8

Introduction
1.6 Soft Computing
The wo major problem-solving technologies include:

1. hard computing:
2. soft computing.

Hard computing deals with precise models where accurate solutions are achieved quickly. On the other
hand. soft computing deals with approximate models and gives solution to complex problems. The two
problem-solving technologies are shown in Figure 1-2.
Soft computing is a relatively new concept, the term really entering general circulation in 1994. The term
soft computing" was introduced by Professor Lotfi Zadeh with the objective of exploiting the tolerance
for imprecision, uncertainty and partial truth to achieve tractability, robustness, low solution cost and better
rapport with reality. The ultimate goal is to be able to emulate the human mind as closely as possible. Soft
computing involves partnership ofseveral fields, the most inmportant being neural networks, GAs and FL. Also
included is the ficld of probabilistic reasoning, employed for its uncertainty control techniques. However, this
field is not examined here.
Soft computing uses a combination of GAs, neural networks and FL. A hybrid technique, in fact, would
inherit all the advantages, but won't have the less desirable features of single sofi computing components. It
has to possess a good learning capacity, a better learning time than that of pure GAs and less sensitivity to
theproblem of local extremes than neural networks. In addition, it has to generate a fuzy knowledge base,
which has a
linguistic representation and avery low degree of computational complexity.
An important thing about the constituents of soft computing is that they are
complementary, not compet
itive, offering their own advantages and techniques to partnerships to allow solutions to otherwise unsolvable
problems. The constituernts of soft computing are examined in turn, following which existing applications of
partnerships are described.
"Negotiation is the communication process of a group of agents in order to reach a mutually accepted
agreement on some matter. This definition is typical of the research being done into negotiation and co-
ordination in relation to sofrware agents. It is an obvious necessity that when multiple agents interact, they
will be required to co-ordinate their efforts and attempt to sort out any conflicts of resources or interest.
lt is important to appreciate that agents are owned and controlled by people in order to complete tasks on
their behalf. An example of a possible multiple-agent-based negotiation scenario is the competition between

HARD COMPUTING SOFT COMPUTING

Precise models Approximate models

Symbolic Traditional Functional

logic numerical Approximate approximation
reasoning modeling and reasoning and randomized
(traditional Al) search search

Figure 1-2 Problem-solving technologies.

 1.7 Summary

long-distance phone call providers. When the consumer picks up the phone and dials, an agent will com-
municate on the consumer's behalf with all the available network providers. Each provider will make an
offer that the consumer agent can accept or reject. A realistic goal would be to select the lowest avail
able price for the call. However, given the first round of offers, nerwork providers may wish to modity
their offer to make it more competitive. The new offer is then submitted to the consumer agent and the

process continues until a conclusion is reached. One advantage of this process is that the provider
can

dynamically alter its pricing strategy to account for changes in demand and competition, theretore max
imizing revenue. The consumer will obviously benefit from the constant competition berween providers.
Best of all, the process is entirely autonomous as the agents embody and act on the beliefs and con
straints of the parties they represent. Further changes can be made to the protocol so that providers
can bid low without being in danger of making a loss. For example, if the consumer choses to go
with the lowest bid but pays the second lowest price, this will take away the incentive to underbid or
overbid.
Much of the negotiation theory is based around human behavior models and, as a result, it is often trans-
lated using Distributed Artificial Intelligence techniques. The problems associated with machine negotiation
and
are as dificult to solve as they are with human negotiation and involve issues such as privacy, security
deception.

1.7 Summary
neural networks whose ability to learn by example makes them
The computing world has a lot to gain from
very flexible and powerful.
In case of neural networks, there is no need to devise an algorithm to performa
internal mechanisms of that task. Neural networks are
specific task, i.e., there is no need to understand the times, which are due
also well suited for real-time systems because of their fast response and computational
to their parallel architecture.
as neurology and psychology. They are
Neural networks also contribute to other areas of research such
and to investigate the internal mechanisms of the brain.
regularly used to modei parts of living organisms networks
networks is the possibility that someday "conscious"
Perhaps the most exciting aspect of neural that
believe that consciousness is a "mechanical" property and
might be produced. Today, many scientists
"conscious" neural networks area realistic possibility.
and handling data but has proven to be an excellent
Fuzzy logic was conceived as a better method for sorting
choice for many control system applications since it mimics human
control logic. It can be built into anything
It uses an inprecise but very
from small, hand-held products to large, computerized process control systems. when
like a human operator. It is robust and often works
descriptive language to deal with input data more
first implenented with little or no tuning
GAs can be used to search
When aplied to optimize ANNs for forecasting and classification problems,
for the right combination of inpur data, the most suitable forecast
horizon, the optimal or near-optimal
and the control parameters (learning rete,
network interconnection patterns and weights among the neurons,
criteria. Like ANNs,
momentum rate, tolerance level, etc.)
based on the training data used and the pre-set
an acceptable solution
cases, you can arrive at
GAs do not always guarantee you a perfect solution, but in many
without the time and expense of an exhaustive search.
term really entering general
circulation in 1994, coined by
Soft computing is a relatively new concept, the
it several fields of comput
Professor Lotfi Zadeh of the University of California, Berkeley, USA, encompasses
GAs. Neural networks are
The three that have been examined in this chapter are neural networks, FL and
ing. and GAs
FL for its exploitation of partial truth and imprecision,
important for their ability to adapt and learn,
 ntroduction to Fuzzy Logic,
lassical Sets and Fuzzy Sets 7
Learning Ubjectives
Definition of classical sets and fuzzy sets. Solved problems performing the operations
The various operations and properties of and properties of fuzzy sets.

clasical and fuzzy sets.

