second type and odd degree, then x = 1 is
always a roots. If we remove the factor x 1 corresponding to this root, we
obtain a standard reciprocal equation.
22
736 60
+ 2 =0
x
x
1
60 x 2 + 2 736x+ + 1433 = 0
x
Putting y = x + 1
60 y 2 736 y + 1313 = 0
On solving, we get y =
When y =
101 13
or
10
6
101
1 101
,x + =
10 x 2 101x + 10 = 0
10
x 10
i.e., x = 10,
1
10
Similarly when y =
13
3
, we get x = , 2
6
2 3
1 3 2
, ,
10 2 3
2. Solve :
x 2 5x 2 + 9 x3 9 x 2 + 5x 1 = 0
Solution:
This is a second type reciprocal equation of odd degree. So x = 1 is a root.
On division by the corresponding factor x 1, we obtain the other factor as
x 4 4 x 3 + 5x 2 4 x + 1 = 0 , which is a standard reciprocal equation.
1 i 3
3 5
or x =
2
2
23
1 i 3 3 5
,
2
2
-43
+76
+25
-100
+12
-62
+28
+106
-31
+14
+53
+6
+12
-38
-48
-19
-24
+5
+12
-14
-7
-38
+12
6
+5
6
6 x 4 + 5x 3 38x 2 + 5x + 6 = 0 is the required equation, which is a standard
reciprocal equation.
It can be written as
6 x 2 + 2 + 5 x + 1 38 = 0
x
x
Putting x +
1
10 5
= y and solving for y, we get y =
or
x
3
2
1
5
1
When y = 5 , we have x +
= . On solving we get: x = 2,
2
x
2
2
When y =
10
, we have 3x 2 + 10 x + 3 = 0 or x = 3 or 1
3
3
Thus the roots of the original equation are 4, 5 , 1, 5 ( by adding 2 to each of the
2
3
above roots)
24
---------- (1)
so that other two roots (real or complex) can then be found by polynomial division
and the quadratic formula. The solution proceeds in two steps. First, the cubic
equation is depressed; then one solves the depressed cubic.
a1
with n = 3 (i.e., the degree of the polynomial equation), a0 = a, a1 = b;
na0
so that h =
b
b
. Set x h = y or x = y + h = y .
3a
3a
b
to the cubic equation (1), and obtain
3a
b
b
b
a y + b y + c y + d = 0.
3
a
3
a
3
a
b2
2b3 bc
ay 3 + c y + d +
= 0,
3a
27 a 2 3a
(2)
had been discovered earlier by Italian mathematician Scipione dal Ferro (1465-1526).
The procedure is as follows:
First find s and t so that
and
3st = A
s3 t 3 = B
....... (3)
........(4)
25
This is true since we can simplify the left side using the binomial formula to obtain
s3 t 3 .
Now to find s and t satisfying (3) and (4), we proceed as follows: From Eq.(3), we
have s =
A
and substituting this into Eq.(4), we obtain,
3t
A
3
3t t = B
A3
= 0,
27
A3
= 0.
27
From this, we can find a value for u by the quadratic formula, then obtain t,
afterwards s. Hence the root s t can be obtained.
Illustrative Examples:
1. Using the discussion above, find a root of the cubic equation
2 x3 30 x 2 + 162 x 350 = 0.
Solution:
Comparing with
ax 3 + bx 2 + cx + d = 0, ........(5)
x= y
b
= y + 5 in (5), expanding and simplifying, we obtain
3a
y 3 + 6 y 20 = 0.
s 3 t 3 = 20.
(6)
(7)
Solving for s in (6) and substituting the result into (7) yields:
8 3
t = 20,
t3
26
s 3 = 20 + t 3 = 20 + 10 + 108 = 10 + 108
By Equation (7),
s = 3 10 + 108.
and hence
is given by
2.
x = y + 5 = 3 10 + 108 3 10 + 108 + 5.
Solution.
Since the term x 2 is absent, the given equation is in the depressed form.
Here x3 2 x = 5, and hence 3st = 2 and s 3 t 3 = 5.
