Professional Documents
Culture Documents
Set:
A classical (crisp) set can be defined (within a universe of discourse
or universal set U) by :
1. naming all its members (used only for finite sets).
e.g. A={a1, a2, .., an}
2. a property satisfied by its members,
A={x| p(x)}.
p(x): property of x.
3. by a characteristic function, A(x), as follows
i .e .
iff
x A
1
iff
x A
0
: U {0 , 1 }
(x ) =
Fuzzy set :
1-2
A (T )
hot
1
50
A=
{( x ,
( x )) | x U }
A :U M .
Ex2-1:
1. U:R, crisp set A represents real numbers 5
A =
and
{( x ,
(x ) =
( x )) |
xU
1-3
,x < 5
,x 5
A ( x) =
1
1 + 10 ( x 5 )
1-4
Remark:
the assignment of the membership function of a fuzzy set is
subjective. in the above example, we may assign :
A ( x) =
1
1 + ( x 5)
Supp ( A ) = { x U | A ( x ) > 0}
1-5
Ex 2.2:
U:{10, 20,,100}. Fuzzy sets A High Score,
B Medium Score, C Low Score.
Membership function :
Numerical Score
10
20
30
0.1
0.9
40
0.5
0.7
50
0.1
0.8
0.5
60
0.3
0.3
70
0.5
0.8
0.1
80
0.8
0.5
90
100
1-6
Then:
U=[0,100]
A ( x ) = H ig h S co re
1
( x ) = ( x 4 0 ) 2
1 +
5
S u p p ( A ) = ( 4 0,1 0 0 ]
Fuzzy singleton:
x1
0 x 40
40 < x 100
1-7
Crossover point:
{ x U | ( x ) = 0.5}
A
Kernel:
Ker ( A) = { x U | A ( x ) = 1}
Height:
1-8
A={(x1, A(x1)), (x2, A(x2)) ,, (xn, A(xn))}.
e.g. :
B= Medium Score in Ex 2.2.
={(10, 0), (20, 0), (30, 0.1), (40, 0.5), (50, 0.8), (60, 1), (70, 0.8),
(80, 0.5), (90, 0), (100, 0)}
A = 1 x1 + 2 x 2 +
where
+ n xn =
xi
i =1
i = A ( xi ) > 0
e.g. : B=0.1/30+0.5/40+0.8/50+1/60+0.8/70+0.5/80.
~ if U is R, then A is written as
A =
(x )
e.g. : A in Ex 2.1
A=
R 1 + 10 ( x 5 )
1-9
-cut (or -level set) of A:
A crisp set A s. t.
A = { x U | A ( x ) } ,
( 0,1] .
U
A
level set of A:
the set of all levels (0,1] that represents distinct -cut of a given
fuzzy set A,
i.e.
A = | A ( x ) = , for som e x U .
1-10
Ex 2.3:
In Ex 2.2.
, then A A
A
A
Thm 2.1:
Let A be a fuzzy set in U. Then A(x) can be expressed as
A ( x ) = sup A ( x ) .
( 0,1]
x U
1-11
where denotes the min operation and
1
A ( x ) =
0
x A
iff
otherwise
sup
( 0 ,1 ]
=
=
A(x)
x
A
sup
(0 ,
A ( x )
( x )
sup
(0 ,
( x )
s u p [
(0 ,
1]
( x )
sup
s u p [
( x ),1
A ( x ),1
( x )
0]
( x )
( x ).
1-12
Resolution principle (decomposition thm. or representation thm.):
A ( x ) = A ( x ) ,
x U
A=
or
A = A
0
where U: union.
Pf: by Thm 2.1
sup ( x ) = sup ( x ) = ( x )
A
A = A
1-13
0.3 A
Ex:
0.3
0.3A0.3
0.3
x
A(x)
0.6 A
0.6
0.6
0.6
0.3
0
A1
A0.6
A0.3
0.6A0.6
x
x
1 A
1A1
Ex 2.4:
In Ex 2.2. we have A=0.1/50+0.3/60+0.5/70+0.8/80+1/90+1/100.
