n
A Note on Convex Fuzzy Sets
Yudong Zou Kaigi
ce, Dalian Maritime University, Dalian, 116026, P-R. China
Department of Basic S
Abstract
This note is to give the additional properties which are required to characterize convex sets.
Keywords: Convex fuzzy sets, the monotonicity.
1, Introduction
Some properties of convex fuzzy sets were studied by Lowen [1], Liu [2] and yang [3,4]. In
this note some additional properties which are required to characterize convex fuzzy sets are
discussed. These additional properties are expressed in terms of the following line segment
characterization of convex fuzzy sets and the monotonicity.
2. Prelimi
Similar to [3] and [4], throughout this paper E will denote the n-dimensional Euclidean
space R*, J denotes the interval [0,1], /°denotes ( 0,1) and [x, y] denotes set {ax+(1-a)y | ae}
Fuzzy sets and values in J will be denoted by lower case Greek letters and we shall make no
difference between notations for a fuzzy set with a constant value and that value itself.
Definition 1. The fuzzy set A on E is said to be a convex fuzzy set if and only if for all xyeE
and ael,
Max+(1-aly) = AA)AKG)
It is easy to sce that 2 is a convex fuzzy set if and only if for all a€/, A[a,1] is convex.
3. Main results
This section gives the main results of this note. That is, we obtain some additional
properties which are required to characterize convex fuzzy sets.
Theorem 1. The fuzzy set A on £ is a convex fuzzy set if and only if for all line /in E, A isa
convex fuzzy set on INE.
‘The proof is straightforward by Definition 1
‘Theorem 2, The fuzzy set 2 on E is a convex fuzzy set if and only if for all x, ye £, thereB
exist intervals A (0eA) and B (0€B), such that fla) = Max +(I-a)y) is a increasing function on
A, is a descent function on B, and AUBDI.
Proof. Let A be a convex fuzzy set on E.
(i) For V x, ye E, let
A= {acl |S fla), W Osbsa} a
and
supd =a, @
Without loss of generality, suppose that a, >0, now we prove that (0, a, f(b). Form (2), there exists ab, such that b,>b. Consequently, we have (by) 2fb,) >
Ab), that is,
Ab) < flb,) » lb) @)
Now, let ¢ 1%, =xtb,(y-x) and y, =x +b,(y-2), we have ce? and
eyyt(L-e)e, = xy Felyy-x;) = x +60)
Also, 2. is a convex fuzzy set, it follows that
Ab) = Mx +b(y-x)) = Meyy+(1-c)x, 2 Mexy)AA(2) = Alby) a Alb.)
contradicting (3). Thus, [0, a, )cA.
Gi) For V xye E, let
A= {bel| Ay +b(x-y)) 2 My +b, (wy), V 05,5)
and
supB=b,
Similar to (i), we can show that (0, b, )f (a) by aé, there exists a.a”e/ and
such that f(a") >f(a) by a€B. So we have f(a)f(a')af(a")
Similar to (i), this leads to a contradiction.
‘Theorem 3. Let A be a convex fuzzy set on E. If there exists a a,¢/° such that
Magx+(l-aa)y )2 Mx)aMy), Vx, ve EB
Then set
M= (ae? | Maxt(1-a)y 2 AQx)AAQ), Wx, ye E}
is infinite and dense everywhere in
Proof. Vx, ye E, since a¢M and
M(-a)tagy ) = Magy+(1-aq)x )2 A(x) 0A(y)4
we have 1-aeM . Now, agrt(I-a.)ye E, and
aolaoxt(1-ay)y}+(1-aa)y = a5 x#(1-a5 Jy
Therefore, Aa} x+(1-a2 WyI2Aagr+(I-aly AMIE MAAK)
that is, a eM. It follows that at, (i-a!)eM, Vkeg¥~ by mathematical induction. Thus,
Mis infinite,
Now, we prove that M is dense everywhere in P' If that is not true, then there exist @
a, =supV(a)
thatis, (4,2, )oM=@ (4)
Choose ae (a) and a,¢U(a) such that J,