How functional mapping of crisp set can be
carried out.

7.1 Introduction to Fuzzy Logic

In general, the entire real world is complex, and the complexiry arises from uncertainty in the form of
ambiguity. One should closely look into the real-world complex problems to find an accurate solution, amidst
the existing uncertainties, using certain methodologies. Henceforth, the growth of fuzzy logic approach, to
is a form of
handle ambiguity and uncertainty existing in the complex problems. In general, fuzzy logic
This is in contradiction
multi-valued logic to deal with reasoning that is approximate rather than precise.
wIth crisp logic" that deals with precise values. Also, binary sets have binary or Boolean logic (either 0 or 1),
have a truth value that
Which finds solution to a particular set of problems. Fuzzy logic variables may ranges
berween 0 and 1 and is not constrained to the two truth values of classic propositional logic. Also, as linguistic
functions.
arables are used in fuzzy logic, these degrees have to be managed by specific
9 complexity ofa system increases, it becomes more difficult and eventually impossible to make a precise

the fuzay
where logic method born in
ement about is behavior, eventually arriving point ofcomplexiry
at a

biumans is the only way to get at the problem.

Ph.D., University of California, Berkeley)
(Originallyidentified and set forth by Lotfi A. Zadeh,
with uncertainty.
Fuzzy lo, Introduced in the year 1965 by Lotfi A. Zadeh, is a mathematicalcorrelated:dealing
tool for
I "The closer looksone
Dr. Zade states that the Principle of complexity and imprecision are

ucn otfers soft computing paradigm

at a r becomes its solution." Fuzzy logic
problem, the fuzzier to deal with imprecision
and
the .world with words. It provides a technique
such as
inf tnt concept of computing mechanism for representing linguistic
constructs

provides a
O n granularity. The fuzzy theory
that enables

In general, fuzzy logic provides an inference structure crisp

medium," "tall," "many."
describes
the traditional binary set theory
On the contrary, it an event will
numan reasoning capabilities. theory to explain
ren do or do not occur. It
uses probability is based
a Is, events that either The theory of fuzzy logic
OC is expected tooccur.
The utility
the chance with which given
event
a
ung functions of cognitive processes.
so are the
upon n o t i o n of relative graded membership and
 252 Introduction to Fuzzy Logic, Classical Sets and
and F-
Fuzzy Sets

Imprecise and vague data Decisions

Fuzzy Logic System

Figure 7-1 A fuzzy logic system accepting imprecise data and providing a decision,

offuzzy sets lies in their ability to model uncertain or ambiguous data and to provide suitable decisian.
Figure 7-1. ons as in

Though fuzzy logic has been applied to many fields, from control theory to artificial intelligence
remains controversial among most statisticians, who preter Bayesian still
logic, and some control engineers, who
prefer traditional two-valued logic. In fuzzy systems, values are indicated by a number (called a
ranging from 0 to 1, where 0.0 represents absolute falseness and 1.0 represents absolute truth.truth valuel
range evokes the idea of probability, fuzzy logic and fuzzy sets operate quite
While thie
difterently
Fuzzy sets that represent fuzzy logic provide means to model the uncertainty associated with
from probability,

imprecision and lack of information regarding a problem or a plant or a system, etc. Consider thevagueness,
of a "short person". For an individual X, a short meaning
person may be one whose height is below 425". For
individual Y, a short person may be one whose other
height is below or equal to 3 90". The word "short is called a
linguistic descriptor. The term "short" provides the same meaning to individuals Xand Y, but it can
that they both do not provide a be seen
unique definition. The term "short" would be conveyed effectively
a
computer compares he given height value with the pre-assigned value of "short". This only when
called as linguistic variable which variable "short" is
represents the imprecision existing in the system.
The basis of the theory lies in
to 1.0. The
making the membership function lie over a range of real numbers from 0.0
fuzzy set is dharacterized by (0.0,0,1.0). Real world
is vague and assigning rigid values to linguistic
variables means that some of the
arise from ignorance, from chance and
meaning and semantic value is
invariably lost. The uncertainty is found to
the fuzziness
randomness, due to lack of knowledge, from
vagueness (unclear), like
existing in our natural language. Dr. Zadeh proposed the set
decisions when uncertainty occurs. Consider
the "short" example discussed
membership idea to make suitable
a
height equal to or less than 4 feet, then 3 90" would easily become the previously. If we take "short as
will not be a member of the set member of the set "short" and 4 2
"short." The membership value is "1" if
is not a member of the set. it belongs to the set and "0 it t
Thus membership in a set is
member of a set or not. It can be found to be
binary. that is, either the element is 2
indicated as

XAo)= *EA
0, x A
where XA (x) is the
membership of element x in the
Ifit is said that the height is 56" (or 168 cm), oneset A and A is the entire set on the universe
it asshort or not short (i.e., might think a bit before deciding whether to ider
tall). Moreover, one
make the statement
"John is short", and give it a might reckon it as short for a man
truth value of 0.70. If 0.70
ma but tall for a woman.
alue, it
would be read as "There is a 70%
short or not short, and there
chance that John is short," represented a
proba ither
exists 70% chance meaning
of knowing which group he
that itisis still
s believed thatJohn is
actually translates to "John's degree of belongs to. But
ru
inology
that if all the
(fuzzy set of) short people membership
are
in the set of short
considered and lined up. people
is
0.70, y vay to the
shortest. In conversation, it is John positioned 700 o
is
demarcation between short and generally said that John is "kind of" short s n o
d e t i n i t e

tal. This could be and recognize that theto 0.70,

where u is the stated
membership function. mathematically as SHORTEN
 7.1
Introduction to
Fuzzy Logic

