6/9/2015Intelligent Systems and Soft Computing1 Lecture 5 Fuzzy expert systems: Fuzzy inference...

04/18/23 Intelligent Systems and Soft Computing 1

Lecture 5 Lecture 5 Fuzzy expert systems: Fuzzy expert systems: Fuzzy inferenceFuzzy inference

Mamdani fuzzy inferenceMamdani fuzzy inference

Sugeno fuzzy inferenceSugeno fuzzy inference

Case studyCase study

Summary Summary


The most commonly used fuzzy inference technique The most commonly used fuzzy inference technique is the so-called Mamdani method. In 1975, is the so-called Mamdani method. In 1975, Professor Professor Ebrahim Mamdani Ebrahim Mamdani of London University of London University built one of the first fuzzy systems to control a built one of the first fuzzy systems to control a steam engine and boiler combination. He applied a steam engine and boiler combination. He applied a set of fuzzy rules supplied by experienced human set of fuzzy rules supplied by experienced human operators.operators.

Fuzzy inferenceFuzzy inference


Mamdani fuzzy inferenceMamdani fuzzy inference

The Mamdani-style fuzzy inference process is The Mamdani-style fuzzy inference process is performed in four steps:performed in four steps:

fuzzification of the input variables,fuzzification of the input variables,

rule evaluation;rule evaluation;

aggregation of the rule outputs, and finallyaggregation of the rule outputs, and finally

defuzzification.defuzzification.


We examine a simple two-input one-output problem that We examine a simple two-input one-output problem that includes three rules:includes three rules:

Rule: 1 Rule: 1 Rule: 1 Rule: 1 IFIF x x is is AA3 3 IF IF project_funding project_funding is is adequate adequate OR OR y y is is BB1 1 OR OR project_staffing project_staffing is is small small THEN THEN z z is is CC1 1 THENTHEN risk risk is is lowlow

Rule: 2 Rule: 2 Rule: 2 Rule: 2 IF IF x x is is AA2 2 IF IF project_funding project_funding is is marginal marginal AND AND y y is is BB2 2 AND AND project_staffing project_staffing is is large large THEN THEN z z is is CC2 2 THEN THEN risk risk is is normalnormal

Rule: 3 Rule: 3 Rule: 3 Rule: 3 IF IF x x is is AA1 IF 1 IF project_funding project_funding is is inadequateinadequate THEN THEN z z is is CC3 3 THENTHEN risk risk is is highhigh


Step 1: FuzzificationStep 1: Fuzzification

The first step is to take the crisp inputs, The first step is to take the crisp inputs, xx1 and 1 and yy11 ( (project fundingproject funding and and project project staffingstaffing), and determine), and determine thethe degree degree to which these inputs belong to each of theto which these inputs belong to each of the appropriateappropriate fuzzy sets.fuzzy sets.

CrispInput

0.1

0.71

0y1

B1 B2

Y

CrispInput

0.20.5

1

0

A1 A2 A3

x1

x1 X(x =A1) = 0.5(x =A2) = 0.2

(y =B1) = 0.1(y =B2) = 0.7


Step 2: Rule EvaluationStep 2: Rule Evaluation

The second step is to take the The second step is to take the fuzzified inputs, fuzzified inputs, ((xx==AA1) 1) = 0.5, = 0.5, ((xx==AA2) 2) = 0.2, = 0.2, ((yy==BB1) 1) = 0.1 and = 0.1 and ((yy==BB 2) 2) = =

0.7, and apply them to the antecedents of the fuzzy 0.7, and apply them to the antecedents of the fuzzy rules. If a given fuzzy rule has multiple antecedents, rules. If a given fuzzy rule has multiple antecedents, the fuzzy operator (AND or OR) is used to obtain a the fuzzy operator (AND or OR) is used to obtain a single number that represents the result of the single number that represents the result of the antecedent evaluation. This number (the truth value) antecedent evaluation. This number (the truth value) is then applied to the consequent membership is then applied to the consequent membership function.function.