A = 0 .1 5 0 + 0 .1 6 0 + 0 .1 7 0 + 0 .1 8 0 + 0 .1 9 0 + 0 .1 1 0 0
+ 0 .3 6 0 + 0 .3 7 0 + 0 .3 8 0 + 0 .3 9 0 + 0 .3 1 0 0
+ 0 .5 7 0 + 0 .5 8 0 + 0 .5 9 0 + 0 .5 1 0 0
+ 0 .8 8 0 + 0 .8 9 0 + 0 .8 1 0 0
+ 1 90 + 1 100
= 0 .1 (1 5 0 + 1 6 0 + 1 7 0 + 1 8 0 + 1 9 0 + 1 1 0 0 )
+ 0 .3 (1 6 0 + 1 7 0 + 1 8 0 + 1 9 0 + 1 1 0 0 )
+ 0 .5 (1 7 0 + 1 8 0 + 1 9 0 + 1 1 0 0 )
+ 0 .8 (1 8 0 + 1 9 0 + 1 1 0 0 )
+ 1 (1 9 0 + 1 1 0 0 )
= 0 .1 A0 .1 + 0 .3 A0 .3 + 0 .5 A0 .5 + 0 .8 A 0 .8 + 1 A1
=
1-14
1-15
On the other hand, given
A =
A =
A
{ 0 .1, 0 .4 , 0 .8 ,1 ]
= 0 .1 A 0 .1 + 0 .4 A 0 .4 + 0 .8 A 0 .8 + 1 A1
= 0 .1 (1 1 + 1 2 + 1 3 + 1 4 + 1 5 )
+ 0 .4 (1 2 + 1 3 + 1 5 )
+ 0 .8 (1 2 + 1 3 )
+ 1 (1 3 )
= 0 .1 1 + 0 .8 2 + 1 3 + 0 .1 4 + 0 .4 5 .
1-16
Convex set:
A set S is convex if x1 , x 2 S . R
the point x 1 + (1 ) x 2 S
e.g. :
x1
x2
x2
convex
nonconvex
if the -cuts of a fuzzy set are convex for all ( 0,1] then the fuzzy
set is convex.
e.g. :
1
2
A1
A2
1-17
Thm: A fuzzy set A on R is convex iff
A ( x1 + (1 ) x2 ) min A ( x1 ) , A ( x2 )
for all x1 , x2 R, [ 0,1] .
A
x1
x2
x1
x2
Pf:
(1)
let = A ( x1 ) A ( x2 ) . then x1 , x2 A .
by the convexity of
A. x1 + (1 ) x2 A . [ 0,1]
A ( x1 + (1 ) x2 ) = A ( x1 ) = min A ( x1 ) , A ( x2 )
1-18
(2)
Here , we need to prove that ( 0,1] , A is convex.
x1 , x 2 A i.e. A ( x1 ) , A ( x 2 ) . [ 0,1] .
Remark:
(1) Convex function:
A function f defined on a convex set S is said to be convex
if, x1 , x2 S , [ 0,1] , there holds
f ( x1 + (1 ) x2 ) f ( x1 ) + (1 ) f ( x2 ) .
(2) The convexity of a fuzzy set does not mean that the
membership function of a convex fuzzy set is a convex function.
1-19
e.g. :
x1
x2
x1
x2
Fuzzy number:
xm
1 + a ,
m x
,
A ( x ; m , a , b ) = 1 +
b
0,
m a x m
m x m +b
o th e rw ise
1-20
~Trapezoid membership function
0,
m+a x
m x
, m x x+a
A (x; a, m ) = 1 +
a
1,
otherwise
0
m
m+a
exp(1)
0
m
m +
1-21
Cardinality:
~ fuzzy set:
x U
A ( x ).
rel
A
U
, where U is finite.
