Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
Lecture#6
Fuzzy Relations & Compositions
Dr Mahmud Iwan S
SCHOOL OF ENGINEERING
UCSI UNIVERSITY
~EK406 DrMIS~
Outlines
Introduction
Fuzzy Relations
Fuzzy Compositions
~EK406 DrMIS~
Introduction
Basic Concept
Mathematically, a relation implies the presence of an
association between elements of different sets (at least
two fuzzy sets).
~EK406 DrMIS~
Introduction
Basic Concept
~EK406 DrMIS~
Introduction
Classical Relation
Relations are basic notion in engineering and
science which is essentially the discovery of relation
between observations and variables.
Classical relations are structure that represent the
presence or absence of correlation or interaction
among various sets.
Two degrees of relationship in a crisp relations:
Completely related
Not related
~EK406 DrMIS~
Introduction
Classical Relation
The strength of this relationship is measured by
characteristic function as follows:
1, ( x, y ) XxY
XxY ( x , y ) =
0, ( x, y ) XxY
=1, is associated with complete relationship
=0, is associated with no relationship
~EK406 DrMIS~
Introduction
Classical Relation
A binary relation is a classical relation between two sets X
and Y denoted by R(X,Y).
Matrix relation R
~EK406 DrMIS~
Introduction
Classical Relation
Example
Let ={Adam, Bernhard, Charles} and ={Diana, Eva,
Francisca}, then the relation married to on is for
example:
~EK406 DrMIS~
Introduction
Classical Relation
Example of classical relation representation:
Let X={Adam, Bernhard, Charles} and Y={Diana, Eva,
Francisca}, then the relation married to on is for
example D E F
A 0 1 0
B 0 0 1
R=
C
1 0 0
Matrix relation
~EK406 DrMIS~
Fuzzy Relations
Basic Concept
Fuzzy relation is based on the philosophy that
everything is related to some extend or unrelated.
Unlike crisp relations, the strength of the relation
between ordered pair of two universes is not
measured with the characteristic function, but rather
with a membership function expression expressing
various degree of strength of the relation on the unit
interval [0,1].
Crisp relations and fuzzy relations can be defined in
terms of subsets.
~EK406 DrMIS~
Fuzzy Relations
Example 1
Let X={x1,x2}={NYC, TKO} and Y= ={y1,y2}={TPE, HKG, BJI}.
Let R represent the relation very close.
Crisp relation
y1 y2 y3
x1 0 0 0
x2 1 1 1 ?
~EK406 DrMIS~
Fuzzy Relations
~EK406 DrMIS~
Fuzzy Relations
~EK406 DrMIS~
Fuzzy Relations
Example 1
Let X={x1,x2}={NYC, TKO} and Y= ={y1,y2}={TPE, HKG, BJI}.
Let R represent the relation very close.
Fuzzy relation
y1 y2 y3
x1 0.3 0.1 0.1
x2 1 0.7 0.8
~EK406 DrMIS~
Fuzzy Relations
R = {((x , y ), R (x , y ))}
or
R (x i , y i )
R= For discrete
( x i , y i )XY (x i , y i )
R (x i , y i ) For continuous
R=
X Y
(x i , y i )
~EK406 DrMIS~
Fuzzy Relations
Example 1
Let X={x1,x2}={NYC, TKO} and Y= ={y1,y2}={TPE, HKG, BJI}.
Let R represent the relation very close.
Fuzzy relation
y1 y2 y3
R( X , Y ) = .... ? by listing ordered pair
x1 0.3 0.1 0.1
x2 1 0.7 .8
~EK406 DrMIS~
Fuzzy Relations
Example 1
Let X={x1,x2}={NYC, TKO} and Y= ={y1,y2}={TPE, HKG, BJI}.
Let R represent the relation very close.
R ( x1 , y1 ) R ( x1 , y 2 ) L R ( x1 , y m )
( x , y ) ( x , y )
L R ( x2 , y m )
R 2 1 R 2 2
R( X , Y ) =
M M O M
R ( xn , y1 ) R ( xn , y 2 ) L R ( xn , y m )
Matrix Relation
~EK406 DrMIS~
Fuzzy Relations
Example 2
Let X={x1,x2}={NYC, TKO} and Y= ={y1,y2}={TPE, HKG, BJI}.
Let R represent the relation very close.
Fuzzy relation
y1 y2 y3
x1 0.3 0.1 0.1 R( X , Y ) = .... ? by matrix relation R
x2 1 0.7 .8
~EK406 DrMIS~
Fuzzy Relations
~EK406 DrMIS~
Fuzzy Relations
1 1 L 1
1
1 L 1
RE =
M M O M
1 1 L 1
~EK406 DrMIS~
Fuzzy Relations
0 0 L 0
0
0 L 0
RE =
M M O M
0 0 L 0
~EK406 DrMIS~
Fuzzy Relations
~EK406 DrMIS~
Fuzzy Relations
Example 3:
Consider two fuzzy sets A and B given below, determine the
Cartesian product between these sets.