253
1
Tall
Membership
0.5

150 180 210

Figure 7-2 Graph Height (cm)
showing mem
mbership functions for fuzzy set "tall.
Fuzzy logic operates on the
rranslated concept of
member of themembership.
Elizabeth
be tra as is a For example, the
statement
the membership
vhere fu is
function that can
set of old
people and can be written "Elizabethasis old ca
mbership. In Figure 7-2, the objective termreturn a value between 0.0 and 0.1 symbolically ulOLD ot
an does not belong to the
person
"tall" has been
fuzzy class while for above
depending on the
assigned fuzzy values. At 150 cm and degree
However, between 150 and 180 cm, the 180, the person
certainly below, a
urve varying linearly between 0 and 1. Thedegree of membership for the class "tall" belongs
to
category "tall.
cu
can be
fuzzy concept "tallness" can be extended into assigned from the
all as shown in Figure 7-3. This is different
and
from the "short," "medium
nrobability of an occurrence of an event (Elizabeth probability approach that gives the degree o
The membership was extended to being old, in this instance).
possess various "degrees of
o.11.Zadeh formedfuzay sets as the sets on the universe membership" on the real continuous interval
Xwhich can accommodate
The concept of a fuzzy set contrasts with the classical of
"degrees of membership."
concept a bivalent set (crisp set) whose boundary is
required to be precise, that is, a crisp set is
collection of things for which it is known
a

anv given thing is inside it or not. Zadeh irrespective of whether

generalized the idea of a crisp set by extending a valuation set {1,0}
definitely in/definitely out) to the interval of real values (degrees of membership) between 1 and 0, denoted
s 0,1]. We can say that the degree of
membership of any particular element ofa fuzzy set expresses the degree
of compatibility of the element with a concept represented by fuzzy set. It means that a fuzzy set A contains
an object.x to degree alt), that is, als) = Degree(r e A), and the mapa: X-Membership Degrees) is called
The fuzzy setA can be expressed as A= {(% als)}, x E X; it
it unction or a
membership funetion. imposes
elastic constrain of the possible values of elements x E X, called the possibility distribution. Fuzzy sets tend to
an

Short Medium Tall

Membership
0.5

150 180 210

Height (cm)
"medium"' and "tall.
for fuzzysets "short,"
Figure 7-3 Graph showing membership functions
 Introduction to Fuzzy Logic, Classical Set
254 a Sets and
X- universe of discourse
Fuzy Sets

Figure 7-4 Boundary region of a fuzzy set.

capture vagueness exclusively via membership functions that are mappings from a given universe ofdice
Xto a unit interval containing membership values. It is important to note that membership can ta ourse
berween 0 and 1. values
Fuzziness describes the ambiguity event and randomness describes the
of an uncertainty in the occurrenceof
an event. It can be generally seen in classical sets that there is no
uncertainty, hence they have crisp boundaries
but in the case of a fuzzy set, since uncertaintry occurs, the boundaries may be ambiguouslyspecitied
From Figure 7-4 it can be noted that "d" is clearly a member of fuzzy set P, "e" is not a
of fuzzy set Pand the membership of "" is found to be vague. Hence "d can take membership value 1,
clearly memie
can take
membership value 0 and "b' can take membership value between 0 and 1 [0 to 1], say 0.4, 0.7, etc
This is said to be a partial membership of fuzzy set P
The membetship function for a set maps each element of the set to a membership value berween 0
and 1 and uniquely describes that set. The values 0 and I describe "not belonging to" and "belonging to' a
conventional set, respectively; values in between represent "fuzziness." Determining the nmembership function
to varying
degrees of the
is subjective depending
on the situation. It on an individua>'s
depends perception
in and does randomness. This concept is important and distinguishes fuzy
data question
theory from probability theory
depend
not on set

Fuzzy logic also consists of fuzzy inference engine or fuzzy rule-base to perform approximate reasonl
somewhat similar (but much seems
to
primitive than) that of the human brain. Computing with wo
more
to be a
slightly futuristic phrase today since only certain aspects of natural language can be repie
the calculus of
fuzzy sets; still fuzzy logic remains one of the most practical ways to mimc pert
h
in a realistic manner.
The fuzzy approach uses a premise that humans dont
represent
eg clas of bald men" or the "class of numbers which are much greater than 50") as fully dis
non
atner as sets in which there may be grades of membership intermediate between full membct
n t u a t n v e

membership. Thus, a fuzzy set works as a that makes it possible fuzziness in

a

manner.
concept to
treat fu2
breat
"IFAD
Fuzzy sets form the building blocks for fuzzy IF-THEN rules which have the
the general
form
lHEN Y s B, "where A and B are genct to systes
that.

fuzzy sets. The term "fuzzy systems reters mostly

systems" refers mosuy ereas theTHEN
governed by fuzzy IF-THEN rules. The lF
part is called a part of an implication is called the ba The

A
consequent. fuzzy system is a set of 0
to outputs
outputs. combines

configuration of apure fuzzy system is shown in Figurefuzzy

rules that converts p inputs
o u r p u tspace

fuzzy IF-THEN rules into a 7-5. The fuzzy inference enge

mapping from fuzzy sets in the input space A O
 Sets (CnispSets)
255
C A s S C a S e t s
i n

Fuzzy Rule Base

sets in X Fuzzy Inference Engine

Fuzzy Fuzzy sets in Y

Figure 7-5 Configuration of a

pure fuzzy system.