/ A
A
Ex 2.5:
Consider the fuzzy set in Ex 2.2 (p 1-5).
A = High score = 0.1 + 0.3 + 0.5 + 0.8 + 1 + 1 = 3.7
A ref =
A
A
=
= 0.37
U 10
1-22
Set-theoretic definitions and operations for fuzzy set
1. Complement:
A ( x ) [ 0 ,1 ] .
A (x)
1 A ( x ).
A : c o m p le m e n t o f A
x U
2. Intersection:
denoted as AB.
AB ( x ) min A ( x ) , B ( x ) A ( x ) B ( x )
x U
Remark : A B A and A B B
3. Union:
A B, AB ( x) max A ( x) , B ( x) A ( x) B ( x) x U
Re mark : A A B and B A B
1-23
1
0.5
0
(b) A B.
(a ) A.
(c) A B.
4. Equality:
A and B are equal iff A ( x ) = B ( x ) , x U
Remark: To check the degree of equality of two fuzzy sets, we
can use the similarity measure:
A B
E ( A, B ) deg ree ( A = B )
A B
when A = B , E ( A, B ) = 1.
when
E ( A, B ) = 0.
1-24
5. Subset:
A is a subset of B, i.e.
A B iff
A ( x) B ( x),
A is a proper subset of B if A
Remark: subsethood measure:
x U
B.
s ( A, B ) = deg ree ( A B )
A B
A
( A B )
but
= A B ,
( A)
A = A.
( A B )
= A B ,
1-25
7. De Morgans laws:
A B = A B,
A B = A B.
Remark:
1. the law of the excluded middle i.e. E E = U and
the law of contradication i.e. E E = are no longer
true in fuzzy sets. That is,
A A U , A A .
2. The above definitions are not unique, we will see other
definitions later in this chapter.
8. Cartesian product:
Let A1, A2, , An be fuzzy sets in U1, U2, , Un, respectively.
The Cartesian product of A1, A2, , An is a fuzzy set in the
product space U1 U 2 U n with
A A
1
An
( x1 , x2 ,
x1 U1 , x2 U 2
, xn )
min A1 ( x1 ) , A2 ( x2 ) ,
, An ( xn )
, xn U n.
1-26
9. Algebraic sum:
A + B.
A + B ( x ) A ( x ) + B ( x ) A ( x )i B ( x ) .
A B.
AiB ( x )
A ( x )i B ( x ) .
A B.
A B ( x ) min 1, A ( x ) + B ( x ) .
m a x 0 , A ( x ) B ( x ) .
Ex2.7:
1-27
Other types of fuzzy sets:
A ( x ) = [ 1 , 2 ]
2
1
0
Type 2 fuzzy set:
A (x ) =
f (u )
f : [ 0 ,1 ]
u,
1-28
u [ 0 ,1 ].
[ 0 ,1 ] .
A ( x)
u
1
f x1 ( u )
x1
x2
f x2 ( u )
1-29
Level-2 fuzzy set:
A : p (U ) [ 0,1]
~ Different functions are defined for fuzzy sets operation and can be
divided into two categories. One is non-parametric function and the
other is parametric function in which parameters are used to adjust
the strength of the corresponding operations.
~ Complement:
is specified by a function
c : [0,1] [0,1]
s.t.
(x)=c(A(x)).
where c( ) satisfies:
c1: Boundary conditions : c(0)=1, c(1)=0
1-30
c2: Monotonic property:
x1 , x2 U. if A(x1) A(x2), then
c(A(x1)) c(A(x2)).
C3: Continuity: c( ) is a continuous function.
C4: Involution: c( c( A (x)) )= A (x), x U
Based on the above conditions, typical functions are:
1. Negation complement:
A : A ( x) = c( A ( x)) 1 A ( x), x U
A : A ( x ) = c ( A ( x ) )
1 A ( x)
.