To evaluate the disjunction of the rule antecedents, To evaluate the disjunction of the rule antecedents, we use the we use the OR fuzzy operationOR fuzzy operation. Typically, fuzzy . Typically, fuzzy expert systems make use of the classical fuzzy expert systems make use of the classical fuzzy operation operation unionunion::

AABB((xx) = ) = max max [[AA((xx), ), BB((xx)])]

Similarly, in order to evaluate the conjunction of the Similarly, in order to evaluate the conjunction of the rule antecedents, we apply the rule antecedents, we apply the AND fuzzy operation AND fuzzy operation intersectionintersection::

AABB((xx) = ) = min min [[AA((xx), ), BB((xx)])]


A31

0 X

1

y10 Y

0.0

x1 0

0.1C1

1

C2

Z

1

0 X

0.2

0

0.2 C11

C2

Z

A2

x1

Rule 3: IF x isA1 (0.5)

A11

0 X 0

1

Zx1

THEN

C1 C2

1

y1

B2

0 Y

0.7

B10.1

C3

C3

C30.5 0.5

OR(max)

AND(min)

OR THENRule 1: IF x isA3 (0.0)

AND THENRule 2: IF x isA2 (0.2)

y isB1 (0.1) z isC1 (0.1)

y isB2 (0.7) z isC2 (0.2)

z isC3 (0.5)

Mamdani-style rule evaluationMamdani-style rule evaluation


Now the result of the antecedent evaluation can be Now the result of the antecedent evaluation can be applied to the membership function of the applied to the membership function of the consequent.consequent.

The most common method of correlating the rule The most common method of correlating the rule consequent with the truth value of the rule consequent with the truth value of the rule antecedent is to cut the consequent membership antecedent is to cut the consequent membership function at the level of the antecedent truth. This function at the level of the antecedent truth. This method is called method is called clippingclipping. Since the top of the . Since the top of the membership function is sliced, the clipped fuzzy set membership function is sliced, the clipped fuzzy set loses some information. However, clipping is still loses some information. However, clipping is still often preferred because it involves less complex and often preferred because it involves less complex and faster mathematics, and generates an aggregated faster mathematics, and generates an aggregated output surface that is easier to defuzzify.output surface that is easier to defuzzify.


While clipping is a frequently used method, While clipping is a frequently used method, scaling scaling offersoffers a better approach for preserving the original a better approach for preserving the original shape of the fuzzy set. The original membership shape of the fuzzy set. The original membership function of the rule consequent is adjusted by function of the rule consequent is adjusted by multiplying all its membership degrees by the truth multiplying all its membership degrees by the truth value of the rule antecedent. This method, which value of the rule antecedent. This method, which generally loses less information, can be very useful in generally loses less information, can be very useful in fuzzy expert systems.fuzzy expert systems.


Clipped and scaled membership functionsClipped and scaled membership functions

1.0

0.0

0.2

Z Z

C2

1.0

0.0

0.2

C2

Degree of Membership



Step 3: Aggregation of the rule outputsStep 3: Aggregation of the rule outputs

Aggregation is the process of unification of the Aggregation is the process of unification of the outputs of all rules. We take the membership outputs of all rules. We take the membership functions of all rule consequents previously clipped or functions of all rule consequents previously clipped or scaled and combine them into a single fuzzy set. scaled and combine them into a single fuzzy set.

The input of the aggregation process is the list ofThe input of the aggregation process is the list of clipped or scaled consequent membership functions, clipped or scaled consequent membership functions, and the output is one fuzzy set for each output and the output is one fuzzy set for each output variable.variable.


00.1

1C1

z isC1 (0.1)

C2

0

0.2

1

z isC2 (0.2)

0

0.5

1

z isC3 (0.5)

ZZZ

0.2

Z0

C30.5

0.1

Aggregation of the rule outputsAggregation of the rule outputs


Step 4: DefuzzificationStep 4: Defuzzification

The last step in the fuzzy inference process is The last step in the fuzzy inference process is defuzzification. Fuzziness helps us to evaluate the defuzzification. Fuzziness helps us to evaluate the rules, but the final output of a fuzzy system has to be rules, but the final output of a fuzzy system has to be a crisp number. The input for the defuzzification a crisp number. The input for the defuzzification process is the aggregate output fuzzy set and the process is the aggregate output fuzzy set and the output is a single number.output is a single number.