Auuzy logic principles. From a knowledge representation viewpoint, a fuzzy IF-THEN rule is a
/ h a s e d

capturing knowledge that involves imprecision. The main feature of reasoning using these rules is
sheme.
for«

maiching capabiliry, which enables an inference to be made from a fuzzry rule even when the rules
is only partially satisfied.
ndiion
a stems, on one hand, are rule-based systems that are constructed from a collection of linguistic
n the oher hand, fuzzy systems are nonlinear mappings of inputs (stimuli) to outputs (responses), that
rtin ypes of fuzzy systems can be written as compact nonlinear formulas. The inputs and outputs can
Dumbers o r vecto vectors of numbers. These rule-based systems can in theory model any system with arbitrary
aCy, that is, they work as universal approximators.
The Achilles heel of a fuzzy system is its rules; smart rules give smart systems and other rules give less
maf Of cven dumb systems. The number of rules increases exponentially with the dimension of theinpu
AEe (number of system variables). This rule explosion is called the curse of dimensionality and is a general
noblem for mathematical models. For the last 5 years several approaches based on decomposition, (cluster)

nerging and fusing have been proposed to overcome this problem.

Hence, fuzy models are not replacements for probability models. The fuzzy models are sometimes found
to work better and sometimes they do not. But mostly fuzzy logic has evidently proved that it provides better
sluions for complex problems.

72 Classical Sets (Crisp Sets)

Basically, a set is defined as a collection of objects, which share certain characteristics. A classical set is a
colecion of distinct objects. For example, the user may define a classical set of negative integers, a set of
parsons with height less than 6 feet, and a set of students with passing grades. Each individual entity in a set
8 called a member or an element of the set. The classical set is defined in such a way that the universe of
dicourse is splitted into two groups: members and nonmembers. Consider an object x in a crisp set A. This
oyect.x is either a member or a nonmember of the given set A. In case ofcrisp sets, no partial membership
LuSS. A crisp set is defined by its characteristic function.
Let universe of discourse be U. The collection of elements in the universe is called whole set. The total
numbe of elements in universe Uis called cardinal number denoted by nu. Collections of elements within
e s e are called sets, and collections of elements within a set are called subsets.
We know that for a crisp set A in universe U:
L. An
object x is a member ofgiven set A (x ¬ A), i.e., x belongs to A.
2. An
object x is not amember
r ofgiven set A (x ¢ A), ie, x does not belong to A.
 256
Introduction to Fuz2y Logic,
There
lassical Sets
Classical
and
several for defining
are ways a set. A set
may be defined using one of
of following:
the foll. FU77 Sets
1. The list of all the members of a set may be given. Example
A =
{2,4,6, 8, 10)
2. The
properties of the set elements may be specified. Example
A= {xlx is prime number< 20}
3. The formula for the definition of a set may be mentioned. Example
A == =1 10,
to
where x;a
4. The set
may be defined on the basis of the results
of a logical operation. Example
A= {x{x is an element
5. There exists a
belonging to PAND Q}
by the letter p and the
membership function, which may also be used to define
membership function for set A is given by (for allset. The membership iss denored
a
deno
a

values of x
Al)= 1 ifx ¬A
The set with
A)0 ifx¢A
of an
no elements is
defined as an empty set or
impossible
The set which
event is denoted
by a null
set, and the
null set. It is denoted
by symbol o. The occurence
consists of all occurrence of a certain event
possible subsets of a given set A is called a indicates a whole set
power set and is denoted as
PA) = {xl*CA)
For
crisp sets A and B
containing some elements in universe X, the
notations used are given
below
xEAx
A>
x
belongs to A
does not x

xEX>xbelongs tobelong
to A

For classical sets A and B on universe X

X, we also have some notations:
ACBA is
ACB A is completely
contained
contained in B (i.e., if x ¬
in or is A, then x E B)
A
BAC Band BCA equivalent
=
to B

7.2.1 Operations on
Classical
Classical Sets
sets can
be
difference. All these manipulated through numerous i n t e r s e c t i o n , complementand

operations are
defined and exr
perations such as union,
ope
explained in the following tions.
 Sets)
(Cmsp

ical
Sets
257

Figure 7-6 Union of two sets.

t 2 1 . 1 Uhion

nion berwe
een (wo sets gives all thoSe elements in the
universe that belong to either set A or set B or
The and
R The
B. The union operation can be termed as a
logical OR operation. The union of two sets A
othsets
A
and
B1s
givenas

AUB= {alx ¬ Aor xE B}

The
union
ers Aand Bis illustrated by the Venn diagram shown in Figure 7-6.
12.1.2 I n t e r s e c t i o n

The intersection berween two sets represents all those elements in the universe that simultaneously belong to
Luh the sets. The intersection operation can be termed as a logical AND operation. The intersection of sets
boch the .

Aand Bis given by

AnB= {z|x ¬ A and xe B}
The intersection of sets A and B is represented by the Venn diagram shown in Figure 7-7.

121.3 Complement
The complement of set A is defined as the collection of all elements in universe X that do not reside in set A,
ie, the entities that do not belong to A. It is denoted by A and is defined as

A= Mx 4 A,x EX}
where X is the universal set and A is a given set formed from universe X. The complement operation of set A
shown in Figure 7-8.

(0.)
Figure 7-7 Intersection of rwo sets.

Figure 7-8 Complement of set A.

 258 Introduction to Fuzzy Logic, Classical
eical Sets anct
and
Fuzz Sets

(A) (B)
Figure 7-9 (A) Difference A|B or (A- B); (B) difference B\A or (B- A).

