Fuzzy Inference Systems

Components of a
Fuzzy Inference System

Four main parts

1. Fuzzification Module
2. Knowledge Base
3. Inference Engine
4. Defuzzification Module
Fuzzy Inference System
Constructing a Fuzzy Inference system
Steps
1. Specify the problem and define linguistic variables.
2. Identify the membership functions for the linguistic
variables and construct the fuzzy sets for them.
3. Convert crisp data into fuzzy data sets using membership
functions. (fuzzification)
4. Construct knowledge base of fuzzy rules.
5. Evaluate rules in the rule base by applying implication
function to the rule. (Inference Engine)
6. Combine results from each rule. (Inference Engine)
7. Convert output data into non-fuzzy values.
(defuzzification)
8. Evaluate and tune the system.
 Problem
• Create a fuzzy control system which models
how you might choose to tip at a restaurant.
• When tipping, you consider the service and
food quality, rated between 0 and 10.
• You use this to leave a tip.
 Inputs (2) and Output(1)
• service
– Universe (crisp value range): How good was the
service of the wait staff, on a scale of 0 to 10?
– Fuzzy set (fuzzy value range): poor, acceptable,
amazing
• food quality
– Universe (crisp value range): How tasty was the food,
on a scale of 0 to 10?
– Fuzzy set (fuzzy value range): bad, decent, delicious

• tip
– Universe: How much should we tip
– Fuzzy set: low, medium, high
Construct membership functions

Gaussian Membership function

Construct membership functions

Trapezoidal Membership function

Construct membership functions

Triangular Membership function

Construct knowledge base of rules
• IF the service is excellent or the food is
delicious, THEN the tip will be high.
• IF the service is good, THEN the tip will be
medium.
• IF the service is poor or the food is awful
THEN the tip will be low.
 Rule Table
Rule No Service Food Tip

1 Excellent Delicious High

2 Excellent Moderate Medium

3 Good Delicious High

4 Poor Awful Low

Rule Antecedent
Rule Consequent (Implication)

Implication operators: min,

product
 Obtain Aggregate fuzzy value
• Combine the results of evaluation by all the
rules to form a final result
• Some of the most commonly used aggregation
operators are maximum, sum and probabilistic
sum
• This result is a fuzzy value
 Perform defuzzification
• Defuzzification converts the fuzzy output set to a
crisp number.
• Commonly used techniques: centroid, mean of
maximum methods.
Percent of bill to add as tip
 Tuning fuzzy systems
• Review model input and output variables, and if
required redefine their ranges.
• Provide sufficient overlap between neighboring sets.
(triangle-to-triangle and trapezoid-to-triangle fuzzy sets should overlap
between 25 and 50 per cent of their bases)
• Review the existing rules, and if required add new
rules to the rule base.
• Revise shapes of the fuzzy sets.
Types of Fuzzy Inference Systems
• Mamdani FIS
• Sugeno FIS
 Sugeno Style Inference
• The format of the Sugeno-style fuzzy rule is
IF x is A AND y is B THEN z is f (x, y) where A and B
are fuzzy sets on universe of discourses X and Y,
respectively; and f (x, y) is a mathematical function
– R1: if X is small and Y is small then z= -x+y+1
– R2: if X is small and Y is large then z= -y+3
• Sugeno Style Inference is computationally efficient.
• Works well with optimization and adaptive
techniques.
• Well-suited for mathematical analysis.
 Applications of Fuzzy Systems
• Fuzzy expert systems
1. Pattern recognition
2. Financial systems
3. Data analysis
• Fuzzy control systems
1. Fuzzy car
2. Fuzzy washing machine
3. Fuzzy vacuum cleaner
4. Fuzzy air conditioner
5. Fuzzy camcorder
6. Fuzzy Automatic Train Operation systems
7. Fuzzy automatic container crane operations