There are several defuzzification methods, butThere are several defuzzification methods, but probably the most popular one is the probably the most popular one is the centroidcentroid techniquetechnique. It finds the . It finds the point where a vertical linepoint where a vertical line would slice the aggregate set into two equal masses. would slice the aggregate set into two equal masses. Mathematically this Mathematically this centre of gravity (COG) centre of gravity (COG) cancan be expressed as:be expressed as:

b

aA

b

aA

COG

x x dx

x dx


Centroid defuzzification method finds a pointCentroid defuzzification method finds a point representing the centre of gravity of the fuzzy set, representing the centre of gravity of the fuzzy set, AA, , on the interval, on the interval, abab. .

A reasonable estimate can be obtained by calculatingA reasonable estimate can be obtained by calculating it over a sample of it over a sample of points.points.1.0

0.0

0.2

0.4

0.6

0.8

160 170 180 190 200

a b

210

A

150X


Centre of gravity (COG):Centre of gravity (COG):

4.675.05.05.05.02.02.02.02.01.01.01.0

5.0)100908070(2.0)60504030(1.0)20100(

COG

1.0

0.0

0.2

0.4

0.6

0.8

0 20 30 40 5010 70 80 90 10060

Z

DegreeofMembership

67.4


Mamdani-style inference, as we have just seen, Mamdani-style inference, as we have just seen, requires us to find the centroid of a two-dimensional requires us to find the centroid of a two-dimensional shape by integrating across a continuously varying shape by integrating across a continuously varying function. In general, this process is not function. In general, this process is not computationally efficient.computationally efficient.

Michio Sugeno Michio Sugeno suggested to use a single spike, a suggested to use a single spike, a singletonsingleton, as the membership function of the rule , as the membership function of the rule consequent. A consequent. A singletonsingleton,,, or more precisely a , or more precisely a fuzzy fuzzy singletonsingleton, is a fuzzy set with a membership , is a fuzzy set with a membership function that is unity at a single particular point on function that is unity at a single particular point on the universe of discourse and zero everywhere else.the universe of discourse and zero everywhere else.

Sugeno fuzzy inferenceSugeno fuzzy inference


Sugeno-style fuzzy inference is very similar to the Sugeno-style fuzzy inference is very similar to the Mamdani method. Sugeno changed only a rule Mamdani method. Sugeno changed only a rule consequent. Instead of a fuzzy set, he used a consequent. Instead of a fuzzy set, he used a mathematical function of the input variable. The format mathematical function of the input variable. The format of theof the Sugeno-style fuzzy rule Sugeno-style fuzzy rule isis

IF IF x x is is A A AND AND y y is is B B THEN THEN z z is is f f ((x, yx, y) )

where where xx, , y y and and z z are linguistic variables; are linguistic variables; A A and and B B are are fuzzy sets on universe of discourses fuzzy sets on universe of discourses X X and and YY, , respectively; and respectively; and f f ((x, yx, y) is a mathematical function.) is a mathematical function.


The most commonly used The most commonly used zero-order Sugeno fuzzy zero-order Sugeno fuzzy model model applies fuzzy rules in the following form:applies fuzzy rules in the following form:

IF IF x x is is A A AND AND y y is is B B THEN THEN z z is is kk

where where k k is a constant. is a constant.

In this case, the output of each fuzzy rule is constant. In this case, the output of each fuzzy rule is constant. All consequent membership functions are represented All consequent membership functions are represented by singleton spikes.by singleton spikes.