7.2.1.4 Difference (Subtraction)

The difference of set A with respect to set B is the collection of all elements in the
universe that belono
but do not belong to B, i.e., the difference set consists of all elements that belong to A but do nor A
B.It is denoted by A|B
or A B and is given by
-

ng to

AB or (A B) = {xlx E A and x B} = A - (ANB

The vice of it also be

versa can
performed
BA or (B -

A) =
B--(BNA) =
{rlx ¬ Band x ¢ A}}
The above operations are shown in Figures 7-9(A) and (B).
7.2.2 Properties of Classical Sets
The important properties that define classical sets and show their similarity to
fuzzy sets are as follows:
1. Commutativity

AUB= BUA; AnB =BNA

2. Associativity

AU(BUC) =
(AU B)UC; An (Bn C) =
(ANB)nC
3. Distributivity

AU(BNC) (AU B)nAUC) =

An(BUC) =(AnB) U (ANC)

4. Idempotency

AUA =A; AnA =A

5. Transitivity

IfAC BC Cthen
6. Identity AgC

AUD= A, An¢=¢
AUX=X, AnX=X
 (Crisp Sets)
Classicalsets
19 259
Involution
(doubl
negation)
7.

=A
excluded middle
8.
Law of
AUA = X
contradiction
9, Law of
AnA =¢
10. DeMorgan's law

ANB= AUB; AU BI = AnB

From the properties mentioned above, we can observe the duality existing by replacing o, U, n with
yO.U, respectively.
lt is important to know the law
of excluded middle and the law of contradiction.

7.2.3 Function Mapping of Classical Sets

Mapping is a rule of correspondence between set-theoretic forms and function theoretic forms. A classical set
is represented by its characteristic function x (), where x is the element in the universe.
Now consider X and Yas two different universes of discourse. If an element x contained in X
corresponds
to an element y contained in Y, it is called mapping from Xto Y, i.e..
f: X> On the basis of this
Y. mapping.
the characteristic function is defined as

XACr)=, xeA

XA)=0, x4 A
where XA the membership in set A for element x in the universe. The membership
is concept represents
mapping from an element x in universe X to one of the two elements in universe Y (either to element 0
or 1). There exists a function-theoretic set called value set VA) for any set A defined on universe X, based on
the mapping of characteristic function. The whole set is assigned a membership value 1, and the null set is
assigned a membership value 0.
Let A and B be rwo sets on universe X. The function-theoretic forms of operations performed between
these two sets are given as follows:
1. Union (AUB)

XAUB) =XAl) VXBlx) = max{xa(), xBl*)}

Here V is the maximum operator.

2. Intersection (A n B)

XAnBx)=XAx)AXBx) =
min{xal), xsl)}
Here A is the minimum
3. Complement (A)
operator.

xa) = 1-xalx)
 260 Introduction to Fuzzy Logic, Classical
gical Seta
Sets and
4. Containment
Fuzzy Seta

IfAC B, then xalx) Xn(«)

7.3 Fuzzy Sets

Fuzzy sets may be viewed as an extension and generalization of the basic
concepts of
property of fuzzy set is that it allows partial membership. A fuzzy set is a set
crisp sets. An in
berween 1 and 0. The having degrees of importa
be member of other
membership in a fuzzy set need not be complete, 1.e., member of
one fuzzy c An also
set
fuzzy sets in the same universe. Fuz7y sets can be analogous to the thinking of
people. If a person has to be classified as friend or enemy, intell1gent people will not in
classification as friend or enemy. Rather, they will resort to n t
classify the person somewhere between two evt
friendship and enmity. Similarly, vagueness is introduced in fuzzy set by
that divide members from nonmembers in the eliminating the sharn ho mes of
group. There a gradual transition between full
is ndaries
and
nonmembership, not abrupt transition. memhe ip
A fuzzy setA in the universe
of discourse U can be defined as a set of ordered pairs and it is
given hu

where (x) is the degree of membership of x in Aand it indicates the

ofmembership ua(«) assumes values in the range from 0 to 1, i.e., the degree
that x belongs to The
4. deore
ree
or
4x) e [0, 1]. membership is set to unit
interval [0. 11
There are other ways of
representation of fuzzy sets; all representations allow partial membership to be
expressed. When the universe of discourse Uis diserete and finite, fuzzy set 4 is given as follows:

*2 3

where "" is a finite value. When the universe

of discourse Uis continuous and infinite, fuzzy set 4 is
given by

In the above
representations of fuzzy sets for discrete and continuous universe, the horizontal bar s
two
not a
quotient butdelimiter. The numerator in each
a
representation is the
universe present in the denominator. Formembership
is associated with the element value in set tn
discourse U, the summation
of the discrete and finite unive of
but indicates the collection
symbol in the
representation of fuzzy set A does not denote algebraic sumu ation

but rather it is a discrete

of each element. Thus the summation
sign ("+") used is not the algebral
0, the
function-theoretic union. Also, for continuous and infinite
integral sign in the representation of universe u
or

function-theoretic union for continuous variables. fuzzy set 4 is not an

algebraic integral but is a
l o u s

A fuzzy set is universal

fuzzy set ifand only if the value of the members
Any fuzzy set Adefined on a universe U ismembership
under consideration. for all the
function is 1tota
and Bare said to be sesd
equal fuzzy sets if Hslx) =uglx) for all e U.a subset of that universe. Iwo fuzzys
if and only if the value of the A fuzzy setA is said be
x

membership function is 0 for all to c niversal niversal

fuzzy set can also be called whole fuzzy set. posible members considercu.
 3 F U Z z yS e t s

261
c o l l e c t i o n
n
of allfuz
afall fuzzy sets and
fuzzy subsets on universeU is called
The
erlap,
can
dinality the ca of the fuzzy power set PMU).
fuzzy power set, npU) is infinite, Since all the
h u r z y
sets

basis oftheabove discussion we have i.e., npU) 0. =

the

4SU Hix) <uuu)