3
Now substituting s =
2
2
in s 3 t 3 = 5, we obtain t 3 = 5
3t
3t
or
8
t 3 = 5 or
3
27t
t 6 + 5t 3 +
8
= 0.
27
u=
5 25
2
8
= 0 with
27
48
27 = 5 4.88 = 4.94 or 0.06.
2
We take the cube of root of the value with largest absolute value, and obtain
t = u1/ 3 = 1.7031 .
2
, we obtain s = 0.3914 .
3t
27
(8)
i.e.,
y 3 + 3Hy + G = 0 ,
i.e.,
(9)
where H = ac b 2 and G = a 2 d 3abc + 2b3 . The equation (9), where the term in y 2
absent, is the standard form of the cubic.
Now to solve (9) using Cardanos method, assume that the roots are of the form
p1/ 3 + q1/ 3 ; where p and q are to be determined.
Hence,
p + q = G
pq = H 3 .
and
p q = [( p + q )2 4 pq ]1/ 2
Now,
i.e.,
(11)
p q = G 2 + 4H 3
(12)
G + G 2 + 4 H 3
,
2
q=
G G 2 + 4 H 3
2
G + G 2 + 4 H 3
) and is one of the imaginary cube roots of unity. But
2
Hence
x 2 3 p1/ 3 q1/ 3 x ( p + q) = 0
(13)
p1/ 3 q1/ 3 = 3
Hence pq = 27 . Now ,
Hence
p q = 26
(14)
4, 3 2 , 2 3 .
Another Method
Since 4 is a root of the given cubic, x + 4 is a factor of the polynomial in the given
cubic equation. Removing the factor x + 4 , the cubic equation yields the quadratic
equation
x2 4 x + 7 = 0
Hence
x=
4 12
= 2i 3
2
Solve x3 6 x 2 + 3x 2 = 0
29
i.e.,
Hence pq = 27
(15) and
and
p1/ 3 q1/ 3 = 3
p q = ( p + q )2 4 pq = 6
(16)
(17)
Hence
( ) 2 ( )2 ( )2 = 27(G 2 + 4 H 3 ). (18)
Then nature of the roots , , of Eq. (17) can be obtained by a consideration of the
product in Eq. (18). Since imaginary roots occur in pairs, equation (17) will have
either all real roots, or one real and two imaginary roots. The following cases can
occur.
Case 1: The roots , , are all real and different. In this case
( ) 2 ( )2 ( )2 is positive. Therefore by Eq. (18), G 2 + 4 H 3 is negative.
Case 2: One root, say , is real and the other two imaginary. Let and be m in .
Then
( ) 2 ( )2 ( )2 = (2in)2 (m in ) 2 (m + in ) 2
= 4n 2 {(m ) 2 + n 2 }2 ,
30
when G 2 + 4 H 3 < 0 , the roots of the cubic in Eq. (17) are all real;
(ii)
when G 2 + 4 H 3 > 0 , the cubic in Eq. (17) has two imaginary roots;
(iii) when G 2 + 4 H 3 = 0 , the cubic in Eq. (17) has two equal roots; and
(iv) when G = H = 0 , all the roots of the cubic in Eq. (17) are equal.
Remark
On substituting the values of G and H, it can be seen that
G 2 + 4 H 3 a 2 {a 2 d 2 6abcd + 4ac3 + 4b3 d 3b 2 c 2 } .
The expression in brackets is called the discriminant of the general cubic in Eq. (7),
and is denoted by .
Such a function is sometimes called a biquadratic function, but the latter term can
occasionally also refer to a quadratic function of a square, having the form
ax 4 + bx 2 + c,
(19)
u = ( x 2 + px + s )2 (mx + n) 2
(20)
Hence,
10 = 2 ps 2mn, 3 = s 2 n 2 .
p = 1, 12 = 2 s + 1 m 2 , 10 = 2 s 2mn, 3 = s 2 n 2 .
Thus p = 1, m2 = 2 s + 13, mn = s 5, n 2 = s 2 3.
To eliminate m and n , we note that (mn) 2 = m2 n 2
which gives ( s 5)2 = (2s + 13)( s 2 3) or s 2 + 10 s + 25 = 2 s 3 6 s + 13s 2 39
which is the cubic equation 2 s 3 16 s + 12 s 2 64 = 0
or s 3 + 6 s 2 8s 32 = 0.