A3

1

0 X

1

y10 Y

0.0

x1 0

0.1

1

Z

1

0 X

0.2

0

0.2

1

Z

A2

x1

z is k3 (0.5)Rule 3: IF x is A1 (0.5)

A11

0 X 0

1

Zx1

THEN

1

y1

B2

0 Y

0.7

B10.1

0.5 0.5

OR(max)

AND(min)

OR y is B1 (0.1) THEN z is k1 (0.1)Rule 1: IF x is A3 (0.0)

AND y is B2 (0.7) THEN z is k2 (0.2)Rule 2: IF x is A2 (0.2)

k1

k2

k3

Sugeno-style rule evaluationSugeno-style rule evaluation


z isk1(0.1) z isk2(0.2) z isk3(0.5) 0

1

0.1Z 0

0.5

1

Z0

0.2

1

Zk1 k2 k3 0

1

0.1Zk1 k2 k3

0.20.5

Sugeno-style aggregation of the rule outputsSugeno-style aggregation of the rule outputs


655.02.01.0

805.0502.0201.0

)3()2()1(

3)3(2)2(1)1(

kkkkkkkkk

WA

0 Z

CrispOutputz1

z1

Weighted average (WA):Weighted average (WA):

Sugeno-style defuzzificationSugeno-style defuzzification


Mamdani method is widely accepted for capturing Mamdani method is widely accepted for capturing expert knowledge. It allows us to describe the expert knowledge. It allows us to describe the expertise in more intuitive, more human-like expertise in more intuitive, more human-like manner. However, Mamdani-type fuzzy inference manner. However, Mamdani-type fuzzy inference entails a substantial computational burden.entails a substantial computational burden.

On the other hand, Sugeno method is On the other hand, Sugeno method is computationally effective and works well with computationally effective and works well with optimisation and adaptive techniques, which makes optimisation and adaptive techniques, which makes it very attractive in control problems, particularly it very attractive in control problems, particularly for dynamic nonlinear systems.for dynamic nonlinear systems.

How to make a decision on which method How to make a decision on which method to apply – Mamdani or Sugeno?to apply – Mamdani or Sugeno?

More Examples for Mamdani Fuzzy Models

Example #1

Single input single output Mamdani fuzzy model with 3 rules:

If X is small then Y is small R1

If X is medium then Y is medium R2

Is X is large then Y is large R3

X = input [-10, 10] Y = output [0,10]

Using centroid defuzzification, we obtain the following overall input-output curve

04/18/23 Intelligent Systems and Soft Computing25

Single input single output antecedent & consequent MFs

04/18/23 Intelligent Systems and Soft Computing26Overall input-output curve

Example #2 (Mamdani Fuzzy models )

Two input single-output Mamdani fuzzy model with 4 rules:

If X is small & Y is small then Z is

negative large

If X is small & Y is large then Z is

negative small

If X is large & Y is small then Z is

positive small

If X is large & Y is large then Z is

positive large04/18/23 Intelligent Systems and Soft Computing27

Two-input single output antecedent & consequent MFs04/18/23 28Intelligent Systems and Soft Computing

X = [-5, 5]; Y = [-5, 5]; Z = [-5, 5] with max-min composition & centroid defuzzification, we can determine the overall input output surface

Overall input-output surface04/18/23 29Intelligent Systems and Soft Computing

More Examples for Sugeno Fuzzy Models

Example 1: Single output-input Sugeno fuzzy model with three rules

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

If “small”, “medium” & “large” are nonfuzzy sets then the overall input-output curve is a piece wise linear


04/18/23 31Intelligent Systems and Soft Computing


However, if we have smooth membership functions (fuzzy rules) the overall input-output curve becomes a smoother one

Example 2:

Two-input single output fuzzy model with 4 rules

R1: if X is small & Y is small then z = -x +y +1

R2: if X is small & Y is large then z = -y +3

R3: if X is large & Y is small then z = -x +3

R4: if X is large & Y is large then z = x + y + 2



Overall input-output surface


Building a fuzzy expert system: case studyBuilding a fuzzy expert system: case study

A service centre keeps spare parts and repairs failed A service centre keeps spare parts and repairs failed ones.ones.

A customer brings a failed item and receives a spare A customer brings a failed item and receives a spare of the same type.of the same type.

Failed parts are repaired, placed on the shelf, and Failed parts are repaired, placed on the shelf, and thus become spares.thus become spares.

