You are on page 1of 46

Intelligent Control

1
Lecturer: VU VAN PHONG

Dr. Van-Phong Vu-Department of Automatic Control


2

 Instructor: Dr. Vu Van Phong


 Email: phongvv@hcmute.edu.vn
 Websites: https://phongvuac.wixsite.com/vuphong/

Dr. Van-Phong Vu-Department of Automatic Control


3

Chapter 2: Fuzzy sets

Dr. Van-Phong Vu-Department of Automatic Control


4

 Contents
 Definition of fuzzy set
 Basic Operations on Fuzzy sets

Dr. Van-Phong Vu-Department of Automatic Control


5

2.1 Definition of Fuzzy Sets

Dr. Van-Phong Vu-Department of Automatic Control


6 2.1 Definition of Fuzzy Sets
 Consider two Sets as follows
Fuzzy boundary
b
Clear boundary

c
a
a
b
Set A
Set B
Point (a): belongs to set A Point (a): belongs to set B
Point (b): does not belong to set A Point (b): does not belong to set B
Point (c): is on the boundary of set B

Dr. Van-Phong Vu-Department of Automatic Control


7 2.1 Definition of Fuzzy Sets
 Definition of Universe Set (tập cơ sở):
 a collection of objects all having the same
characteristics.
 For example: temperature, pressure, speed of motor,
and so on.

Dr. Van-Phong Vu-Department of Automatic Control


8 2.1 Definition of Fuzzy Sets

 Definition of Fuzzy Sets


 Fuzzy set A on the universe set X is a set in which each
elements of fuzzy set is a pair of value, where and is a
mapping:

 is membership function of fuzzy set A

Dr. Van-Phong Vu-Department of Automatic Control


9 2.1 Definition of Fuzzy Sets
 Continuous fuzzy set

Dr. Van-Phong Vu-Department of Automatic Control


10 2.1 Definition of Fuzzy Sets
 Discrete time fuzzy set

 Universe set:

 A+ + +…}
Dr. Van-Phong Vu-Department of Automatic Control
11 2.1 Definition of Fuzzy Sets
 Fuzzy Sets should satisfy objectivity and reasonable

Dr. Van-Phong Vu-Department of Automatic Control


12 2.1 Definition of Fuzzy Sets
 Membership function:
 The membership function maps each element of X to a
continuous membership value between 0 and 1.
 Typical Membership function:
 Triangle MFs

Dr. Van-Phong Vu-Department of Automatic Control


13
 Trapezoidal MFs

 Gaussian MFs

Dr. Van-Phong Vu-Department of Automatic Control


14  Generalised bell MFs

 Signmoid MFs

Dr. Van-Phong Vu-Department of Automatic Control


15  Definition of Height of fuzzy set:

A(x)
1 High(A(x))

 Support of fuzzy set Support of A(x)

Dr. Van-Phong Vu-Department of Automatic Control


16  ⍺-cut and strong ⍺-cut
 ⍺-cut is a crisp set and not a fuzzy set.
 ⍺-cut 
A: 
A   x A( x)   
 Strong ⍺-cut: 
A   x A( x) (we
 remove two terminal point of
⍺-cut.
 Consider fuzzy set A(x) and
1


Strong ⍺-cut

Dr. Van-Phong Vu-Department of Automatic Control


⍺-cut
 Two fuzzy sets A, B with universal set X ,   [0, 1]
17

A  B   A  B ; A  B  
A  B ;
A  B   A  B ; A  B  
A  B .

 A

‧  ‧
A
‧ 

B

Dr. Van-Phong Vu-Department of Automatic Control


18  Normal Fuzzy sets:
 A fuzzy set A(x) on universe X is called normal fuzzy set if only if
A(x) is a convex set and height of A(x) is equal to 1.