(21)
( Remarks :
The reducing cubic gives three values of s . These do not however lead
to three different sets of roots for the quartic equation. They only give
three different methods of factorizing the left hand side of the quartic.
Hence it is enough to find any one root of the reducing cubic. )
Hence,
n 2 = s 2 3 = (2) 2 3 = 1.
mn = s 5 = 2 5 = 3.
u = {x 2 + (1 + 3) x 2 1}{x 2 + (1 3) x 2 + 1}
u = {x 2 + 2 x 3}{x 2 4 x 1} , or
or
u = ( x 1)( x + 3){x 2 4 x 3}
(22)
(23)
Then, x 4 3 x 2 6 x 2 = ( x 2 + px + s )2 (mx + n) 2
Equating coefficients of like powers of x, we obtain
p = 0, m2 = 2 s + 3, mn = 3, n 2 = s 2 + 2.
2 s 3 + 3s 2 + 4 s 3 = 0. (24)
In order to find a root of (24) we proceed as follows: If , , are the roots of (24),
then the equation with roots 2 , 2 , 2 is given by
3
y
y
y
2 + 3 + 4 3 = 0
2
2
2
Or
y 3 + 3 y 2 + 8 y 12 = 0
(25)
i.e., s =
1
is a solution of the reducing cubic (24).
2
2
m2 = 2
1
9
1
+ 3 = 4, n 2 = + 2 = .
2
4
2
33
Hence
1
is a root of (24).
2
3
2
3
2
Putting p = 0, s = , m = 2, n = , in
u = {x 2 + ( p + m) x + s + n}{x 2 + ( p m) x + s n} , we get u = {x 2 + 2 x + 2}{x 2 2 x 1} .
3. Solve x 4 + 3 x3 + x 2 2 = 0.
Solution.
Let
3
5
p = , m2 = 2s + ,
2
4
3
mn = s, n 2 = s 2 + 2. To eliminate m and n , we note that (mn) 2 = m 2 n 2 , which gives
2
2
5 2
3
2 s = 2s + 4 ( s + 2 ) and is the cubic equation
4 s 3 2s 2 + 8s + 5 = 0.
(27)
In order to find a root of (27) we proceed as follows: If , , are the roots of (24),
then the equation with roots 4 , 4 , 4 is given by
3
y
y
y
4 2 + 8 + 5 = 0.
4
4
4
Or
y 3 2 y 2 + 32 y + 80 = 0.
(28)
If we let 4 = 2 then =
reducing cubic (27).
m2 = 2s +
1
is a root of (27).
2
Hence
5
5 1
= 1 + =
4
4 4
n2 = s2 + 2 =
1
9
+2= .
4
4
34
i.e.,
s=
1
is a solution of the
2
3
2
3
4
9
4
3
2
in
terms
of
finite
number
of additions, subtractions,
as solution.
35
, which
If the coefficients of a polynomial equation are all positive, the equation has
no positive root; for example, the equation
x 4 + 3x 2 + 3 = 0
coefficients of the odd powers are all of the opposite sign, the equation has no
negative root; thus for example, the equation
x8 + x 7 + x 5 2 x 4 + x 3 3 x 2 + 7 x 3 = 0
cannot have a negative root.
3. If the equation contains only even powers of x and the coefficients are all of
the same sign, the equation has no real root; thus for example, the equation
x8 2 x 4 3 x 2 3 = 0
x 7 + x5 + 3x 3 + 8 x = 0
has no real root except x = 0.
Suppose that the signs of the terms in a polynomial are + + + + + ; here
the number of changes of sign is 7.
+++++
+
__________________
+++++
+++++++
__________________
+++++
Here in the last line the ambiguous sign is placed wherever there are two different
signs to be added.
Here we see that in the product
(i) an ambiguity replaces each continuation of sign in the original polynomial;
(ii) the signs before and after an ambiguity or set of ambiguities are unlike;
(iii) a change of sign is introduced at the end.
Let us take the most unfavourable case (i.e., the case where the number of
changes of sign is less) and suppose that all the ambiguities are replaced by
continuations; then the sign of the terms become
+ + + + + +,
and the number of changes of sign is 8.