~EK406 DrMIS~
Fuzzy Relations
Exercise: (T.J. Ross, page 60)
Suppose we have two fuzzy sets, T defined on a universe of
three discrete temperature, X={x1,x2,x3), and P defined on a
universe of two discrete pressures, Y={y1,y1}. Let say, T
represent ambient temperature and P the near optimum
pressure and the Cartesian Product might represent the
efficient condition of operation. Find the efficient condition ?
Introduction
Let R be a relation that relates, or maps elements from
universe X to universe Y, and let S be another relation that
relates elements from universe Y to universe Z.
~EK406 DrMIS~
Fuzzy Compositions
X Y Y Z
x1 y1 y1
x2 y2 y2 z1
x3 y3 y3 z2
y4 y4
R S
T??
~EK406 DrMIS~
Fuzzy Compositions
X Z
x1 y1 y1
x2 y2 y2 z1
x3 y3 y3 z2
y4 y4
T??
~EK406 DrMIS~
Fuzzy Compositions
Composition is very important for differencing procedure
used in linguistic description and is particularly useful in
fuzzy control and expert system as we shall discuss in next
classes.
There are several type compositions as
Max-min composition (most common in application)
Min-max composition
Max-product composition
Max-average composition
The computation of fuzzy composition is like a matrix
multiplication with replacement as:
for example max-min composition, sum is replaced by
max and product is replaced by min operator.
~EK406 DrMIS~
Fuzzy Compositions
Max-
Max-min Composition
Suppose we have two fuzzy relations R1(x,y) and R2(y,z)
defined over the Cartesian product.
[
R1 ( x, y ) R2 ( y, z ) ]
y
R1 o R2
X Z
( x, z )
~EK406 DrMIS~
Fuzzy Compositions
Max-
Max-min Composition
R o R ( x, z ) = [ R ( x, y ) R ( y, z )]
1 2 1 2
y
~EK406 DrMIS~
Fuzzy Compositions
Min-
Min-max Composition
Suppose we have two fuzzy relations R1(x,y) and R2(y,z)
defined over the Cartesian product.
[
R1 ( x, y ) R2 ( y, z ) ]
y
R1 R2
X Z
( x, z )
~EK406 DrMIS~
Fuzzy Compositions
Max-
Max-product Composition
Suppose we have two fuzzy relations R1(x,y) and R2(y,z)
defined over the Cartesian product.
[
R1 ( x, y ) R2 ( y, z ) ]
y
R1 R2
X Z
( x, z )
~EK406 DrMIS~
Fuzzy Compositions
Max-
Max-product Composition
R R ( x, z ) = [ R ( x, y ) R ( y, z )]
1 2 1 2
y
~EK406 DrMIS~
Fuzzy Compositions
Max-
Max-average Composition
Suppose we have two fuzzy relations R1(x,y) and R2(y,z)
defined over the Cartesian product.
1
(
R1 ( x, y ) + R2 ( y, z ) )
R1 + R2
y 2
X Z
( x, z )
~EK406 DrMIS~
Fuzzy Compositions
Max-
Max-average Composition
1
( )
R1 + R 2 (x , z ) R1 ( x , y) + R 2 ( y, z)
y 2
~EK406 DrMIS~
Fuzzy Compositions
Example 4
Let X={Peter, Mary, John} represents students who plan to take four
courses (Z) offered based on their preferences for Y={y1,y2,y3,y4}={theory,
application, hardware, programming}.
Z={Fuzzy Theory, Fuzzy Control, Neural Network, Expert System}.
Assume the students interests is represented by the fuzzy relation P(X,Y).
The properties of the course are indicated by the fuzzy relation Q(Y,Z}.
Use max-min composition to make decision.
~EK406 DrMIS~
Fuzzy Compositions
Example 5
~EK406 DrMIS~
Fuzzy Compositions
Exercise
0.9 0.0 0.3 0.4
0.2 1.0 0.8 0.0
0.1 0.2 0.0 1.0 0.7
R = 0.8
R 1 = 0.3 0.5 0.0 0.2 1.0 2 0.0 0.7 1.0
0.8 0.0 1.0 0.4 0.3 0.4 0.2 0.3 0.0
0.0 1.0 0.0 0.8
Find the max-min composition, max-product composition, max-
average composition.
~end~
~EK406 DrMIS~