Dr. Van-Phong Vu-Department of Automatic Control


19

2.2 Operation on Fuzzy sets

Dr. Van-Phong Vu-Department of Automatic Control


20 2.2 Operation on Fuzzy sets
 2.2.1 Compliment:
 A fuzzy set A is defined in the universal set X . A(x) represents the
degree of how much does an element , belong to A .
 Ac is usedx toX
represent the complement of fuzzy set A . Ac (x) is
regarded as “the degree of x belongs to Ac “ or “the degree of x does
not belong to A”.
 To simplify the operation, the condition is
necessary. 0  A c ( x)  1

𝑐 ¿ ¿
Dr. Van-Phong Vu-Department of Automatic Control
21
 Continuous time:

Ac(x) Bc(x)
A(x) B(x)
1 1

0 0

Dr. Van-Phong Vu-Department of Automatic Control


22  Discrete time:
 Consider the discrete time fuzzy set A(x) on the Universe set X={1, 2, 3, 4, 5}
0 0.5 1 0.5 0
𝐴 ( 𝑥 ) ={ +
1
+ +
2 3
+
4 5 }
+ }
𝑐 1 0.5 0 0.5 1
𝐴 ( 𝑥 ) ={ + + +
1 2 3 4 5

 Consider the discrete time fuzzy set B(y) on the Universe set Y={y1, y2, y3, y4, y5 , y6, y7}

𝐵 ( 𝑥 )={
0 0. 25 0.75 1 0 .75 0.25 0
𝑦1
+
𝑦2
+ +
𝑦 3 𝑦4
+
𝑦5
+ +
𝑦6 𝑦7 }
Dr. Van-Phong Vu-Department of Automatic Control
23  2.2.2 Intersection (t-norm)
 Consider two fuzzy set and B on the same universe X
 Standard Intersection :

𝑡 ( 𝜇 𝐴(𝑥), 𝜇 𝐵 ( 𝑥 ) )=min ( 𝜇 𝐴(𝑥), 𝜇 𝐵 ( 𝑥 ) )

 Algebraic Product :
𝑡 ( 𝜇 𝐴 ( 𝑥 ) ,𝜇 𝐵 (𝑥 ) )=𝜇 𝐴 (𝑥) 𝜇 𝐵 ( 𝑥 )

 Bounded Difference
𝑡 ( 𝜇 𝐴 ( 𝑥 ) ,𝜇 𝐵 (𝑥 ) )=max ⁡(0 , 𝜇 ¿ ¿ 𝐴 ( 𝑥 ) +𝜇 𝐵 ( 𝑥 ) −1)¿

Dr. Van-Phong Vu-Department of Automatic Control


 Example:
24
 Continuous:
B(x) B(x)
A(x) A(x)
1 1

0 0

Standard intersection Algebraic Product

Dr. Van-Phong Vu-Department of Automatic Control


25  Discrete time

𝐴 ( 𝑦 )={
0 .1 0.2 0. 8 1 0 .7 0.2 0
𝑦1
+ + + + +
𝑦2 𝑦3 𝑦 4 𝑦5 𝑦6 𝑦7
+
}
0 0. 25 0.75 1 0 .75 0.25 0
𝐵 ( 𝑦 )={ +
𝑦1 𝑦2
+ +
𝑦3 𝑦4
+
𝑦5
+ +
𝑦6 𝑦 7 }
Standard intersection:

𝐴 ( 𝑦 ) ⋂ 𝐵 ( 𝑦 ) ={
0 0 .2 0. 75 1 0 .7 0.2 0
+
𝑦1 𝑦 2
+
𝑦3
+ + + +
𝑦4 𝑦5 𝑦 6 𝑦7 }
Product:

𝐴 ( 𝑦 ) ⋂ 𝐵 ( 𝑦 ) ={
0 0 .05 0. 6 1 0 .525 0.05 0
𝑦1
+
𝑦2
+ +
𝑦3 𝑦 4
+
𝑦5
+
𝑦6
+
𝑦7 }
Dr. Van-Phong Vu-Department of Automatic Control
26  Consider two fuzzy set A(x) on universe set X={1, 2, 3, 4, 5} and
B(y) on the universe set Y={a1, a2, a3}