We conclude that if a polynomial is multiplied by a binomial (corresponding to a
positive root) whose signs are + , there will be at least one more change of sign in
the product than in the original polynomial.
If then we suppose the factors corresponding to the negative and imaginary
roots to be already multiplied together, each factor x a corresponding to a positive
root introduces at least one change of sign; therefore no equation can have more
positive roots than it has changes of sign.
Again, the roots of the equation
opposite to them in sign; therefore the negative roots of f ( x ) = 0 are the positive
roots of f ( x ) = 0; but the number of these positive roots cannot exceed the number
of changes of sign in f ( x); that is, the number of negative roots of f ( x ) = 0 cannot
exceed the number of changes in sign in f ( x ).
37
1.8. Exercises
1. Solve the equation x 4 + x 3 x 2 2 x 2 = 0 given that one root is
2.
, ,
1
1
5 = 3 2 .
5
6
1
2
and ( ) .
+
10. Prove that the sum of the ninth powers of the roots of x 3 + 3x + 9 = 0 is zero.
11. If , , are the roots of x 3 7 x + 7 = 0 , find the value of 4 + 4 + 4 .
12. Find the equation whose roots are the roots of the equation
3x 4 + 7 x 3 15x 2 + x 2 = 0 , each increased by 7.
16. If , , are the roots of the equation x 3 + px + q = 0 , find the equation whose
roots are
+ , + , + .
39
23. Find the greatest possible number of real roots of the equation
x 5 6x 2 4x + 5 = 0
40
Module-II
FUZZY SETS
Prepared by:
Santhosh P.K.
Asst. Professor
Dept. of Mathematics
Govt. Brennen College, Thalassery.
1
From Crisp Sets to Fuzzy Sets: A
Grand Paradigm Shift
Since its inception in 1965, the theory of fuzzy sets has advanced in a variety of ways and in many disciplines. Applications of this theory can be
found, for example, in artificial intelligence, computer science, medicine, control engineering, decision theory, expert systems, logic, management science,
operations research, pattern recognition, and robotics. Mathematical developments have advanced to a very high standard and are still forthcoming to
day. In this course material, the basic mathematical framework of fuzzy set
theory is described. Applications of fuzzy set theory to real problems are
abound.
that elements are non repeated and given an element, it must be possible to
determine whether it belongs or does not belongs the given collection.
The following are some examples of crisp sets.
1. Collection of boys in a class. That is, {x : x is a boy in the class}.
2. Collection of positive integers between 1 and 10. That is, {1, 2, 3, 4, 5
,6, 7, 8, 9, 10}.
The following are some non example of sets as they cannot be uniquely
represented.
1. Collection of beautiful girls in a class.
2. The collection of real numbers much greater than 8.
To indicate an individual object x is a member or element of a set A, we
write x A, else x
/ A. A set can be represented using three ways, naming
its elements, called List Method; defined by a property of its elements, called
Rule Method and the last by, using a function, called Characteristic Function.
A set A X is defined by its characteristic function A as follows,
1 ; f or x A
A (x) =
0 ; f or x
/A
That is, a characteristic function maps elements of X to its elements of the
set {0, 1}, which is formally denoted by A : X {0, 1}.
General Principle of Duality in Set Theory:
For each valid equation in set theory that is based on the union and
intersection operation, there corresponds a dual equation , also valid, that is
obtained by replacing , , by X, and in order.
The membership value 0 indicates that the element lies entirely outside
the set, where as 1 indicates that the element lies entirely inside the given
set A. Any value between 0 and 1 indicates a degree of partial membership
to the set. In Fuzzy logic every thing is a matter of degree. In this logic, an
element can have different degree of membership in different sets.
The following figures represents the fuzzy set set of real numbers close
to 2 corresponds to different membership functions.
Fig. 1.1:
Exercise: By describing each of the fuzzy sets mentioned above, discuss the
merits and demerits of each such. Provide examples to each.
Remark 1.3.1. F(X) denotes the fuzzy power set of X( the set of all ordinary fuzzy sets of X).
Next we introduce some basic concepts and terminology of fuzzy sets.
