Page 4-1 Chapter 4 Fuzzy Arthmetic 4.1 Fuzzy Numbers ( or Fuzzy intervals ) --- Fuzzy sets defined on R Membership function A:R[0,1] e.g., i, close to a real number ii, around a given interval of real numbers ◎ Applications : fuzzy control, decision making approximate reasoning, optimization, statistics with imprecise probabilities. ◎ Required properties of a fuzzy number A ( i ) A : a normal fuzzy set (ii) (iii) A 0 : a convex set, 0,1 A : bounded support Page 4-2 ◎ Cases of fuzzy numbers Crisp fuzzy an ordinary member an interval of numbers ◎ Membership functions of fuzzy numbers need not be symmetric large number small number Page 4-3 ◎Membership function may be piecewise-defined Theorem 4.1 : Let A F ( R) A : a fuzzy number fuzzy power iff a, b Normal b. Convex c. bounded x a, b 1 A( x) l ( x) r ( x) Where a. s.t. a. x , a ----- ( 4.1 ) x b, b, Monotonic increasing l : , a 0,1 : c, Continuous from right d , l ( x) 0, x (, w ) 1 Monotonic decreasing r : (b, ) 0,1 : Continuous from left r ( x) 0, x ( w ) 2, Proof: I, hecessity => 1. A : normal 1A Let 1 A a, b 2. prove that the left part of A agree with l (a )l : ( a , ) 0 , 1 (b) Monotonic increasing (c) Continuous from right (d )l ( x) 0, x (, w ) 1 Page 4-4 a, 0 A( x) 1, x (, a) ( A (a ) 1 ) with l : (, a) 0,1 A : ( a, ) 0 , agree 1 b. Let x y a, A( y) min A( x), A(a) A( x) Theorem1.1 (1) A(a) 1 (A:convex) A : increasing agree with l : monotonic increasing c. assume A(x): not continuous from the right Let x0 (, a) , be a discontinuous pt. a sequence xn , s.t. xn x0 , n and lim xn x0 n increasing Let lim A( xn ) A( x0 ) n A : closed interval ( x0 A A : convex) contradict A( x0 ) A( x) : continuous from right d. A : fuzzy number 0 A : bounded w1 , s.t. A( x) 0, x , w1 Page 4-5 3, prove r : monotonic decreasing Continuous from the left r : (b , ) [ 0 , 1] r ( x) 0 x (2w , ) II. Sufficiency i. A defined by (4.1) is normal -----(a) ii. iii. and its support 0 A : bounded----(b) prove (0,1], A : a closed interval (i.e. A: convex) proof : let X inf{x | l ( x) , x a} ----(c) Y s u p {x r| x( ) x, ---(d) b } Show that A : closed (i.e., A [ X , Y ] ) (A) prove A [ X , Y ] proof : x0 A , if x0 a A( x0 ) l ( x0 ) i.e., x0 {x | l ( x) , x a} x0 i n f {x l| x( ) x, a }---(i) x Page 4-6 if x0 b A( x0 ) r ( x0 ) i.e., x0 {x | r ( x) , x b} x0 s u p {x r| x( ) x, b } y x0 [ X , Y ] A [ X , Y ] (B) prove [ X , Y ] A (idea [ x , y ] A :) proof : from (c) , i.e. X inf{x | l ( x) , x a} a sequence {xn } in {x | l ( x) , x a} s.t. lim xn x n l : continuous from right l ( x ) x ( l ixm ll xi nm ( ) n ) n n x A similarly [ x , y ] ( A ) (B ) x[ y A A y, ] :Aa closed interval (i)(ii)(iii)=> complete sufficiency (I)(II) => complete theorem ※ Every fuzzy number can be represented in the form of (4.1) Page 4-7 Examples (a) w1 a b w2 1.3 x (,1.3), l ( x) 0 x (1.3, ), r ( x) 0 (b) w1 a 1.25, w2 b 1.35 x (,1.25), l ( x) 0 x (1.35, ), r ( x) 0 (c) a b 1.3, w1 1.2, w2 1.4 l ( x) r ( x) x (,1.2) 0 1 0 x( 1 .3 ) 1 x [ 1 . 