1 + A ( x )
-1< < .
A :
( x ) = c ( A ( x ))
(1 ( x ) )
w
A
1
w
, 0<w<
observation : w=1, c ( A ( x ) ) = 1 A ( x ) .
1-31
Remark: The equilibrium of a fuzzy complement c is defined as
any value for which c(a)=a, e.g. a=0.5 in standard
complement operation. Every fuzzy complement has at
most one equilibrium due to the monotonic property.
~ Intersection:
often referred to as triangular norms (t- norms):
t- norms are of the form:
t : [0,1] [0,1] [0,1]
s.t.
AB(x) = t [A(x), B(x)]
where t( , ) satisfies:
1-32
t 3. Monotonicity:
if A ( x ) C ( x ) and B ( x ) D ( x ) , then
t ( A ( x ) , B ( x ) ) t ( C ( x ) , D ( x ) ) .
t 4. Associativity:
t A ( x ) , t ( B ( x ) , C ( x ) ) = t ( t(A ( x ) , B ( x )), C ( x ) ) .
Let a =A(x), b=B(x), typical non-parametric t-norms are:
1. Intersection: ab = min(a , b).
2. Algebraic product: ab = ab
3. Bounded product: ab = max( 0 , a+b-1 )
b
4. Drastic product: ab=
0
b=1
a=1
a, b < 1
5. Yager intersection:
t w ( a,b ) = 1 min 1 ,
(1 a ) + (1 b )
w
1
w w
w ( 0, )
1-33
~ Union:
often referred to as triangular conorms (t- conorms).
t- conorms are of the form.
S: [0,1][0,1] [0,1].
s .t.
A B ( x ) = s A ( x ) , B ( x ) .
w h e re
s ( i , i ) s a tis f ie s :
s 1 . B o u n d a r y c o n d itio n s :
s (1 ,1 ) = 1, s ( A
s 2 . C o m m u ta tiv ity :
( x ) , 0 ) = s (0 , A ( x )) = A ( x ).
s ( A ( x ) , B ( x )) = s ( B ( x ) , A ( x )).
s 3 . M o n o to n ic ity :
if
A ( x ) C ( x ),
s ( A
B ( x ) D ( x ) , th e n
( x ) , B ( x )) s ( C ( x ) , D ( x )).
s 4 . A s s o c ia tiv ity :
s A
( x ) , s ( B ( x ) , C ( x ))) =
s s ( A
( x ) , B ( x )) , C ( x )).
1-34
~ typical nonparametric t- conorms are
1. Union: ab = max(a,b)
2. Algebraic sum: a + b = a + b ab
3. Bounded sum: a b = min(1, a+b)
4. Drastic sum: ab
= b
1
,b=0
,a=0
, a, b > 0
1-35
pf of (1) :
2nd inequality:
By boundary condition,
t (a, 1) = a, t (1, b)= b.
By monotonicity condition, t (a, b) t (a, 1) = a.
t (a, b) t (1, b)= b
t (a, b) min(a, b)
..
t (a , b) a b.
Remark: the Yager intersection and Yager union become
tmax and Smin as w and become
tmin and Smax as w 0.
Aggregation operation:
defined by
s.t.
A ( x ) = h ( A ( x ) , A ( x ) ,
1
, An ( x ) . x U.
min ( a1 , a2 ,
, an ) h ( a1 , a2 ,
where ai = Ai ( x ) , i = 1, 2,........., n
, an ) max ( a1 , a2 ,
, an )
a1 + a2 + + an
h ( a1 , a2 , , an )
, R. 0
n
- h become min ( a1 , a2 , , an )
h become max ( a1 , a2 , , an ) .