Definition 1.3.1. cuts and strong cuts
Let A be a fuzzy set defined on X and [0, 1]. Then cut of A, denoted
by A is defined as the crisp set
A = {x : A(x) }
A = {x : A(x) > }.
1
when x 20
35 x
A1 (x) =
when 20 < x < 35
15
0
when x 35
0
when either x 20 or 60
x 20
A2 (x) =
when 20 < x < 35
15
1
when 35 x 45
0
when x 45
x 45x
A3 (x) =
when 45 < x < 60
15
1
when x 60
A complete list of cut and strong cuts for the fuzzy sets A1 , A2 and
A3 are given below.
Remark 1.3.2. It is immediate form the definition that, for any fuzzy set A
and pair 1 , 2 [0, 1], of distinct values such that 1 < 2 , we have
1
A 2 A
1 +
and
A 2 + A.
A 2 A = 2 A,
A 2 A = 1 A
and
1 +
A 2 + A = 2 + A,
1 +
A 2 + A = 1 + A.
Thus all cuts and all strong cuts of any fuzzy set form two distinct
families of nested crisp sets.
Definition 1.3.3. Support of a Fuzzy Set
The support of a fuzzy set A defined on X is the crisp set that contain
all elements of X that have non zero membership grades in A. Obviously
support of A is exactly the same as the strong cut of A for = 0. Support
of A is generally denoted either by S(A) or by Supp(A) or by
0+
A.
10
11
of A. Consequently,
A(x1 + (1 )x2 ) = A(x1 ) = min[A(x1 ), A(x2 )].
Conversely, assume A satisfies the given equation. We need to prove that
for any (0, 1], A is convex. Now for any x1 , x2 A, and for any
[0, 1], by given equation,
A(x1 + (1 )x2 ) min[A(x1 ), A(x2 )] min(, ) = ;
That is, x1 + (1 )x2 A. There fore, A is convex for any (0, 1].
Hence A is convex.
Definition 1.3.6. Cutworthy Property, Strong Cut cutworthy Property
Any property generalized from classical set theory into the domain of fuzzy
set theory that is preserved in all (0, 1] in the classical sense is called
cutworthy property. If it is preserved in all strong cuts for all (0, 1], it
is called strong cut worthy property.
Convexity of fuzzy sets defined above is an example of cut worthy property.
Question: Prove that convexity in fuzzy sets is a strong cut worthy property.
1.4
12
Basic fuzzy set operations includes fuzzy complement, fuzzy union and fuzzy
intersection.
Definition 1.4.1. Fuzzy Complementation
Let A be fuzzy set defined over the universal set X. Then for any x X,
A = 1 A(x).
Elements of X for which A(x) = A(x) is called equilibrium points.
For example, complements of the fuzzy sets defined perviously by the
membership functions A1 ,A3 are given below. Draw the complement of the
13
Remark 1.4.1. The law of contradiction, ie, AA = is not valid for fuzzy
14
15
A(x).
xX
X
1
(|A|
max[0, A(x) B(x)])
|A|
xX
It can be simplified as
S(A, B) =
|A B|
|A|
A A;
A + A;
iii. (A B) = A B and (A B) = A B.
iv.
(A B) = + A + B and
16
(A B) = + A + B.
v. (A) = (1)+ A.
For a detailed proof, students are requested to refer your text.
Remark 1.5.1. Result v stated above shows that fuzzy complement is neither
cutworhty nor strong cutworhty. That is,
(A) 6= A and
(A) 6= + A.
It is illustrated as below.
Remark 1.5.2. In spite of its negative connotation, (i.e, an idea that is implied or suggested), property v describes an interesting feature of the standard
fuzzy complement: The cut of the complement of A is always the same as
the complement of the strong (1 ) cut of A.
17
Ai =
Ai
Ai ( Ai ) and
1.
iI
2.
S +
iI
Ai = +
iI
iI
iI
Ai
and
iI
T +
Ai +
iI
Ai
iI
Proof. We will prove the second part of the above result. First part is left to
the reader. First we prove the equality of the set union.
x
S +
iI
Ai
iI
Ai )(x) >
iI
A + B.