2 , 1 . 3 ) 1 0 ( 1 . 3x ) 1 x( 1 . 3 , 1 . 4 ] x (1.4, ) 0 (d) a 1.28, b 1.32 w1 1.2, w2 1.4 l ( x) r ( x) 0 1 2 . 5x ( x (,1.2) 1 .2 8 ) 1 2 . 5 ( 1 . 2x 8 0 ) 1x [ 1 . 2 , 1 . 2 8 ) 1x ( 1 . 3 2 , 1 . 4 ] x ( 1 . 4, ) Page 4-8 a = 90, b=100 w1 77.5, w2 100 x , 77.5 0 l ( x) 0.08( x 90) 1 x 77.5, r ( x) 0, x (100, ) ◎ Fuzzy Cardinality A A. a fuzzy number defined on N A Example A Page 4-9. | D2 | 0.13 0.27 0.4 0.53 0.67 0.8 0.93 1.0 19 17 15 13 11 9 7 5 e.g. x {30,50}, D2 ( x) 0.67 50 30 1 11 2 B. Scalar Cardinality | A | | A | | A | A( x) xsup( A) Example : | D2 | 0 . 1 3 * 2 0 . 27 * 2 0 . 4 * 2 0 . 5 3 * 2 0 . 67 ** 22 10*. 85 * 21 2 .04. 69 3 Page 4-10 4.3 Arithmetic Operations on Fuzzy Numbers Idea: Define operations in terms of arithmetic operations on α-cuts which are a subject of interval analysis of classical mathematics. ※ Fuzzy number can uniquely be represented by theirα-cuts. ⊙Classical interval analysis ‧ 4 operations: +,-,‧,/ Let * denote any of the 4 operations→ a, b d , e f g | a f except a, b/d , e when b, d g e 0 [ d , e] ※ The result of an operations on closed intervals is a closed interval. Page 4-11 ●Definitions of operations [a, b] [d , e] [a b, b e] [a, b] [d , e] [a b, b e] [a, b] [d , e] [min( ad , ae, bd , be), max( ad , ae, bd , be)] [a, b] /[ d , e] [a, b] [1 / e,1 / d ] [min( where 0 [d , e] ●Examples: 1.[2,5]+[1,3]=[3,8] 2.[2,5]-[1,3]=[-1,4] 3.[3,4]‧[2,2]=[6,8] 4..[4,10]/[1,2]=[2,10] a a b b a a b b , , , ), max( , , , )] d e d e d e d e Page 4-12 ●Properties of Operations Let A [a1 , a2 ] , B [b1 , b2 ] , C [C1 , C2 ] , 0 [0,0] , 1 [1,1] A B B A 1. Commutativity A B B A ( A B) C A ( B C ) 2. Associativity ( A B) C A ( B C ) A 0 A A 0 3. Identity A 1 A A 1 4. Subdistributivity A ( B C ) A B A C 5. Distributivity If b B, c C , b c 0, A ( B C ) A B A C 6. 0 A A,1 A / A,0 A 7. Inclusion monotonicity If A E , B F A B A B A B A / B EF EF EF E/F Page 4-13 ● Prove Subdistrubutivity: ( A ( B C ) A B A C ) Proof: A ( B C ) {a (b c) | a A, b B, c C} {a b a c) | a A, b B, c C} Less Re striction (loose ) {a b a c) | a, a A, b B, c C} A B A C ( A (B C) A B A C) ● Example: Let A=[0,1], B=[1,2], C=[-2,-1] A B [0,1] [1,2] [0,2] A C [0,1] [2,1] [2,0] B C [1,2] [2,1] [1,1] A ( B C ) [0,1] [1,1] [1,1] A B A C [0,2] [2,0] [2,2] A (B C) A B A C Page 4-14 ˙ Prove Distributivity : ( f b B, c C , b c 0, A (B C) A B A C ) Proof : Let A [a1 , a2 ], B [b1 , b2 ], C [c1 , c2 ] 1. assume B,C i.e , b1 , c1 0 0 Case 1 : a1 0 A ( B C ) [a1 (b1 c1 ), a2 (b2 c2 )] [a1b1 a1c1 ), a2b2 a2c2 )] [a1b1 , a2b2 ] [a1c1 , a2c2 )] A [ ] 0 Case2 : a2 0 ( A) [a2 , a1 ] ( A) ( B C ) )( A) B ( A) C A A (B C) A B A C [ ] 0 Case3 : a1 0, a2 0 A ( B C ) [a1 (b2 c2 ), a2 (b2 c2 )] [a1 b2 , a2 b2 ] [ a1 c2, a2 c]2 A B A C A [ ] 0 B,C , i.e., b2 , c2 0 2. Similarity , 0 Page 4-15 4.4 . Arithmetic Operations on Fuzzy Numbers Fuzzy Arithmetic : Interval arithmetic Extension principle ⊙Interval arithmetic method ( A B) A B see theorem 4.2 (c) *: arithmetic operation From Theorem 2.5 A B ( A B) [0,1] ˙Example : Let A(x) = x 1, x 3 0 1 x 1 1 x 3 (x+1)/2 (3-x)/2 B(x) = x 1, x 5 0 1 x 3 (5-x)/2 3 x 5 A [2 1,3 2], B [2 1,5 2] now, (x-1)/2 ( A B) [4,8 4] ( 0 , 1 ] ( A B) [4 6, 2 4] ( 0 , 1 ] ( A B) = ( A / B) = 0.5 Both A, B 0 [4 2 12 5, 4 2 16 15] [4 1, 4 16 15] [ ( 2 1 ) / ( 2 1 ) , ( 3 2 ) / ( 2 [ ( 2 1 ) / (5 2 ) , ( 3 2 ) / ( 2 2 2 ( 0 , . 5 ] ( . 5 , 1 ] 1)(]0 , . 5 ] 1)(]. 5 , 1 ] Page 4-16 From [4,8 4] f (4 ) g ( 8 4 ) f ( ) 4 g ( ) 8 4 f ( x) x 4 g ( x) 8 x 4 0 1 0 4 4 i.e., 0 x 4 (A+B)(x)= 0 1 0 4 4 0 4 4 8 8 4 4 i.e., 4 x 8 0 x 0, x 8 x 4 0 x4 8 x 4 4 x 8 Page 4-17 ( A B)( x) 0 x 0, x 8 x 4 0 x4 8 x 4 ( A B )( x) ( A B )( x) 4 x 8 0 x 6, x 2 x6 4 6 x 2 2 x 4 2 x 2 x 5, x 15 0 1 2 3 ( 4 x ) 2 5 x 0 1 ( 1 x 2) 2 0 x3 1 2 4 ( 1 x ) 2 3 x 1 5 Page 4-18 0 [3 (4 X )1 / 2 ] / 2 Show ( A B)( X ) 1/ 2 (1 X ) / 2 [4 (1 X )1 / 2 ] / 2 X 5, X 15 5 X 0 0 X 3 3 x 15 [4 2 12 5,4 2 16 15], (0,0.5] From: ( A B) [4 2 1,4 2 16 15], (0.5,1] A. Let x 4 2 12 5 4 2 12 (5 x) 0 by 公式: x b b 2 4ac 2a 3 4 x (0,0.5] 2 As 00 0 3 4 x 3 4 x 其中 (不能成立)( 3 4 x 0, 4 x 3, 不可能) 2 2 3 4 x x 5 2 As 0.5 0.5 3 4 x 1 3 4 x ,0.5 (不能成立) 2 2 2 1 3 4 x x0 2 2 5 x 0 0.5 B. Let x 4 2 16 15 4 2 16 15 x 0 4 1 x (0,0.5] 2 As 0 0 0 4 1 x 4 1 x ,0 (不能成立) 2 2 4 1 x x 15 2 As 0.5 0.5 4 1 x 1 4 1 x ,0.5 (不能成立) 2 2 2 1 4 1 x x8 2 2 8 x 15 0.5 Page 4-19 x 4 2 16 15 4 2 16 15 x 0 4 (1 x) 2 (0.5,1] x 4 2 1 4 2 1 x 0 C. Let 4 (1 x) 2 4 (1 x) ,1 2 2 2 1 4 (1 x) 0.5 x8 2 2 4 (1 x) 5 8 x 15 x x8 2 11 (1 x) 2 (0.5,1] As 0.5 0.5 (1 x) 1 2 2 x 0 0 x 3 D. Let x 4 2 16 15 4 2 16 15 x 0 4 (1 x) 2 (0.5,1] As 0.5 , 0.5 0.5 4 (1 x) (不能成立) 2 1 4 (1 x) x8 2 2 As 1 1 4 (1 x) 2 4 (1 x) ,1 (不可能成立) 2 2 2 2 4 (1 x) x3 2 2 3 x 8 1 By (A),(B),(C),(D) (A).when 5 x 0 x 3 (4 x) 2 (B).when 8 x 15 x 4 (1 x) 2 (C).when 0 x 3 x (D).when 3 x 8 x i.e 4 (1 x) 2 x 5, x 15 0 3 ( A B) (1 x) 2 ( 4 x ) 2 ( 1 x ) 2 4 ( 1 x ) 2 5 x 0 0 x3 3 x 1 5 Page 4-20 (A/B)(x)= 0 Derivation : from x let x= (5x+1)/(2x+2) 0 (3-x)/(2x+2) 1/ 3 x 3 ( A / B) [ x 0 -1 2 1 3 - 2 , ] 2 + 1 2 + 1 2 1 3 - 2 , ] 5 2 2 + 1 1/3 x [0,0.5] [0.5,1] 2 1 , (2 +1)x=2 1 2x +x=2 -1 2 +1 ( 2x 2) 1x as 0 0 1 x 2 2x , (0, 0.5] 1 x x 1 ( ) 2 2x as 0.5 0.5 ii, Let , x3 (x+1)/(2-2x) [ i. -1 1 x 1 x 1 x x 0 [ ] 2 2x 3 2 3 x , 2 1 2x 2 (0.5,1] 1 x 3 x f3 ( x) 2 1 5x 1 , 5 2 2 2x 1 (0.5,1] 0 x 3 x iii, Let f 4 ( x) 3 2 3 x , 2 1 2x 2 1 (0.5,1] x 1 3 x iv, Let f 4 ( x) 1 x3 3