U
forall
rE
Ao.
Hox) = 0; ulx) = 1

113.1 Fuzzy Set Operations

tions on classical sets to
ralization of operatio
he
general operations fuzzy sets is not unique. The fuzzy set
on
erations being discu ussed in this section are termed standard fuzzy set operations. These are the operations
used in engineering ations. Let A and B be fuzzy sets in the universe of discourse U. For a given
p e r a

O an the
o n the
universe, the following unction theoretic operations of union, intersection and complement
mentx
and B on Uu
e defined for fuzzy
sets

23.1.1 Union

The union of fuzzy sets 4 and B, denoted by 4 UB, is defined as

max[ustr), ju4gl*)] =4,l*)V4glr) for all x e U
HAUglx) =

and Bis shown in

V indicates max operation. The Venn diagram for union operation of fuzzy sets
where
Figure 7-10.

23.1.2 Intersection

e intersection offuzzy sets 4 and 3,

denoted by 4n B, is defined by
for all e U
min[uglr), Hglx)] =u,t«) Ap()
A x

4anela) =

and Bis shown

diagram for intersection operation of fuzzy sets A
where A indicates min operator. The Venn
inFigure 7-11.

KA

sets 4 and .
Figure
7-10 Union offuzzy
 Introduction to Fuzzy Logic, Classical Sets and Fir:
262 Fuzzy Set
K

X
0

Figure 7-11 Intersection offuzzy sets and .

7.3.1.3 Complement
When 4z) e [0, 1), the complement of 4, denoted as 4 is defined by

zr) =
1-ug() for all x ¬ U

The Venn diagram for complement operation of fuzy set A is shown in Figure 7-12.

7.3.1.4 More Operations on Fuzzy Sets

1. Algebraic sum: The algebraic sum (+ offuzzy sets, fuzzy sets 4 and B is defined as

Ha+alor)=4)+gle)-ua)-ugl)
2. Algebraic product: The algebraic product (4 ) of two fuzzy sets 4 and B is defined as

Heel)=4aa)ngle)

****

******* *******

Figure 7-12 Complement of fuzzy set A.

 263

The bounded sum (4 ®) oftwo fuzzy sets 4 and B is defined as

S u m :

H4Ge) = min{1, ua(«)+/4p(x)}

bounaeaa rnThe
bounded diffcrence (4 O oftwo fuzzy sets 4 and Bis defined as
HyOE)=max{0, A,ls)-1Hg&))
Sets
Properties of Fuzzy
78.2
ofcontradiction.
follow the same properties
as crisp sets except for the law ofexcluded middle and law
Kuzyset
tor fuzzy
set4
Thatis,
4U4U 4n4#¢

Fræquentlyused properties
offuzzy sets are given as follows:

I.Commutativity

4U8=BU4 4nB=Bnd

2. Associativity
AUBUC) = 4UBUG
AnBnG) = 4nBng

3. Distributivity
AUBnG) = 4UÐn AUC)
AnBUC) = 4nBU4ng)

4.ldempotency
4Ud 4 4nd=4

5. ldentiry
Uuniversal set)
4UU=
AUd=4 and
4n= and 4nU=4

6. Involution (double negation)

7. Transitivity

If4SBE G
then dEC

De Morgan's law
JUR-nEJ0=4uB
 264
Introduction to Fuzzy Logic, Cassical Sets and Fuzzy Ss
Sets
7.4 Summary
In this
chapter, we
have discussed the basic
definitions, properties and operations on classical sets
Fuzzy sets are the tools that convert the
sets.
and furz
concept of fuzzy logic into algorithms.
partial membership, they provide
computer with such algorithms that extend binary
Since fuzzy sets allow
take human-like decisions.
In other words, logic and enable it to
fuzzy sets can be
thinking is transferred to a computer. One difference between thought of as a media
through which the human
do not follow the law
of excluded middle and law of
fuzzy sets and classical sets is
that the former
intersection and union contradiction. Hence, it we want to
choose fuzz
and
operations which satisfy these laws, then the operations will not
idempotency. Except the difference of set membership being an infinite valued satisfy distributiviry
binary valued quantity, fuzzy sets are treated in the same mathematical form as
quantiy instead of
classical sets.

7.5 Solved Problems

1. Find the power set and
X= cardinality of the given set (b) Intersection
{2,4,6). Also find cardinality of power set.
Solution: Since set X contains three
cardinal number is
elements, so its 4nB= min{uglo), ugla}
= 03
nx= 3
The power set of X is (c) Complement
given by
PX) =
{¢ ,
{2).14),{6}. (2,4), 4=1-4)= 9 0.5 0.8
4,6,(2,6, (2,4,6}} 6 8
The cardinality of power set PMX), denoted by npx)»
is found as

nAX)= 2 = 23 = 8
B=1-4e)=t6t3
(d) Difference
2. Consider two
given fuzzy sets
dl 4nB= | 2+4 6 8

-
0.4 Bld Bn=+
3. Given the
Perform union, intersection, two fuzzy sets
difference and com-
plement over fuzzy sets A and B. 0.75 0.3 0.15 0
Solution: For the given fuzzy
following
sets we have the 1.015t20t25 t 30
0.6 0.2 0.1
(a) Union 1.5 2.0T 25 T 3.0o
AUB= max{uar), ug(x)) find the following

040.5 (a) B UB: (b) 8nB:: () B:

(d) B2: ()BilB: ( U
 15Solved Problems
265

ip B nB: (h) Bi nB: () B U: Table 1

(k) B: UB Gain Detection level Detection level
)B nB2:
given fuzzy sets, we have the
setting of sensor 1 of sensor 2
Solution: Fo
0
following: 10 0.2 0.35
0.75 0.3 20
t 0.35 0.25
B 1.0 1.5 2.0 2.5 3.0 30 0.65 0.8
40 0.85 0.95
b)B8=1++0.20.1, 30
o 50

Now given the universe of

discourse X={0, 10,
0.25 0.7 0.85 1 20,30,40, 50} and the membership functions
e-10152.0 2.5 + 3.0 for the wo sensors in discrete form as

0.4 0.8 0.9 0.85 50

D- 10t 20
a -10 15+ 20+2 30
00.35 0.250.80.9
e)B11B2 = Bi2 10 20 30 40 50
1.05+ +0 find the following membership functions:

0.25 (a) upup); (6) upnee); () D ):

B- 0, 0.7 0.85
15 3.0 (d) ):

Bn-16+2+
Solution: For the given fuzzy sets we have
a &-1+152025t30 (a) Dup)
) BUBi =
0.75 0.7 0.85 =
max
{p, (), Px))
.5 T 2.0 2.5 - .35 + 0.95
30 40
) 0.2 0.1
Bn=+ 20+ 25t3.0 (b) upn )
= min {up, ) , up, ()}
( 0.8 0.9
u=
. It is
1t1.5t 2.0 2.5
necessary to compare two sensors baSed upon
3.0 -1
0.2
10
0.25
20
0.65
30
0.85
40
+
50
their
or
detection levels and gain
settings. The table (c)D
gain settings and sensor detection levels with
a
standard item being monitored = 1-up ()

cal membership values

providing typi 0.8 0.65 0.35 0.15
evels for each
sensor is
to

given
represent the detection
in Table l.
-

10 20
, 30 40 +50
 266 Introduction to Fuzzy Logic, Classical Sets and Fuzzy
Seto
(d) ) Find the following:
= l - u p , ()

0.65 0.75 0.2 0.05 0

(a) Plane U Train; (6) Plane n Train
-
10 20 30 40 T50 (c) Plane; (d) Train;

(e)pUD (e) PlaneTrgin; (F) Plane U

Train;
=
max{up, (x), u)} Plgne n Train: (h) Plane U
Plane;
= 0.65 0.65 0.85 1
) Plane n Plane; (G) Train U Train:
10 200 30 40
(k)Train U Train

min{p, (),"p}} Solution: For the given fuzzy sets we

have the
- + 2 0.35 0.35 following:
0.15 o ) uowiIng
20 30 40 50 (a) Plane U Train
=
max{wplane (*), uTrain ()}
max(4pa (), 4Dz*)} 1.0 0.5 0.4 0.8 0.2
.65 0.750.809+
20 0,95
bike boat Plane house
30 40 50 (b) Plane n Train
(h) p D )
= min{up), up(a)}
=
min{uplane (x), juTrain (t)}
=0.2, 0.2 0.3 0.5 0.1
0.35 0.25, 0.2 +

(i) upilD
-0 20 50
0.05
40
0
50
()Plane =1-4plane )
train bike boat plane house

0.8 0.5 0.7 0.2 0.9

Hpn)= min{up, ),) } train bike boat plane house
-J90.2 0.35 0.2 0.05
20 30 40 50 d) Train=1-4Train )
0.8 0.6 0.5 0.8

h p n D ) = min{up(), 4D lx)) train bike boat plane house

0 0.35 0.25
0.35 0.15 0 (e) Plane|Train
10 20
30 40350 = Plane Train
5. Design computer software to perform image
a

processing to locate objects within a scene. The

=
min{uPlane (x), Tin)
0.5 0.3 0.5 0.1
fuzzy
rwo

image are:
sets
representing plane a and a train
-in train bike boat plane
house

Plane=0.2
+
0.5 0.3 0.8 o.1 (f) Plane UTrzin
bike boat plane house =l-
0.2 0.4
max{4plane (x), UTrain (*)}
0.5 0.2 0.8
Traintrain bike boat = 4 0.5 0.6 0.2
plane T house
plane house trainbike T boat
 15SolvedProblems

267
n Train Solution: For the
Plane
(g -1 - min|u Plane 7),jATHain(a)) two given fuzzy sets we have the
following:
0.8 0.8 0.7 0.5 0.9 (a) 4UB
train Dike boat *plane *house
= max{4al), ugl«)}
U Plane
(h) Plane = 0.75 0.5
max|4Plane(W),MPlne t)} 0.64 0.645
0.8 0.5 0.7 0.8 0.9 0.65 0.655 0.66
0.9
trainbike boat plane house (b) 4nB
)PlanePlane = min{ug),. pgl)}

min{4Pane), 4lnel)] 0 0.25 0.75 0.5 0

0.20.5 0.3 0.2 0.1 0.64 0.645 0.650.655
train Tbike boat +
plane house|
0.66
(A=1-4
6) Train U Train
max{4 Train (t), Train 0.25 0 0.5
- 1.0 0.80.6 0.5 0.8 0.64 0.645 0.650.655 0.66
rain bike boat plane house

(k) Trgin Train (d)=1-Hgl)