The objective here is to advise a manager of the The objective here is to advise a manager of the service centre on certain decision policies to keep service centre on certain decision policies to keep the customers satisfied.the customers satisfied.


Process of developing a fuzzy expert systemProcess of developing a fuzzy expert system

1. Specify the problem and define linguistic variables.1. Specify the problem and define linguistic variables.

2. Determine fuzzy sets.2. Determine fuzzy sets.

3. Elicit and construct fuzzy rules.3. Elicit and construct fuzzy rules.

4. Encode the fuzzy sets, fuzzy rules and procedures 4. Encode the fuzzy sets, fuzzy rules and procedures to perform fuzzy inference into the expert system.to perform fuzzy inference into the expert system.

5. Evaluate and tune the system.5. Evaluate and tune the system.


There are four main linguistic variables: average There are four main linguistic variables: average waiting time (mean delay) waiting time (mean delay) mm, repair utilisation , repair utilisation factor of the service centre factor of the service centre (is the ratio of the (is the ratio of the customer arrival day to the customer departure rate) customer arrival day to the customer departure rate) number of servers number of servers ss, and initial number of spare parts , and initial number of spare parts nn..

Step Step 11: Specify the problem and define : Specify the problem and define linguistic variables linguistic variables


Linguistic variables and their rangesLinguistic variables and their ranges

Linguistic Va lue Notation Numerical Range (normalised)Very ShortShortMedium

VSSM

[0, 0.3][0.1, 0.5][0.4, 0.7]

Linguistic Va lue Notation

Notation

Numerical Range (normalised)SmallMediumLarge

SML

[0, 0.35][0.30, 0.70]

[0.60, 1]

Linguistic Va lue Numerical RangeLowMediumHigh

LMH

[0, 0.6][0.4, 0.8][0.6, 1]

Linguistic Va lue Notation Numerical Range (normalised)Very SmallSmallRather SmallMediumRather LargeLargeVery Large

VSS

RSMRLL

VL

[0, 0.30][0, 0.40]

[0.25, 0.45][0.30, 0.70][0.55, 0.75]

[0.60, 1][0.70, 1]

Linguistic Variable: Mean Delay, m

Linguistic Variable: Number of Servers, s

Linguistic Variable: Repair Utilisation Factor,

Linguistic Variable: Number of Spares, n


Step Step 22: Determine fuzzy sets: Determine fuzzy sets

Fuzzy sets can have a variety of shapes. However, Fuzzy sets can have a variety of shapes. However, a triangle or a trapezoid can often provide an a triangle or a trapezoid can often provide an adequate representation of the expert knowledge, adequate representation of the expert knowledge, and at the same time, significantly simplifies theand at the same time, significantly simplifies the process of computation.process of computation.


Fuzzy sets of Fuzzy sets of Mean Delay mMean Delay m

0.10

1.0

0.0

0.2

0.4

0.6

0.8

0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1Mean Delay (normalised)

SVS M



Fuzzy sets of Fuzzy sets of Number of Servers sNumber of Servers s

0.10

1.0

0.0

0.2

0.4

0.6

0.8

0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

M LS



Fuzzy sets of Fuzzy sets of Repair Utilisation Factor Repair Utilisation Factor

0.10

1.0

0.0

0.2

0.4

0.6

0.8

0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1Repair Utilisation Factor

M HL

Degree ofMembership


Fuzzy sets of Fuzzy sets of Number of Spares nNumber of Spares n

0.10

1.0

0.0

0.2

0.4

0.6

0.8

0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

S RSVS M RL L VL

Degree ofMembership

Number of Spares (normalised)


Step Step 33: Elicit and construct fuzzy rules: Elicit and construct fuzzy rules

To accomplish this task, we might ask the expert to To accomplish this task, we might ask the expert to describe how the problem can be solved using the describe how the problem can be solved using the fuzzy linguistic variables defined previously. fuzzy linguistic variables defined previously.