𝐴 ( 𝑥 ) ={
0 .1 0 .2 1 0.5 0
1
+
2
+ +
3 4
+
5 }
𝐵 ( 𝑦 )={
0 .2 1 0.1
+ +
𝑎1 𝑎2 𝑎 3 }
 Find -Standard intersection

++

Dr. Van-Phong Vu-Department of Automatic Control


27  2.2.3 Union(S- norm):
 Consider two fuzzy set and B on the same universe X
 Standard Union :

 Algebraic Sum :

 Bounded Sum :

Dr. Van-Phong Vu-Department of Automatic Control


28

B(x) B(x)
A(x) A(x)
1 1

0 0

Dr. Van-Phong Vu-Department of Automatic Control


 Discrete time:
29
𝐴 ( 𝑦 )={
0 .1 0.2 0. 8 1 0 .7 0.2 0
𝑦1
+ + + + +
𝑦2 𝑦3 𝑦 4 𝑦5 𝑦6 𝑦7
+
}
𝐵 ( 𝑦 )={
0 0. 25 0.75 1 0 .75 0.25 0
𝑦1
+
𝑦2
+ +
𝑦3 𝑦4
+
𝑦5
+ +
𝑦6 𝑦 7 }
 Standard union:

𝐴 ( 𝑦 ) ∪ 𝐵 ( 𝑦 )={
0.1 0 .25 0. 8 1 0 .75 0.25 0
𝑦1
+
𝑦2
+ + +
𝑦3 𝑦4 𝑦5
+ +
𝑦 6 𝑦7 }
 Find the union based on the Algebraic Sum and Bounded Sum

Dr. Van-Phong Vu-Department of Automatic Control


30

 Consider two fuzzy set A(x) on universe set X={1, 2, 3, 4, 5} and B(y) on the
universe set Y={a1, a2, a3}

𝐴 ( 𝑥 ) ={
0 .1 0 .2 1 0.5 0
1
+
2
+ +
3 4
+
5 }
𝐵 ( 𝑦 )={
0 .2 1 0.1
+ +
𝑎1 𝑎2 𝑎 3 }
 Find 𝐴 ( 𝑥 ) ∪ 𝐵 ( 𝑦 ) based on standard union

Dr. Van-Phong Vu-Department of Automatic Control


31

 DeMorgan’s laws:
 For classical set:

A  B  A  B

 A B  A  B

 For fuzzy set:


Dr. Van-Phong Vu-Department of Automatic Control


B(x)
A(x)
32  Find 𝐴( 𝑥 )∩ 𝐵 ( 𝑥 ) 1

0
 Find 𝐴( 𝑥 )∪ 𝐵 ( 𝑥 )

B(x)
1 A(x)

Dr. Van-Phong Vu-Department of Automatic Control


33  Consider two discrete time fuzzy sets on universe set Y={y1, y2, y3, y4,
y5, y6, y7}:

𝐴 ( 𝑦 )={
0 .4 0 .6 0. 8 1 0 .7 0.2 0
+ + + + +
𝑦1 𝑦 2 𝑦3 𝑦4 𝑦5 𝑦6 𝑦7
+
}
𝐵 ( 𝑦 )={
0.3 0. 5 0. 9 1 0 .8 0.1 0
+ + + + +
𝑦1 𝑦2 𝑦 3 𝑦4 𝑦5 𝑦 6 𝑦7
+
}
 Find 𝐴 ( 𝑦 ) ∩ 𝐵 ( 𝑦based
) on standard intersection and union

