FUZZY SYSTEMS

Fuzzy Sets Operations

Intersection (A  B)

In classical set theory the intersection of two sets contains

those elements that are common to both

The value of the elements in the intersection:

µA  B(x) = min [µA(x), µB(x)]

e.g. Tall = (0/5, 0.1/5.25, 0.5/5.5, 0.8/5.75, 1/6)

Short = (1/5, 0.8/5.25, 0.5/5.5, 0.1/5.75, 0/6)
Tall  Short = (0/5, 0.1/5.25, 0.5/5.5, 0.1/5.75, 0/6)
= Medium
1
 FUZZY SYSTEMS

Fuzzy Sets Operations

Union (A  B)

In classical set theory the union of two sets contains those

elements that are in any one of the two sets

The value of the elements in the union:

µA  B(x) = max [µA(x), µB(x)]

e.g. Tall = (0/5, 0.1/5.25, 0.5/5.5, 0.8/5.75, 1/6)

Short = (1/5, 0.8/5.25, 0.5/5.5, 0.1/5.75)
Tall  Short = (1/5, 0.8/5.25, 0.5/5.5, 0.8/5.75, 1/6)
= not Medium
2
 FUZZY SYSTEMS

Fuzzy Sets Operations

Complement (A)

The value of complement of A is:

µ  A(x) = 1 - µA(x)

e.g. Tall = (0/5, 0.1/5.25, 0.5/5.5, 0.8/5.75, 1/6)

 Tall = (1/5, 0.9/5.25, 0.5/5.5, 0.2/5.75, 0/6)

3
 FUZZINESS IS NOT PROBABILITY

• Probability is used, for example, in weather forecasting

• Probability is a number between 0 and 1 that is the

certainty that an event will occur

• Fuzziness is more than probability; probability is a

subset of fuzziness

• Probability is only valid for future/unknown events

• Fuzzy set membership continues after the event

 FUZZY SYSTEMS

Fuzzy Relations
The fuzzy set operators allow rudimentary reasoning about
facts

For example, consider the three fuzzy sets tall, good_athlete

and good-basketballplayer. Now assume

If we know that a good basketball player is tall and is a

good athlete, then which one of Peter or Carl will be the
better basketball player?
5
 FUZZY SYSTEMS

Fuzzy Relations
Through application of the intersection operator, we get

Using the standard set operators, it is possible to

determine that Peter will be better at the sport than Carl

The example above is a very simplistic situation. For

most real-world problems, the sought outcome is a
function of a number of complex events, or scenarios
6
 FUZZY SYSTEMS

Fuzzy Relations
For example, actions made by a controller are determined
by a set of if-then rules. The if-then rules describe situations
that can occur, with a corresponding action that the
controller should execute

It is, however, possible that more than one situation, as

described by if-then rules, are simultaneously active, with
different actions. The problem is to determine the best
action to take

A mechanism is therefore needed to infer an action from a

set of activated situations
7
 FUZZY SYSTEMS

Fuzzy rule based reasoning system

8
 FUZZY SYSTEMS

Fuzzy rule based reasoning system

For fuzzy systems in general, the dynamic behavior of that
system is characterized by a set of linguistic fuzzy rules

These rules are based on the knowledge and experience of a

human expert within that domain. Fuzzy rules are of the
general form
if antecedent(s) then consequent(s)
The antecedents of a rule form a combination of fuzzy sets
through application of the logic operators (i.e. complement,
intersection, union)

The consequent part of a rule is usually a single fuzzy set,

with a corresponding membership function 9
 FUZZY SYSTEMS

Fuzzification
The antecedents of the fuzzy rules form the fuzzy “input
space,” while the consequents form the fuzzy “output
space”

The input space is defined by the combination of input

fuzzy sets, while the output space is defined by the
combination of output sets

The fuzzification process is concerned with finding a fuzzy

representation of non-fuzzy input values. In which, input
values from the universe of discourse are assigned
membership values to fuzzy sets
10
 FUZZY SYSTEMS

Fuzzification

For example the fuzzy sets A and B, and assume the

corresponding membership functions have been defined
already

Let X denote the universe of discourse for both fuzzy sets.