18
3. A = B if and only if A = B.
4. A = B if and only if
A = + B.
Proof. We will prove the first two results. Students are requested to do the
next two.
1. Let A B. Assume that A * B. Then there exists x0 X and
0 [0, 1] such that x0
A and x0
/
A + B.
A and
x0
/ + B. Hence + A * + B, which is a contradiction to the fact that
+
= . A(x).
A,
[0,1]
where
19
union.
Proof. For x X, let a = A(x). Then
!
S
(x) =
sup A(x)
[0,1]
[0,1]
(a,1]
For each (a, 1], we have A(x) = a < and , therefore A(x) = 0.
On the other hand, for each [0, a], we have A(x) = a , therefore,
A(x)
= . Hence
[0,1]
Exercise: Illustrate the above theorem for the fuzzy set A by taking
membership function as below
A(x) =
x1
when x [1, 2]
3 x x [2, 3]
0
otherwise
set defined by
+ A(x)
20
+ A
(x) =
[0,1]
sup + A(x)
[0,1]
+ A(x), sup + A(x)]
[a,1]
[0,a)
= max[ sup
=
sup = a = A(x)
[0,a)
2
Operations on Fuzzy sets
Let A, B be two fuzzy subsets of X. Then for any x X, the standard fuzzy
operations include
1. Complementation : A = 1 A(x)
2. Union : (A B)(x) = M ax[A(x), B(x)]
3. Intersection :(A B)(x) = M in[A(x), B(x)]
22
that b0 > b1 > c(a0 ) for which no a1 [0, 1] xists such that c(a1 ) = b1 .
This contradicts the fact that c is bijective.
23
From this theorem, it can be noted that class of all classical fuzzy complements is contained in class of all involutive fuzzy complements(axionms
c1-c4), which is contained in class of continuous fuzzy complements(Axioms
c1 - c3), which is contained in class of all fuzzy complements(Axioms c1 and
c2).
The axioms c1 and c2 are called axiomatic skeleton. Examples of fuzzy
complements that satisfies only axiomatic skelton are the threshold type
complements defined by
1
c(a) =
0
f or a t
f or a > t
1a
,
1 + a
24
(2.1)
25
ec
(a + ) 2 1 when 6= 0
1
when = 0
2
a [0, 1] defined by
c(d a) d a = a c(a)
(2.2)
26
Theorem 2.1.5. For each a [0, 1], d a = c(a) if and only if c(c(a)) = a,
that is, when complement is involutive.
Proof. Let d a = c(a). Then from equation 2.2, we have
c(c(a)) c(a) = a c(a).
There fore , c(c(a)) = a. For the reverse implication, let c(c(a)) = a. Then
from equation 2.2, we have
c(d a) d a = c(c(a)) c(a)
for d a whose solution is d a = c(a).
Theorem 2.1.6. First Characterization Theorem of Fuzzy Complements
Let c be a function from [0, 1] to [0, 1]. Then c is a fuzzy complement
(involutive) iff there exists a continuous function g from [0, 1] to R such that
g(0) = 0, g is strictly increasing and
c(a) = g 1 (g(1) g(a))
(2.3)
1
ln(1 + a) f or > 1
27
28
Theorem 2.2.1. The standard fuzzy intersection is the only idempotent tnorm.
Proof. Clearly min(a, a) = a for all a [0, 1]. Assume that there exists a a
t-norm such that i(a, a) = a for all a [0, 1]. Then for any a, b [0, 1] with
a b, we have
a = i(a, a) i(a, b) i(a, 1) = a.
Hence i(a, b) = a = min(a, b). Similarly for a b, we have
b = i(b, b) i(1, b) = b
Consequently i(a, b) = b = min(a, b). Hence i(a, b) = min(a, b) for all a, b
[0, 1].
The following are some standard t-norms used as fuzzy intersections(each
defined for all a, b [0, 1]).
1. Standard intersection: i(a, b) = min(a, b).
2. Algebraic Product: i(a, b) = ab.
3. Bounded Product: i(a,b)=max(0, a+b-1).
a when b = 1
0 otherwise.
Exercise: Draw graph of these four fuzzy intersections.