𝐴 ( 𝑦 ) ∩ 𝐵 ( 𝑦 )={
0 .3 0 .5 0. 8 1 0 .7 0.1 0
+
𝑦 1 𝑦2
+ + + + +
𝑦3 𝑦 4 𝑦5 𝑦6 𝑦7 }
𝐴 ( 𝑦 ) ∩ 𝐵 ( 𝑦 )={
0 .7 0 .5 0.2 0 0 .3 0.9 1
+ + + + +
𝑦1 𝑦2 𝑦 3 𝑦4 𝑦5 𝑦 6 𝑦7
+ }
Dr. Van-Phong Vu-Department of Automatic Control
34

2.3 Fuzzy number

Dr. Van-Phong Vu-Department of Automatic Control


35 2.3.1 Definition
 Fuzzy number is a fuzzy set to express the number in which the
fuzzy set needs to satisfy the following conditions:
 Fuzzy set must be a normal fuzzy set
 ⍺-cut with have to be a closed interval
 Support of the fuzzy set is finite.

A(x) A(x)
1 1

0 0
1 2 3 5
Dr. Van-Phong Vu-Department of Automatic Control
Fuzzy number 2 Fuzzy number 5
A(x)
36 0.5

0
1 2 3
Non-fuzzy number

Non-fuzzy number Non-fuzzy number


Dr. Van-Phong Vu-Department of Automatic Control
37 2.3.2 Decomposion theorems
 Define:

 A( x )    A( x)
 Continuous:

A(x)
1 1
⍺A(x)

0 0
1 2 3 1 2 3

Dr. Van-Phong Vu-Department of Automatic Control


38
 Discrete time:

𝐴 ( 𝑦 )={
0 .1 0.2 0. 8 1 0 .7 0.2 0
𝑦1
+ + + + +
𝑦2 𝑦3 𝑦 4 𝑦5 𝑦6 𝑦7
+
}
𝛼 𝐴 ( 𝑦 )={
0 0 0. 5 0.5 0 .5 0
+ + + + + +
𝑦1 𝑦2 𝑦3 𝑦 4 𝑦5 𝑦6 𝑦7
0
}
=0.5

Dr. Van-Phong Vu-Department of Automatic Control


39  First Decomposition Theory (reference WJ Wang)

A

 A
[ 0 ,1]

5 A
5
2 A
2
‧ 2 ‧ x
A
‧5 ‧
A
 Any fuzzy set A can be composed many i A
(each can be A as a stair
 i see
with height)

Dr. Van-Phong Vu-Department of Automatic Control


40 2.3.2 Operation on fuzzy number
 Operation on Interval
 Consider two interval [a,b] and [c, d].
 [a, b]+[c, d]=[a+c, b+d].
 [a, b]-[c, d]=[a-d, b-c]
 Example:
 [1, 3]+[2, 4]=[3, 7]
 [2, 5]-[3, 6]=[-4, 2].

Dr. Van-Phong Vu-Department of Automatic Control


41  Operation on the fuzzy number ( Refer WJ Wang’s book and
lecturer)
 Consider two fuzzy number A and B.
 Based on definition of ⍺-cut, we have:

 where * is the plus (+), minus (-),


 Based on the Decomposition Theory, it yields

 A and B are fuzzy set, therefore A*B is fuzzy set as well.

Dr. Van-Phong Vu-Department of Automatic Control


42  Example:
 Consider two fuzzy number as follows:

 Compute: A+B and A-B?

Dr. Van-Phong Vu-Department of Automatic Control


 Step 1: Draw the two fuzzy sets and carry out ⍺-cut

43

 Step 2: calculate ⍺-cut of A, B and (A+B)

Dr. Van-Phong Vu-Department of Automatic Control


44  Step 3: compute the (A+B)
 Left side: choose ⍺=0 and ⍺=1 we have two points: (0,0) and (4, 1), thus we can
obtain the equation of left side

 Right side: choose ⍺=0 and ⍺=1, we have two points: (4,1) and (8,0), Then we
obtain the equation of right side

Dr. Van-Phong Vu-Department of Automatic Control


45

Dr. Van-Phong Vu-Department of Automatic Control


46

Thanks for Listening

Dr. Van-Phong Vu-Department of Automatic Control

You might also like