1-36
1-37
Yager class
Yager class
sw
tw
Generalized means
-
h
max
min
tmin
tmax
t- norms
smax
Smin
Averaging operations
(Intersection operations)
t- conorms
(Union operations)
Aggregation operations
h ( a1 , a 2 ,
wi 0,
, a n ; w1 , w2 ,
, wn )
(w a
1 1
+ wn a n )
+ w2 a 2 +
= 1.
i =1
Other operations:
1. Fuzzy conjunction: AB
A B ( x, y ) t ( A ( x ) , B ( y ) ) . t: t-norm.
2. Fuzzy disjunction: AB
A B ( x, y ) s ( A ( x ) , B ( y ) ) . s: t-conorm.
3. Fuzzy implication: AB
M aterial implication : A B=s A,B
Propositional calculus : A B=s A , t ( A,B )
Extended propositional calculus : A B=s A B , B
Generalization of modus porens :
(
(
1-38
1-39
Ex: t: algebraic product.
s: algebraic sum.
Material implication is adopted.
A B ( x, y ) = A ( x ) B ( y )
A B ( x, y ) = A ( x ) + B ( y ) A ( x ) B ( y ) .
A B ( x , y ) = s (1- A ( x ) , B ( y ) )
=1- A ( x ) + B ( y ) - (1- A ( x ) ) B ( y )
=1- A ( x ) + A ( x ) B ( y ) .
Extension principle:
1-40
If more than one element of U is mapped to the same element y in
V by f ( i.e. many-to-one mapping ), then the maximum among their
membership grades is taken. That is
f ( A ) ( y ) = sup A ( xi ) .
xi U
f ( xi ) = y
B=
{( y ,
where B ( y ) =
( y )) |
sup
( x1 , x2 , xn ) U
y = f ( x1 , x2 , xn )
y = f ( x1 , x2 ,
xn ) , ( x1 , x2 ,
min A1 ( x1 ) , A2 ( x2 ) ,
, xn ) U ) .
, A n ( xn )
1-41
Ex: U = {1, 2, 3, 9,10}. A= Large is given as
A = Large = 0.5/6 + 0.7/7 + 0.8/8 + 0.9/9 + 1/10.
If y=f(x)=x2, by extension principal, B= "Large"2 can be
calculated as :
B=0.5/36+0.7/49+0.8/64+0.9/81+1/100.
Ex: U={-2 , -1 , 0 , 1 , 2}. A= 0.5/-1 + 0.8/0 + 1/1 + 0.4/2.
y=f(x)=x2 ,
B(x)
A(x) y=f(x)=x2
x
-1
0.5
Max{0.5 , 1.0}=1.0
0.8
0.8
1.0
Max{0.5 , 1.0}=1.0
0.4
0.4
1-42
Ex: f : U1U2V, U1 = {- 1, 0, 1}, U2 = {- 2, 2}, V = {- 2, - 1, 2, 3}.
f(x1 , x2) = x12+x2
A1: defined on U1 and A1 = 0.5/-1 + 0.1/0 + 0.9/1.
A2: defined on U2 and A2 = 0.4/-2 + 1.0/2.
x1 A1 x2 A2 A1A2(x1 , x2)
-1 0.5 -2 0.4 min{0.5 , 0.4}=0.4
y=f(x1 , x2 )= x12 + x2
-1
-1
0.5
1.0
min{0.5 , 1.0}=0.5
0.1
-2
0.4
min{0.1 , 0.4}=0.1
-2
0.1
1.0
min{0.1 , 1.0}=0.1
0.9
-2
0.4
min{0.9 , 0.4}=0.4
-1
0.9
1.0
min{0.9 , 1.0}=0.9
1-43
Appendix: Factual Information about the Impact of Fuzzy Logic
(data from e-News of Berkeley Initiative in Soft Computing (BISC), Nov. 2004)
~ PATENTS
. Number of fuzzy-logic-related patents applied for in Japan: 17,740
. Number of fuzzy-logic-related patents issued in Japan: 4,801
. Number of fuzzy-logic-related patents issued in the US: around 1,700
~ PUBLICATIONS
Count of papers containing the word "fuzzy" in title, as cited in INSPEC
and MATH.SCI.NET databases.