Theorem 2.2.2. For all a, b [0, 1], imin (a, b) i(a, b) min(a, b) where
imin denotes the drastic intersection.
29
f 1 (a) =
when a (, 0)
0
when a (f (0), )
f 1 (a) =
when a (, 0)
0
when a (f (1), )
30
sa 1
, (s >
s1
0, s 6= 1).
Remark 2.2.1. The t-norms for Yager class of decreasing operators is given
by i (a, b) = 1 min(1, [(1 a) + (1 b) ]1/ ), > 0.
Exercise: Prove that imin (a, b) i (a, b) min(a, b) for all a, b [0, 1].
31
32
a when b = 0
1 otherwise.
Exercise: Draw the graph of these fuzzy unions and conclude that for all
a, b [0, 1],
max(a, b) a + b ab min(1, a + b) umax (a, b).
Proof of the following theorem is similar to that in the corresponding case
of t-norm.
Theorem 2.3.2. For all a, b [0, 1], max(a, b) u(a, b) umax(a,b) .
Proof of the following fundamental theorem of t-conorm is omitted from
your syllabus.
Theorem 2.3.3. Characterization Theorem of t-conorm
Let u be a binary operation on the unit interval. Then u is an Archimedian
t-conorm if and only if there exists an increasing generator such that u(a, b) =
g 1 (g(a) + g(b)) for all a, b [0, 1].
Exercise: Compute the t-conorm for the class of increasing generators
described below.
1. Schweizer and Sklar class of increasing generators for the parameter p
defined by fp (a) = 1 (1 a)p , p 6= 0.
2. Yager class of increasing generators, given by f (a) = a , ( 0).
3. Frank class of increasing generators, given by fs (a) = In
s1a 1
,
s1
(s > 0, s 6= 1).
33
(2.4)
(2.5)
and
Let the triple < i, u, c > denote that i and u are dual with respect to c,
and let any such triple be called a dual triple.
Exercise: Prove that the following are dual t-norms and t-conorms with
respect to the standard complement cs . (ie, dual triples)
1. < min(a, b), max(a, b), cs >.
2. < ab, a + b ab, cs >.
3. < imin (a, b), umax (a, b), cs >.
Theorem 2.4.1. The triple < min, max, c > and < imin , umax , c > are dual
with respect to any fuzzy complement c.
34
Proof. Let a b. Then c(a) c(b) for any fuzzy complement c. Hence
max(c(a), c(b)) = c(a) = c(min(a, b)),
min(c(a), c(b)) = c(b) = c(max(a, b)).
The reader is asked to prove the rest.
Theorem 2.4.2. Given a t-norm i and an involutive fuzzy complement c,
the binary operation u on [0, 1] defined by
u(a, b) = c(i(c(a), c(b)))
(2.6)
for all a, b [0, 1] is a t-conorm such that < i, u, c > is a dual triple.
Proof. It is a routine procedure to verify that u is a t-conorm.
Next we have to prove that < i, u, c > is a dual triple. Since c is involutive,
c(c(a)) = a for all a [0, 1]. Using the equation 2.6, we can prove that u
satisfies De Morgans Laws as below.
c(u(a, b)) = c(c(i(c(a), c(b))) = i(c(a), c(b)),
u(c(a), c(b)) = c(i(c(c(a), c(b))) = c(i(a, b)).
Hence i and u are dual with respect to c.
Exercise: Using the above theorem, obtain a t-conorm for the t-norm i(a, b) =
ab using the Sugeno class of fuzzy complements c (a) =
1a
1+a
for 1.
35
(2.7)
(2.8)
(2.9)
36
37
= i(e, c(e))
= 0.
REFERENCES
[1.] George J. Klir and Bo Yuan- Fuzzy Sets and Fuzzy Logic: Theory and
Applications,PHI,2000. (Text Book as per Syllabus)
[2.] Klir, G.J., and T. Folger, Fuzzy Sets, Uncertainity and Information, PHI
1988.
[3.] Dubois D and H. Prade, Fuzzy Sets and Systems: Theory and Applications, Academic Press, 1988.
[4.] Abhram Kandel, Fuzzy Mathematical Techniques with Applications,
Addison.