Required knowledge also can be collected from Required knowledge also can be collected from other sources such as books, computer databases, other sources such as books, computer databases, flow diagrams and observed human behavior.flow diagrams and observed human behavior.

The matrix form of representing fuzzy rules is called The matrix form of representing fuzzy rules is called fuzzy associative memory (FAM).fuzzy associative memory (FAM).


m

s

M

RL

VL

S

RS

L

VS

S

M

MVS S

L

M

S

The square FAM representationThe square FAM representation


The rule tableThe rule table

Rule m s n Rule m s n Rule m s n

1 VS S L VS 10 VS S M S 19 VS S H VL

2 S S L VS 11 S S M VS 20 S S

S

3 M S L VS 12 M S M VS 21 M S

4 VS M L VS 13 VS M M RS 22 VS M H M

M

M

M

5 S M L VS 14 S M M S 23 S M

6 M M L VS 15 M M M VS 24 M M

7 VS L L S 16 VS L M M 25 VS L H

H

H

H

H

H

RL

8 S L

L

L S 17 S L M RS 26 S L

9 M L L VS 18 M L M S 27 M L H RS


Rule Base 1Rule Base 11. If (utilisation_factor is L) then (number_of_spares is S)2. If (utilisation_factor is M) then (number_of_spares is M)3. If (utilisation_factor is H) then (number_of_spares is L)

4. If (mean_delay is VS) and (number_of_servers is S) then (number_of_spares is VL)5. If (mean_delay is S) and (number_of_servers is S) then (number_of_spares is L)6. If (mean_delay is M) and (number_of_servers is S) then (number_of_spares is M)

7. If (mean_delay is VS) and (number_of_serversis M) then (number_of_spares is RL)8. If (mean_delay is S) and (number_of_servers is M) then (number_of_spares is RS)9. If (mean_delay is M) and (number_of_servers is M) then (number_of_spares is S)

10. If (mean_delay is VS) and (number_of_servers is L) then (number_of_spares is M)11. If (mean_delay is S) and (number_of_servers is L) then (number_of_spares is S)12. If (mean_delay is M) and (number_of_servers is L) then (number_of_spares is VS)


VS VS VSVS VS VS

VS VS VS

VL L MHS

VS VS VSVS VS VS

VS VS VSM

VS VS VSVS VS VSS S VSL

s

LVS S

m

MH

VS VS VSLVS S

S

m

VS VS VSM

S S VSL

s

S VS VSMVS S

m

VS S

m

S

RS S VSM

M RS SL

s

S

M M SM

RL M RSL

s

M

M

M

M

Cube FAM of Rule Base 2Cube FAM of Rule Base 2


To accomplish this task, we may choose one of To accomplish this task, we may choose one of two options: to build our system using a two options: to build our system using a programming language such as C/C++ or Pascal, programming language such as C/C++ or Pascal, or to apply a fuzzy logic development tool such as or to apply a fuzzy logic development tool such as MATLAB Fuzzy Logic Toolbox, Fuzzy Clips, or Fuzzy MATLAB Fuzzy Logic Toolbox, Fuzzy Clips, or Fuzzy Knowledge Builder. Knowledge Builder.

Step Step 44: Encode the fuzzy sets, fuzzy rules : Encode the fuzzy sets, fuzzy rules and procedures to and procedures to perform fuzzy perform fuzzy inference into the expert systeminference into the expert system


Step Step 55: Evaluate and tune the system: Evaluate and tune the system

The last, and the most laborious, task is to evaluate The last, and the most laborious, task is to evaluate and tune the system. We want to see whether our and tune the system. We want to see whether our fuzzy system meets the requirements specified at fuzzy system meets the requirements specified at the beginning.the beginning.

Several test situations depend on the mean delay, Several test situations depend on the mean delay, number of servers and repair utilization factor.number of servers and repair utilization factor.

The Fuzzy Logic Toolbox can generate surface to The Fuzzy Logic Toolbox can generate surface to help us analyze the system’s performance.help us analyze the system’s performance.


