19RE613

Soft Computing
202 (3)
Dr. Lekshmi R. R.
Asst. Prof.
Department of Electrical & Electronics
Engineering
Amrita School of Engineering 1
Fuzzy Membership
Function
Fuzzy Membership Function
Set of heights in the region around 6′ − 7′ Tall

1
𝜇𝐴(𝑥)

1 2 3 4 5 6 7 8
𝑥

Core

෩
Region of universe characterized by full membership in set A
Fuzzy Membership Function
Set of heights in the region around 6′ − 7′ Tall

1
𝜇𝐴(𝑥)

1 2 3 4 5 6 7 8
𝑥

Boundaries

Region of universe characterized by non zero membership

but not complete membership in set A෩
Fuzzy Membership Function
Set of heights in the region around 6′ − 7′ Tall

1
𝜇𝐴(𝑥)

1 2 3 4 5 6 7 8
𝑥

Support

Region of universe characterized by non zero membership in

෩
set A
Fuzzy Membership Function
Normal Fuzzy Set

1 1

𝜇𝐴(𝑥) 𝜇𝐴(𝑥)

1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8
𝑥 𝑥

Membership function has at least one element in universe whose membership

value is unity
Fuzzy Membership Function
Convex Fuzzy Set

Membership function whose

1 • membership values are
𝜇𝐴(𝑥) monotonically increasing
• membership values are
monotonically decreasing
• membership values are
1 2 3 4 5 6 7 8 monotonically increasing then
𝑥
monotonically decreasing
Fuzzy Membership Function
Cross Over Point

෩
A ෩
B
1
𝜇𝐴(𝑥)
0.5

1 2 3 4 5 6 7 8
𝑥

Elements in universe for which a particular fuzzy set has values equal to 0.5.
Types of Membership Functions
Gaussian Membership Function

෩
A
1

𝜇𝐴(𝑥)

1 2 3 4 5 6 7 8
𝑥
Types of Membership Functions
Triangular Membership Function

෩
A
1

𝜇𝐴(𝑥)

1 2 3 4 5 6 7 8
𝑥
Types of Membership Functions
Delta Membership Function

෩
A
1

𝜇𝐴(𝑥)

1 2 3 4 5 6 7 8
𝑥
 IF mark is HIGH, THEN ELECT

Mark Elect/Not

LOW AVERAGE HIGH DON ′ T ELECT ELECT

 IF mark is HIGH, THEN ELECT

LOW AVERAGE HIGH DON ′ T ELECT ELECT

 IF mark is HIGH, THEN ELECT

LOW AVERAGE HIGH DON ′ T ELECT ELECT

 IF 𝑥1 is 𝐴෪ ෪ ෪
11 𝑎𝑛𝑑 𝑥2 is 𝐴21 , THEN y is 𝐵1

𝑥1 𝑥2 𝑦

෪1 1
𝐴 𝐴෪
12 𝐴෪
21 𝐴෪
22
෪1
𝐵 ෪2
𝐵
 IF 𝑥1 is 𝐴෪ ෪ ෪
11 𝑎𝑛𝑑 𝑥2 is 𝐴21 , THEN y is 𝐵1

𝑥1 𝑥2 𝑦

෪1 1
𝐴 𝐴෪
12 𝐴෪
21 𝐴෪
22
෪1
𝐵 ෪2
𝐵
 IF 𝑥1 is 𝐴෪ ෪ ෪
11 𝑎𝑛𝑑 𝑥2 is 𝐴21 , THEN y is 𝐵1

𝑥1 𝑥2 𝑦

෪1 1
𝐴 𝐴෪
12 𝐴෪
21 𝐴෪
22
෪1
𝐵 ෪2
𝐵
 IF 𝑥1 is 𝐴෪ ෪ ෪
11 𝑎𝑛𝑑 𝑥2 is 𝐴21 , THEN y is 𝐵1

𝑥1 𝑥2 𝑦

෪1 1
𝐴 𝐴෪
12 𝐴෪
21 𝐴෪
22
෪1
𝐵 ෪2
𝐵
 Fuzzification
Converting crisp quantity to Fuzzy

Low Voltage
1

𝜇𝐴(𝑥)

0.4

1 2 3 4 5 6 7 8
Voltage
IF 𝑥1 is 𝐴෪ ෪ ෪
11 𝑎𝑛𝑑 𝑥2 is 𝐴21 , THEN y is 𝐵1

OR
IF 𝑥1 is 𝐴෪ ෪ ෪
12 𝑎𝑛𝑑 𝑥2 is 𝐴22 , THEN y is 𝐵2
 𝑥1 𝑥2 Mamdani 𝑦
෪1 1
𝐴 𝐴෪
12 𝐴෪
21 𝐴෪
22
෪1
𝐵 ෪2
𝐵
1

0
෪1 1
𝐴 𝐴෪
12 𝐴෪
21 𝐴෪
22
෪1
𝐵 ෪2
𝐵
1

0
 𝑥1 𝑥2 Mamdani 𝑦
෪1 1
𝐴 𝐴෪
12 𝐴෪
21 𝐴෪
22
෪1
𝐵 ෪2
𝐵
1

0
෪1 1
𝐴 𝐴෪
12 𝐴෪
21 𝐴෪
22
෪1
𝐵 ෪2
𝐵
1

0
 𝑥1 𝑥2 Mamdani 𝑦
෪1 1
𝐴 𝐴෪
12 𝐴෪
21 𝐴෪
22
෪1
𝐵 ෪2
𝐵
1

0
෪1 1
𝐴 𝐴෪
12 𝐴෪
21 𝐴෪
22
෪1
𝐵 ෪2
𝐵
1

0
 𝑥1 𝑥2 Mamdani 𝑦
෪1 1
𝐴 𝐴෪
12 𝐴෪
21 𝐴෪
22
෪1
𝐵 ෪2
𝐵
1 0.8

0 Min
0
෪1 1
𝐴 𝐴෪
12 𝐴෪
21 𝐴෪
22
෪1
𝐵 ෪2
𝐵
1
0.8 Implication

0.2 Min

0
෪1
𝐵 ෪2
𝐵

𝑥1 𝑥2 Aggregation Max

Fuzzification
Max Min Defuzzification
 𝑥1 𝑥2 Mamdani 𝑦
෪1 1
𝐴 𝐴෪
12 𝐴෪
21 𝐴෪
22
෪1
𝐵 ෪2
𝐵
1
0.7 0.5

0
෪1 1
𝐴 𝐴෪
12 𝐴෪
21 𝐴෪
22
෪1
𝐵 ෪2
𝐵
1
0.5
0.3

0
෪1
𝐵 ෪2
𝐵

𝑥1 𝑥2
 𝑥1 𝑥2 Sugeno 𝑦
෪1 1
𝐴 𝐴෪
12 𝐴෪
21 𝐴෪
22
෪1
𝐵 ෪2
𝐵
1 0.8

0
0
0 0.5
1
0.8

0.2

0
0 0.5

𝑥1 𝑥2

0 0.5
Thank you