The fuzzification process receives the elements a, b € X, and
produces the membership degrees µA(a), µA(b) and µB(a),
µB(b)

11
 FUZZY SYSTEMS

Inferencing

The task of the inferencing process is to map the fuzzified

inputs to the rule base, and to produce a fuzzified output

For the consequents in the rule output space, a degree of

membership to the output sets are determined based on the
degrees of membership in the input sets and the relationships
between the input sets

The output fuzzy sets in the consequent are then combined to

form one overall membership function for the output of the
rule
12
 FUZZY SYSTEMS

Fuzzy Relations

Classical relation between two universes

U = {1, 2} and V = {a, b, c} is defined as:

a b c

R=UxV= 1 1 1 1
2 1 1 1

13
 FUZZY SYSTEMS

Fuzzy Relations

Fuzzy relation between two universes U and V is defined as:

µR (u, v) = µAxB (u, v) = min [µA (u), µB (v)]

i.e. we take the minimum of the memberships of the two

elements which are to be related

14
 FUZZY SYSTEMS

Fuzzy Relations

Example:

Determine fuzzy relation between A1 and A2

A1 = 0.2/x1 + 0.9/x2
A2 = 0.3/y1 + 0.5/y2 + 1/y3

15
 FUZZY SYSTEMS

Fuzzy Relations
A2
R = R(A1, A2)
a23
(1.0)
= 0.2 0.2 0.2
0.3 0.5 0.9 0.2
a22 0.9
(0.5)
0.2 0.5
a21
(0.3)
0.2 0.3
a11 a12 A1
(0.2) (0.9)

16
 FUZZY SYSTEMS

Fuzzy Relations

Example: The fuzzy relation R is

R = A1 x A2 = 0.2 x 0.3 0.5 1

0.9

= min(0.2, 0.3) min(0.2, 0.5) min(0.2, 1)

min(0.9, 0.3) min(0.9, 0.5) min(0.9, 1)

= 0.2 0.2 0.2

0.3 0.5 0.9

17
 FUZZY SYSTEMS

Fuzzy Rules

Relates two or more fuzzy propositions

If X is A then Y is B

e.g. if height is tall then weight is heavy

X and Y are fuzzy variables

A and B are fuzzy sets

18
 FUZZY SYSTEMS

Fuzzy Associative Matrix

So for the fuzzy rule:

If X is A then Y is B

We can define a matrix M(nxp) which relates A to B

M=AxB

It maps fuzzy set A to fuzzy set B and is used in the fuzzy

inference process

19
 FUZZY SYSTEMS

Fuzzy Associative Matrix

Concept behind M

a1  b1 a1  b2 …
a2  b1 …
.
.
.

If a1 is true then b1 is true; and so on

20
 FUZZY SYSTEMS

• Inference process in fuzzy expert systems has four steps

• FUZZIFICATION
• INFERENCE
• COMPOSITION
• DEFUZZIFICATION

21
 FUZZY SYSTEMS

• Fuzzification :
In the fuzzification subprocess, the membership functions defined
on the input variables are applied to their actual values, to
determine the degree of truth for each rule premise

• Inference :
The truth value for the premise of each rule is computed, and
applied to the conclusion part of each rule. This results in one
fuzzy subset to be assigned to each output variable for each rule

22
 FUZZY SYSTEMS

• Composition :
All of the fuzzy subsets assigned to each output variable are
combined together to form a single fuzzy subset for each output
variable.

• Defuzzification :
Sometimes it is useful to just examine the fuzzy subsets that are
the result of the composition process, but more often, this fuzzy
value needs to be converted to a single number - a crisp value.
This is what the defuzzification subprocess does

23
FUZZY SYSTEMS

24
 FUZZY SYSTEMS

Fuzzy Inference: Composition Operator,

Max-min Inference

25
 FUZZY SYSTEMS

Composition of Fuzzy Relations

Let there be three universes U, V and W

Let R be the relation that relates elements from U to V

e.g. R= 0.6 0.8

0.7 0.9

And let S be the relation between V and W

e.g. S= 0.3 0.1

0.2 0.8
26
 FUZZY SYSTEMS

Composition of Fuzzy Relations

With the help of an operation called “composition” we can

find the relation T that maps elements of U to W

By max-min rule T = R  S = maxvV { min(R(u, v), S(v, w)) }

0.6 0.8 0.3 0.1 = 0.3 0.8

0.7 0.9  0.2 0.8 0.3 0.8
Where
element (1,1) is obtained by max{min(0.6, 0.3), min(0.8, 0.2)}
= 0.3
Note that S  R = 0.3 0.3  R  S
0.7 0.8
27
 FUZZY SYSTEMS