00.2

0.40.6

0.81

0

0.2

0.4

0.6

0.2

0.3

0.4

0.5

0.6

number_of_serversmean_delay

Three-dimensional plots for Rule Base 1Three-dimensional plots for Rule Base 1



00.2

0.40.6

0.81

0

0.2

0.4

0.6

0.2

0.3

0.4

0.5

0.6

utilisation_factormean_delay


00.2

0.40.6

0.81

0

0.2

0.4

0.6

0.15

0.2

0.25

0.3

0.35




00.2

0.40.6

0.81

0

0.2

0.4

0.6

0.2

0.3

0.4


0.5



However, even now, the expert might not be However, even now, the expert might not be satisfied with the system performance.satisfied with the system performance.

To improve the system performance, we may use To improve the system performance, we may use additional sets - additional sets - Rather Small Rather Small and and Rather Large Rather Large – – on the universe of discourse on the universe of discourse Number of ServersNumber of Servers, , and then extend the rule base.and then extend the rule base.


Modified fuzzy sets of Modified fuzzy sets of Number of Servers sNumber of Servers s

0.10

1.0

0.0

0.2

0.4

0.6

0.8

0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1Number of Servers (normalised)

RS M RL LS

Degree ofMembership


VS VS VS

VS VS VS

VS VS VS

VS VS VS

VS VS VS

VS VS VS

VS VS VS

VS VS VS

VS VS VS

VS VS VS

VS VS VS

VS VS VS

VS VS VS

S S VS

S S VS

VL L M

VL RL RS

M M S

RL M RS

L M RS

HS

M

RL

L

RS

s

LVS S

m

M

M

H

VS VS VS

VS VS VS

VS VS VS

S S VS

S S VS

LVS S

S

M

M

RL

L

RS

m

s

S VS VS

S VS VS

RS S VS

M RS S

M RS S

MMVS S

m

VS S

m

S

M

RL

L

RS

s

S

M

M

RL

L

RS

s

Cube FAM of Rule Base 3Cube FAM of Rule Base 3


00.2

0.40.6

0.81

0

0.2

0.4

0.6

0.15

0.2

0.25

0.3

0.35




00.2

0.40.6

0.81

0

0.2

0.4

0.6

0.2

0.3

0.4


0.5



Tuning fuzzy systemsTuning fuzzy systems

1. Review model input and output variables, and if 1. Review model input and output variables, and if required redefine their ranges.required redefine their ranges.

2. Review the fuzzy sets, and if required define 2. Review the fuzzy sets, and if required define additional sets on the universe of discourse. The additional sets on the universe of discourse. The use of wide fuzzy sets may cause the fuzzy system to use of wide fuzzy sets may cause the fuzzy system to perform roughly.perform roughly.

3. Provide sufficient overlap between neighboring sets. 3. Provide sufficient overlap between neighboring sets. It is suggested that triangle-to-triangle and trapezoid-It is suggested that triangle-to-triangle and trapezoid-to-triangle fuzzy sets should overlap between 25% to to-triangle fuzzy sets should overlap between 25% to 50% of their bases.50% of their bases.


4. Review the existing rules, and if required add new 4. Review the existing rules, and if required add new rules to the rule base.rules to the rule base.

5. Examine the rule base for opportunities to write 5. Examine the rule base for opportunities to write hedge rules to capture the pathological behaviour hedge rules to capture the pathological behaviour of the system. of the system.

6. Adjust the rule execution weights. Most fuzzy 6. Adjust the rule execution weights. Most fuzzy logic tools allow control of the importance of rules logic tools allow control of the importance of rules by changing a weight multiplier. by changing a weight multiplier.

7. Revise shapes of the fuzzy sets. In most cases, 7. Revise shapes of the fuzzy sets. In most cases, fuzzy systems are highly tolerant of a shape fuzzy systems are highly tolerant of a shape approximation.approximation.

Date post:	19-Dec-2015
Category:	Documents
Upload:	howard-eaton
View:	228 times
Download:	1 times

6/9/2015Intelligent Systems and Soft Computing1 Lecture 5 Fuzzy expert systems: Fuzzy inference...

Documents