=
min{uTrain (), Train) 0.75 0.25 0 0.5

0.20.4 0.5 0.2 0.64 0.645 0.65 0.655 0.66

T
-
Train bike boat T plane house

6. For aircraft simulator data the determination of (e) AUB

certain changes in its operating conditions is made
on the basis of hard break points in the mach = 1-max{ua*), ug(x)
region. We define two fuzzy sets A and rep
the condition of "near" a mach number
0.25 0
+ 0.5
resenting 0.64 0.645 0.65 0.655 0.66
of 0.65 and "in the region" of a mach numberof
0.65, respectively, as follows
( 4n
= near mach 0.65
= l - min{ug(x), pglx))
0.5
= 0.75 1 0.75 0.25 0.5 +06
|0.64 0.645 0.65 0.655 0.66
0.66
i n the region ofmach 0.65 0.64 0.645 0.65 0.655

0 0.25 0.75
+ 0.5 7.
7. For
o r the
the owo give fuzzysets
twogiven
0.64 0.645 0.65 0.655 0.66

For these two sets find the following

(a)4uB: (b)4nB: (c) 4: tTt,t,ti

()4UB: f) 4nB
 sical Sets
Introductionto Fuzzy Logic, Classical Sets and
268 and F Fuzzy Sets
find the following: Gi) 4B= minlu,*), 14p)}
(a)4U b)4nB: (c)4: 02 +
d) E )4U: (E) 4n4: (G) 4UB = max{ualo),4g»)}
BUB (h)Bn ) 4nB
)4U (k)BnA: )BUA: -0 2
+ 0.7
(m)4UB (n) An
(kn4= min{uglt), uz)}
Solution: For the given sets we have:
- +
(a)4UB=max{4gl), uglx)}
- + ) BUA= max{upl). uZa))
-

(b) 4nB= minlug), uglr)}

(m) 4UB =1 max{u), ng}
-

-
(c) =1-ul)

-
+ + (n) AnB= min{4z), Hzl«)}

d) = 1-4g(*)
8. Let U be the universe of
military aircraft of
interest as defined below:

(e) 4U4 maxlu,l), uzt))

= U= lal0, b52, e130. f2. f9)
=090.80.6 0.6 1) Let be the fuzzy set of bomber class aircraft:
-++2t 0.3 0.4 0.2 0.1
() 4n4= minfu,). 4za)) 4a10T 52 130 2
Let be the fuzzy set of fighter class aircraft
0.1 0.2 0.8 0.7
gBUB= max{ugl«), agla)} a10 52 cl30 2 t9
- +}| Find the following:

(h) (a) 4U: (b)4nB ()4: (d

8n= minl4gl). uglx))
(e) 4le: (f)BId UE
- 00.5 0.3 0.3
(h) 4n )au )BU4
75
Solved Problems

Solution: We have
9. 269
AUB= max(mao), 4g(x)} Consider two fuzzy sets
0.4 0.8
=

a10 b52 0.7

c130 2 f9
-
b) 408-min{ngt),4gl)
Find the
J0.10.2 0.20.1+9
- algebraic sum, algebraic product,
bounded sum and bounded
difference of the given
fuzzy sets.
Solution: We have
0.7 0.6 0.8 0.9 (a) Algebraic sum
a10 b52 c130 f2 Ha+a)= l4gl)+ugl)]-lux)-uglx)]
) =1-4gl«) = 06,051
0.90.8 0.2 0.3 1 0.02 0.06
|al10b52 c130 + 30.08 0.5

(e 41B=AnB= minlug), 4z)}

=.340.4 0.2 0.1
a10 b52 l30tt (b) Algebraic product

(f Bl4 =BnA min{4e), Lz}

=

0.06 0.08 0.5

+
b52 c130 (c) Bounded sum

AUB=1 -max{ulx), pla)} H4eglr)

0.3 = min[1. 44x)+4p)]
-J0.7 0.6 0.2
a105* a30 t
hAnB=1 -min{ksl*), ngl*))
0.8 0.8 0.9 1
.9
a1052t d30t tol (d) Bounded difference

) u= max(4z*), 4go) H4oplx)

0.9 = max[0, u,(v)-ugl«)]
0.8 0.8
0.5
alo52 d30f2 "f9
9 BUA = max{#plx), Hala)}
0.2 0.3
0.8
09
a10 b52
 Introduction to Fuzzy Logic, Classical
Sets and
270
Fuzzy
Sets
membership functions for a (b) Algebraic product
10. The discretized
transistor and a resistor are given below:

=uJtx):up(x)
- 9 -
0.02

(c) Bounded sum

Find the following: (a) Algebraic sum; (b) alge-
braic product; (c) bounded sum; (d) bounded =
min{1, uTlx) +ugl)}
difference.
= min
1+40 1.0
Solution: We have
3
(a) Algebraic sum 1.3
4
HT+Rla)
00.3 1.0,1.0
=
lupt«)+ug(«)] -

[uztr)-ugl«)] 1 3
J0 0.3 1.0, 1.0
-

2+ (d) Bounded difference

0.21 0.16 0.3 HTOpl)
2 T3 4 max{0, 4Tt)-u gl«))
0.51
+ = max

0.28 0.79 0.84, 0.94

- +
4

-190.10.40.605 05
-
2 3t 4t5

7.6 Review Questions

1. Define classical and

2. State the
sets
fuzzy sets. 8. Justify the following statement: "Partual
nem

importance of fuzzy sets. bership is allowed in fuzzy sets."

3. What are the methods of
erties

representation of a 9. Discuss in detail the operations and prop

classical set?
of fuzzy sets.
4. Discuss the Venn

operations of crisp sets. 10. Represent the fuzzy sets operato

using
5. List the
properties of classical sets. diagram.
6. What is meant Whetherd

by characteristic function? 11. What is the cardinality of a fuzy set?

7. Write the function set:
theoretic form power ser can be formed for a fuzzy
of crisp set
operations. representation 12, Apart from basie
12. basic rations,
state
few
ocher

operations involved in
opetrfuzzy sets.