III.

FUZZY SETS AND OPERATIONS

3.1. INTRODUCTION

Classical sets are also called ‘crisp’ sets so as to distinguish them from fuzzy sets. In fact, the
crisp sets can be taken as special cases of fuzzy sets. Let A be a crisp set defined over the
universe X. Then for any element x in X, either x is a member of A or not. In fuzzy set
theory, this property is generalized. Therefore, in a fuzzy set, it is not necessary that x is a full
member of the set or not a member. It can be a partial member of the sets.

The generalization is performed as follows: For any crisp set A, it is possible to define a
characteristic function µ X = {0,1 }. i.e. the characteristic function takes either of the values 0 or
1 in the classical set. For a fuzzy set, the characteristic function can take any value between
zero and one.

Definition 3.1
The membership function µ A (x) of a fuzzy set A is a function µ A : X → [0,1] --- (1)
So every element in x in X has membership degree: µ A ( x) ∈ [0,1] ----(2)
A is completely determined by the set of tuples: A = {( x, µ A ( x)) x ∈ X } -----(3)

Example.3.1. Suppose some one wants to describe the class of cars having the property of
being expensive by considering BMW, Rolls Royce, Mercedes, Ferrari, Fiat ,Honda and
Renault. Some cars like Ferrari and Rolls Royce are definitely expensive and some like Fiat
and Renault are not expensive in comparison and do not belong to the set. Using a fuzzy set,
the fuzzy set of expensive cars can be described as:
{(Ferrari,1),(Rolls Royce,1),(Mercedes, 0.8),( BMW, 0.7), (Honda,0.4)}. Obviously, Ferrari
and Rolls Royce have membership value of 1 whereas BMW, which is less expensive, has a
membership value of 0.7 and Honda 0.4.

Example 3.2: A set of natural numbers ‘close to 6’ can be defined as a fuzzy set. This can be
done, say, buy including all numbers from 3 to 9 as follows:
~
6 = { (3,0.1), (4,0.2), (5,0.5), (6,1), (7,0.5), (8,0.2)m (9,0.1)}
The expression for the characteristic function or membership function can be written as:

1
µ ~6 ( x) =
(1 + ( x − 6) 2
~
The fuzzy set 6 contains the elements like ( 6,1),(5.5,0.8),and (100,0.000113161) .etc.

Zadeh proposed an alternate representation for fuzzy sets, which is more convenient. Suppose
A is a finite crisp set with elements {x1 , x 2 ,.......x n } , then an alternative representation for C is:
C = {x 1 + x 2 + x 3 + ...... + x n }

Here + denotes an enumeration or listing rather than addition. Yet another way is to include
µ ( x)
the characteristic function also into its fold is: where the line between the top and
x
bottom entries is just a delimiter or separator.
The fuzzy set of expensive cars can be now written using this notation as:

1
 1 1 0.8 0.7 0.4 0 0
{ + + + + + + }
Ferrari Rollsroyce Mercedes BMW Honda Fiat Renault
The set F = {( x, µ F ( x)) | x ∈ X } can then be written as:
n
µ F ( x1 ) µ F (xn ) µ F ( xi )
F=
x1
+ ............. +
xn
= ∑
i =1
xi
Where + satisfies the condition:
a b a, b
+ = max( )
x x x
i.e. if the same element has two membership values, say 0.8 and 0.6, then its membership
degree becomes 0.8, the larger of the two. Any countable or discrete universe U allows such a
notation.
µ F ( x)
A= ∑
x∈ X
x
but when the set is uncountable, or continuous, it can be written as :
µ A ( x)
A=∫
x
Here the symbol ∫ denotes a listing or collection rather than integration.
The set ‘close to 6’ can now be re-written as :
~ 0.1 0.2 0.5 1 0.5 0.2 0.1
6 ={ , , , , , , }
3 4 5 6 7 8 9
Example.3.3 : X can be the set of integers described by the membership function :
x−6
1− | | 3≤ x ≤ 6
3
µ F ( x) =
0 otherwise
µ F ( x)
Then F can be expressed as : A=
x ∫ where the operation is on the set of all real
integers.
Example 3.4: Consider the set of old people belonging to the universe of people in the age of
0 to 120. X = [ 0, 120]
We can define the membership function as :

0 ; 0 ≤ x ≤ 60

µ old = ( x -60) / 20 ; 60 ≤ x ≤ 80
1 ; x ≥ 80

120
µ old ( x)
then we can call the set as : ∫0
x

3.2 PROPERTIES OF FUZZY SETS

Fuzzy sets follow the same properties as crisp sets. Since membership values of crisp sets are
a subset of the interval [0,1], classical sets can be thought of as generalization of fuzzy sets:

2
 Commutativity : AUB % % = BUA% %
A% ∩ B% = B% ∩ A%
Associativity: A% ∪ ( B% ∪ C% ) = ( A% ∪ B% ) ∪ C%
A% ∩ ( B% ∩ C% ) = ( A% ∩ B% ) ∩ C%
Distributivity: A% ∪ ( B% ∩ C% ) = ( A% ∪ B% ) ∩ ( A% ∪ C% )
Idempotency: A% ∪ A% ) = A%
A% ∩ A% = A%
Identity : A% ∪ ∅ = A% A% ∩ X = A%
A% ∩ ∅ = ∅ A% ∪ X = X
Transitivity: A ⊆ B% ⊆ C ) then A% ⊆ C% )
% %
~ ~
Involution: A=A

