You are on page 1of 35

FUZZY INFERENCE SYSTEM

It is a popular computing framework based on fuzzy set theory, fuzzy if then rules , and fuzzy reasoning.

3 components:
1. Rule Base: a selection of fuzzy rules 2. Database (or Dictionary): defines the MFs 3. Inference Engine: a reasoning mechanism which performs the inference procedure
11/1/2012 1

Cont..
It can take either fuzzy inputs or crisp inputs but the output it produces are always fuzzy sets.

11/1/2012

MODELS OF FUZZY INFERENCE SYSTEM


Three types
Mamdani fuzzy model Sugeno fuzzy model Larsan fuzzy model

They differ in the consequences of their fuzzy rules and defuzzification methods

11/1/2012

MAMDANI FUZZY MODEL


Proposedd to control a steam engine and boiler combination Mamdani rule base can model the system using rules that have a high correctness.

Correctness-measure of how close our model is to the real one


11/1/2012 4

Cont..
Mamdani model is a crisp model of a system.
It can model a real system where the relation between the inputs and outputs are known. 2 common operators that we use are T-norm and T-conorm operators.
11/1/2012 5

MAMDANI RULE BASE


Can be broken down into 4 part
Fuzzification
Determining the output of each rule given its fuzzy antecedent. Aggregation defuzzification
11/1/2012 6

Fuzzification
Mamdani rule base models a crisp system,it has crisp inputs and outputs. Fuzzifier converts the crisp input into fuzzy variables.

The membership of each fuzzy input variable is evaluated for a given crisp input.
11/1/2012 7

EVALUATING THE RULES


Rules are evaluated based on the membership values. The rule base let the user to determine which type of operation to use like minimum , maximum, product If we use min for T-norm and T-conorm (implication )operators respectively and use max min for composition then the resulting fuzzy reasoning is as,
11/1/2012 8

max-min T-conorm/norm

11/1/2012

AGGREGATING THE RULES


The output of the rule base should be the maximum of the output of each rule. Can use any one of the operators defined on fuzzy sets like maximum , algebraic sum or min.

11/1/2012

10

DEFUZZIFICATION
a method to extract a representative crisp value from a fuzzy set. 5 methods
1. Centroid of area zCOA :

zCOA =Z A(z) z dz / Z A(z) dz

Most widely used strategy.

11/1/2012

11

Cont.

11/1/2012

12

Cont..
2.Bisector of area zBOA : It generates the value z0 which partitions the area into two area with same area. zBOA (z) dz = A zBOA A(z) dz 3.Mean of maximum zMOM :

11/1/2012

13

BOA

11/1/2012

14

CONT
Is the average of the maximizing z at which the MF reaches a maximum .
If A(z) has a single maximum at z=zthen ZMOM=z If A(z) reaches its maximum whenever z[zleft,zright] then ZMOM=(ZLEFT+ZRIGHT)/2.

11/1/2012

15

CONT.
4.SMALLEST OF MAXIMUM ZSOM
It is the minimum of Z of the maximizing .

5.LARGEST OF MAXIMUM ZSOM


ZSOM is the maximum z of the maximizing z ZSOM and LOM are not used as often as the other 3 defuzzification
methods.

11/1/2012

16

Example: Mamdanis Fuzzy Model


Single-input single-output Mamdani fuzzy model
If X is small then Y is small. If X is medium then Y is medium. If X is large then Y is large.
11/1/2012 17

11/1/2012

18

11/1/2012

19

ADVANTAGES AND DISADVANTAGES


Advantages
Its well suited to human input Rules are having high correctness

Disadvantage
Defuzzification requires a large amount of mathematical calculations.

11/1/2012

20

II.LARSEN MODEL
Product operator for a fuzzy implication Max-product operator for the composition

11/1/2012

21

Contd..
The output of Larsen model are also fuzzy sets. Need defuzzification methods to get final output.

11/1/2012

22

II.SUGENO FUZZY MODELS


Also known as the TSK fuzzy model
Introduced in 1984 by
T.Takagi M.Sugeno and K.T.Kang

11/1/2012

23

Motivation of TSK
To reduce the number of rules required by the Mamdani model.
For complex and high dimensional problems Develop a systematic approach to generate fuzzy rules from a given input-output data set. TSK model replaces the fuzzy set (then part) of mamdani rule with function(equation) of the input variables.

11/1/2012

24

TSK fuzzy rule


If x is A and y is B then z=f(x,y)
Where A and B are fuzzy sets in the antecedent ,and Z=f(x,y) is a crisp function in the consequences .
Usually f(x,y) is a polynomial in the input variables x and y.

11/1/2012

25

First order TSK fuzzy model


F(x,y) is a first order polynomial Example: a two-input one-output TSK
If x is Aj and y is Bk then z= px+qy+r The degree the input matches ith rule is typically computed using min operator: wi=min(Aj(x), Bk(y))

11/1/2012

26

Cont..
Each rule has a crisp output Overall output is obtained via weighted average (reduce computation time of defuzzification required in a Mamdani model) Z=i wi zi/ i wi to further reduce computation ,weighted sum may be used. i.e Z= i wi zi
11/1/2012 27

TSK fuzzy model

11/1/2012

28

example
Example: a single input TSK fuzzy model can be expressed as,

If X is small then Y = 0.1X + 6.4 If X is medium then Y = - 0.5X + 4 If X is large then Y = X 2.

11/1/2012

29

11/1/2012

30

Example-2
Two-input single-output Sugeno fuzzy model
If X is small and Y is small If X is small and Y is large If X is large and Y is small If X is large and Y is large then then then then z=-x+y+1. z=-y+3. z=-x+3. z=x+y+2

11/1/2012

31

Zero order TSK


When f is a constant , we have a zero order TSK fuzzy model.

It has minimum computation time.

11/1/2012

32

summary
Overall output via either weighted average or weighted sum is always crisp. Without the time consuming defuzzification operation the TSK fuzzy model is by the far most popular candidate for sample-databased fuzzy modeling.

11/1/2012

33

Advantages
Computationally efficient Works well with linear techniques Has continuity of the output surface Well suited to mathematical analysis.

11/1/2012

34

Reference
Neuro-fuzzy and soft computing J.Jang,C.Sun and E.Mizutani, Prentice Hall 1997 System modelling using a Mamdani Rule BaseBryan Davis,University of Florida Fuzzy systems toolbox,M.Beale and H.Demuth.

11/1/2012

35

You might also like