Compiled by Camille Wanat, Head, Engineering Library, UC Berkeley,
May, 2006
Number of papers in INSPEC and MathSciNet which have "fuzzy" in their titles:
INSPEC - "fuzzy" in the title
1970-1979:
569
1980-1989:
2,403
1990-1999:
23,210
2000-2006(5) 21,147
Total:
47,329
1-44
~ JOURNALS
~ Applications
1-45
The range of application-areas of fuzzy logic is too wide for exhaustive listing.
Following is a partial list of existing application-areas in which there is a record of
substantial activity.
1. Industrial control
2. Quality control
3. Elevator control and scheduling 4. Train control
5. Traffic control
6. Loading crane control
7. Reactor control
8. Automobile transmissions
9. Automobile climate control
10. Automobile body painting control
11. Automobile engine control
12. Paper manufacturing
13. Steel manufacturing
14. Power distribution control
15. Software engineering
16. Expert systems
17. Operation research
18. Decision analysis
19. Financial engineering
20. Assessment of credit- worthiness
21. Fraud detection
22. Mine detection
23. Pattern classification
24. Oil exploration
25. Geology
26. Civil Engineering
27. Chemistry
28. Mathematics
29. Medicine
30. Biomedical instrumentation
31. Health-care products
32. Economics
33. Social Sciences
34. Internet
35. Library and Information Science
~ Product Information
1. Information from SIEMENS:
washing machines, 2 million units sold
fuzzy guidance for navigation systems (Opel, Porsche)
OCS: Occupant Classification System (to determine, if a place in a
car is occupied by a person or something else; to control the airbag as
well as the intensity of the airbag). Here FL is used in the product as well
as in the design process (optimization of parameters).
fuzzy automobile transmission (Porsche, Peugeot, Hyundai)
2. Information from OMRON:
fuzzy logic blood pressure meter, 7.4 million units sold, approximate retail
value $740 million dollars
3. Facts on FL-based systems in Japan (as of 2/06/2004)
Sony's FL camcorders
Total amount of camcorder production of all companies in 1995-1998
times. Sony's market share is the following. Fuzzy logic is used in all
Sony's camcorders at least in these four years, i.e. total production of
Sony's FL-based camcorders is 2.4 millions products in these four years.
1,228K units X 49% in 1995
1,315K units X 52% in 1996
1,381K units X 50% in 1997
1,416K units X 51% in 1998
1-46
1-47
Canon
Canon used (uses) FL in their cameras, camcorders, copy machine, and
stepper alignment equipment for semiconductor production. But, they have
a rule not to announce their production and sales data to public.
Canon holds 31 and 31 established FL patents in Japan and US, respectively.
Minolta cameras
Minolta has a rule not to announce their production and sales data to
public, too, whose name in US market was Maxxum 7xi. It used six FL systems in a
camera and was put on the market in 1991 with 98,000 YEN (body price
without lenses). It was produced 30,000 per month in 1991. Its sister
cameras, alpha-9xi, alpha-5xi, and their successors used FL systems,
too. But, total number of production is confidential.
1-48
FL plant controllers of Yamatake Corporation
Yamatake-Honeywell (Yamatake's former name) put FUZZICS, fuzzy
software package for plant operation, on the market in 1992. It has been
used at the plants of oil, oil chemical, chemical, pulp, and other industries
where it is hard for conventional PID controllers to describe the plan
process for these more than 10 years.
They planed to sell the FUZZICS 20 - 30 per year and total 200 million
YEN.
As this software runs on Yamatake's own control systems, the software
package itself is not expensive comparative to the hardware control
systems.
Others
Names of 225 FL systems and products picked up from news articles in
1987 - 1996 are listed at http://www.adwin.com/elec/fuzzy/note_10.html
in Japanese.)