3.3.OPERATIONS ON FUZZY SETS

~ ~ ~
Let A, B , C be three fuzzy sets defined on the universe of discourse X . For a given element x
of the universe, the following function-theoretic operations of Union, Intersection and
Complements are defined as follows :

Union : µ A~ ∪ B~ ( x) = µ A~ ( x) ∨ µ B~ ( x)
~ ~
Intersection : µ A~ ∩ B~ ( x) = µ A~ ( x) ∧ µ B~ ( x) µ A B
Complement : µ A~ ( x) = 1 − µ A~ ( x)
These are shown graphically in Figs3.8-3.10

Fig.3.1. Union of fuzzy sets

~ ~
µ A B

~ ~
µ A A

Fig.3.2. Intersection of fuzzy sets Fig.3.3. Complement of fuzzy set

~
Any fuzzy set A defined on the universe X is a subset of the universe. Also by definition null
set has a membership 0 and x in X has a membership 1. Note that the null set and the whole
set are not fuzzy sets.
Example3.5:A simple hollow shaft is 1-m radius and has a wall thickness of ( 1/2π) m. The
shaft is built up stacking a ductile section and a brittle section. A downward force P and a
torque T are simultaneously applied to the shaft. The failure properties of the two sections can
be described by the following fuzzy sets A and B for the ductile and brittle sections as
follows:

3
 ~ 1 0.5 0.3 0.2 ~ 0.5 0.7 0.2 0.4
A ={ + + + } and B = { + + + }
2 3 4 5 2 3 4 5

We can see the following:

.
1. The set of loadings for which either material B or material D will be “safe” can be
~ ~
obtained by getting A ∪ B
2. The set of loadings for which one expects that both material B and material D are
~ ~
“safe” can be obtained by forming A ∩ B .
~ ~
3. The complements A and B represents the set of loadings for material D and B are
unsafe.
~ ~
4. A | B gives the set of loadings for which the ductile material is safe but the brittle is
not.
~ ~
5. B | A gives the set of loadings for which the brittle material is safe but the ductile
not.
~ ~ ~ ~
6. De Morgans laws can be used to find A ∩ B .= A ∪ B which asserts that the loadings
that are not safe with respect to both materials are the union of the those that are
unsafe with respect to the brittle material with those that are unsafe for with respect to
the ductile material .
~ ~ ~ ~
7. De Morgans law A ∪ B = A ∩ B asserts that the loads that are safe for neither
material D nor material B are the intersection of those that are unsafe for material D
with those that are unsafe for material B.
Consequently, we can find the following:
~ 1 0 .5 0 .3 0 .2 ~ 0 .5 0 .7 0 .2 0 .4
A ={ + + + } and B = { + + + }
2 3 4 5 2 3 4 5

Complements:
~ 1 0 0.5 0.7 0.8 ~ 0.5 0.3 0.8 0.6
A ={ + + + + } and B = { + + + }
1 2 3 4 5 2 3 4 5

Union :
~ ~ 1 0.7 0.3 0.4
A∪B ={ + + + }
2 3 4 5
Interesection :
~ ~ 0.5 0.5 0.2 0.2
A∩ B ={ + + + }
2 3 4 5
Difference :
~ ~ ~ ~ 0.5 0.3 0.3 0.2
A| B = A∩B ={ + + + }
2 3 4 5
etc.

3.4 SET POINTS AS HYPERCUBES

A geometric analogue can be used to illustrate the idea of a set membership. The fuzzy
~
set A is defined on the a universe X. For a universe with only one element. The
membership function is defined on the unit interval [ 0, 1] ; foe a two element universe,

4
the membership function is defined on the unit square; for a three element universe, the
membership function is defined on a unit cube. All these are shown in Fig. For a universe
of n elements we define the membership on the unit hypercube, I n = [0,1] n . The end points
on the unit interval in the Fig and the vertices of the unit square, unit cube in Fig b and c
respectively represent the possible crisp sets, or collections of the elements of the universe
in each figure. This collection of possible crisp(non fuzzy) subsets of elements in a
universe constitute the power set of the universe. For example. In Fig. C, the universe
comprises three elements, X = { x1 , x 2 , x 3 }. The point (0,0,1) represents the crisp subsets in
3-space where x1 and x 2 have no membership and x 3 has full membership. i.e. the subset
{ x 3 }; the point (1,1,0) is the crisp subset where x1 and x 2 have full membership and
element x 3 has no membership and so on. In general there are 2 n subsets in the power
set of a universe with n elements; geometrically this universe is represented by a
hypercube in n-space, where 2 n vertices represent the collection of sets constituting the
power set. Two points in the diagram bear special notes, as illustrated IN Fig. The point
(1,1,1) where all elements in the universe have full membership is called the whole set X
and the point (0,0,0) where all elements have zero membership is called the null set ∅ .
The centroids of each of the diagrams represent single points where the membership value
for each element equals 0.5. This corresponds to the point of maximum fuzziness. A
membership value 0.5 corresponds to the element belongs to the set as much as it does
not, i.e. it holds equal membership in both fuzzy sets and its complement.

∅ .= (0) X ={ x1 }=(1) { x 2 }= (0,1) X= { x1 , x 2 }=(1,1)

Fig.3.4a.Set as line (n=1) . (0.5, 0.5)

∅ .= (0,0) { x1 }= (1,0)
Fig.3.4b.Set as a square (n=2)
(0,1,1) (1,1,1)

(0,1,0) (1,1,0)

. (0.5,0.5,0.5)

(0,0,1) (1,0,1)

(0,0,0) (1,0,0)
Fig.3.4c. Set as a cube ( n=3)