Example:

If Temperature is normal then Speed is medium

Let A = [0/100, 0.5/125, 1/150, 0.5/175, 0/200]

B = [0/10, 0.6/20, 1/30, 0.6/40, 0/50]

M = 0 0 0 0 0
0 0.5 0.5 0.5 0
0 0.6 1 0.6 0
0 0.5 0.5 0.5 0
0 0 0 0 0

28
 FUZZY SYSTEMS

Max-Min Inference

Let Temperature = 125 o F = Acurrent

Then it may be thought of as a new fuzzy set defined over the
universe of discourse of variable Temperature
Acurrent = [0/100, 0.5/125, 0/150, 0/175, 0/200]

We can find the relationship (i.e. mapping or FAM) between

Acurrent and A as
A’ = Acurrent x A = 0 x [0 0.5 1 0.5 0]
0.5
0 = [0 0.5 0.5 0.5 0]
0 (if we eliminate all 0 rows)
0 29
 FUZZY SYSTEMS

Max-Min Inference

Now the composition of A’ and M will produce a new

relationship, which we call B’

A’  M = B’

0 0 0 0 0
0 0.5 0.5 0.5 0
[0 0.5 0.5 0.5 0]  0 0.6 1.0 0.6 0 = [0 0.5 0.5 0.5 0]
0 0.5 0.5 0.5 0
0 0 0 0 0

30
 FUZZY SYSTEMS

Multi-Premises Rules

If A  B then C CA’
CB’

CA’ = A’  MAC
CB’ = B’  MBC

C’ = CA’  CB’ = min(CA’ , CB’)

CB’
CA’

31
 FUZZY SYSTEMS

Multi-Premises Rules

If A  B then C
C’ = CA’  CB’
= max(CA’ , CB’)

CB’ CB’
CA’ CA’

32
 FUZZY SYSTEMS

Multiple Fuzzy Rules

In a regular rule-based system, if two rules are

simultaneously satisfied, a conflict resolution policy decides
the precedence

The system proceeds sequentially, with one rule firing at a

time

In fuzzy rule based systems, all rules are executed during

each pass through the system

33
 FUZZY SYSTEMS

Fuzzy Inference: Combining Multiple Rules

34
 EXAMPLE – DINNER

Dinner for two: this is a 2 input, 1 output, 3 rule

system

Rule 1 If service is poor or

food is rancid, then
Input 1 tip is cheap

Service (0-10)
Rule 2 If service is good, Output
then tip is average
Tip (5-25%)
Input 2
Food (0-10) Rule 3 If service is excellent or
food is delicious, then
tip is generous

The inputs are crisp (non- All rules are evaluated The results of the rules The result is a crisp (non-
fuzzy) numbers limited to in parallel using fuzzy are combined and fuzzy) number
a specific range reasoning distilled (de-fuzzyfied)

35
 FUZZY SYSTEMS

Multiple Fuzzy Rules

36
 FUZZY SYSTEMS

Fuzzy Inference: Defuzzification

37
 FUZZY SYSTEMS

Defuzzification

In most applications we need to obtain a crisp value after

inferring a fuzzy set B’

The most popular defuzzification technique used is the

fuzzy centroid method

38
 FUZZY SYSTEMS

Defuzzification
p

Fuzzy Centroid Method y

j 1
j A( y j )
yi  p


j 1
A (yj)

Let B’ = [0/10, 0.3/20, 0.5/30, 0.3/40, 0/50]

then y = 10*0 + 20*0.3 + 30*0.5 + 40*0.3 + 50*0

0 + 0.3 + 0.5 + 0.3 + 0

= 6 + 15 + 12 = 33/1.1 = 30
1.1
39