aa r X i v : . [ m a t h . G M ] M a y Fuzzy vectors via convex bodies
Cheng-Yong Du a , Lili Shen b, ∗ a School of Mathematical Sciences and V. C. & V. R. Key Lab, Sichuan Normal University, Chengdu 610068, China b School of Mathematics, Sichuan University, Chengdu 610064, China
Abstract
In the most accessible terms this paper presents a convex-geometric approach to the study of fuzzy vectors. Motivatedby several key results from the theory of convex bodies, we establish a representation theorem of fuzzy vectors throughsupport functions, in which a necessary and su ffi cient condition for a function to be the support function of a fuzzyvector is provided. As applications, symmetric and skew fuzzy vectors are postulated, based on which a Mareˇs core ofeach fuzzy vector is constructed through convex bodies and support functions, and it is shown that every fuzzy vectorover the n -dimensional Euclidean space has a unique Mareˇs core if, and only if, the dimension n = Keywords: fuzzy vector, convex body, support function, symmetric fuzzy vector, skew fuzzy vector, Mareˇs core,Mareˇs equivalence
1. Introduction
Since Zadeh introduced the concept of fuzzy sets [29] in the 1960s, fuzzy numbers , as a special kind of fuzzysubsets of the set R of real numbers, have received considerable attention both in the theory and the applications offuzzy sets [6, 7, 8, 9, 10, 11, 5, 4]. The notion of fuzzy number may be generalized to fuzzy vector (also n-dimensionalfuzzy number ) without obstruction, simply by replacing R with the n -dimensional Euclidean space R n in its definition,which has been widely studied as well [16, 24, 3, 31, 32, 28, 27, 19]. Convex geometry , as an independent branch of mathematics, has a much longer history that dates back to the turnof the 20th century [1], in which several contributions can be even traced back to the ancient works of Euclid andArchimedes. As a well-developed theory in the past decades, convex geometry has been applied to di ff erent areas ofgeometry, analysis and computer science [14, 18].It is well known that fuzzy vectors can be characterized through their level sets (see Theorem 2.2.3, originatedfrom [22, 17]). In particular, each level set of a fuzzy vector is a nonempty, compact and convex subset of R n , whichis precisely a convex body [13, 26] in the sense of convex geometry. It is then natural to consider the possibility ofexploiting the powerful arsenal of convex geometers in the realm of fuzzy vectors, and it is the motivation of this paper.Being tailored to the readership of fuzzy set theorists, the geometric machinery involved in this paper are presented inthe most accessible terms, so that hopefully, even a reader who is not familiar with the extensive apparatus of convexgeometry could easily follow up.Specifically, inspired by several key results from the theory of convex bodies, this paper is intended to representfuzzy vectors through support functions (Section 2) and, as applications, investigate Mareˇs cores of fuzzy vectors(Section 3). The backgrounds and our main results are illustrated as follows. ∗ Corresponding author.
Email addresses: [email protected] (Cheng-Yong Du), [email protected] (Lili Shen) .1. Representation of fuzzy vectors via support functions Support functions play an essential role in the study of fuzzy vectors. Explicitly, the support function [24, 3] of afuzzy vector u is given by h u : [0 , × S n − → R , h u ( α, x ) : = _ t ∈ u α h t , x i , where S n − is the unit sphere in R n , h− , −i refers to the standard Euclidean inner product of R n , and u α is the α -levelset of u . The following question arises naturally: Question 1.1.1.
Can we find a necessary and su ffi cient condition for a function h : [0 , × S n − → R to be the support function of a (unique) fuzzy vector?This question is partially answered by Zhang–Wu in [32], where several su ffi cient conditions are provided, thoughneither of them is necessary. In order to fully solve this question, several key results from the theory of convex bodiesare exhibited in Subsection 2.1: • There exists a hyperplane that supports a convex body at any of its boundary point (Theorem 2.1.3). • A function h : S n − → R is the support function of a (unique) convex body if, and only if, it is sublinear (Theorem 2.1.5).Without assuming any a-priori background by the reader on convex geometry, in Subsection 2.1 we develop allneeded ingredients from scratch for the self-containment of this paper. Then, based on the well-known characteriza-tion of fuzzy vectors through convex bodies (Theorem 2.2.3), a representation theorem of fuzzy vectors via supportfunctions is established (Theorem 2.2.4). Explicitly, it is shown in Theorem 2.2.4 that a function h : [0 , × S n − → R is the support function of a (unique) fuzzy vector if, and only if,(VS1) h ( α, − ) : S n − → R is sublinear for each α ∈ [0 , h ( − , x ) : [0 , → R is non-increasing, left-continuous on (0 ,
1] and right-continuous at 0 for each x ∈ S n − .Therefore, a perfect answer is provided for Question 1.1.1. In the recent works of Qiu–Lu–Zhang–Lan [25] and Chai–Zhang [2], a crucial property of fuzzy numbers regard-ing their
Mareˇs cores is revealed. Explicitly, a fuzzy number u is skew [2] if it cannot be written as the sum of a fuzzynumber and a non-trivial symmetric fuzzy number in the sense of Mareˇs [20]; that is, if u = v ⊕ w and w is symmetric, then w is constant at 0. A fuzzy number v is the Mareˇs core [21, 15] of a fuzzy number u if v isskew and u = v ⊕ w for some symmetric fuzzy number w . The following theorem combines the main results of [25]and [2]: Theorem 1.2.1.
Every fuzzy number has a unique Mareˇs core, so that every fuzzy number can be decomposed in aunique way as the sum of a skew fuzzy number, given by its Mareˇs core, and a symmetric fuzzy number.
2t is then natural to ask whether it is possible to establish the n -dimensional version of Theorem 1.2.1 for generalfuzzy vectors. Unfortunately, a negative answer will be given in Section 3 (Theorem 3.4.10).Based on the representation of fuzzy vectors through convex bodies and support functions in Section 2, Theorem3.1.1 describes the sum of fuzzy vectors defined by Zadeh’s extension principle through the Minkowski sum of convexbodies and the sum of their support functions, which is the cornerstone of the results of Section 3. Then, in Subsection3.2, the notion of symmetric fuzzy vector is defined as symmetric around the origin in accordance with the case of n = skew fuzzy vectors and Mareˇs cores of fuzzy vectorsin Subsection 3.4. However, for the purpose of studying their properties we have to be familiar with the inner parallelbodies of convex bodies, and this is the subject of Subsection 3.3, in which we characterize inner parallel bodiesthrough support functions, and prove that every convex body can be uniquely decomposed as the Minkowski sum ofan irreducible convex body and a closed ball centered at the origin (Theorem 3.3.4).The first part of Subsection 3.4 is devoted to the decomposition u = c ( u ) ⊕ s ( u ) (1.i)of each fuzzy vector u ∈ F n , where c ( u ) is a Mareˇs core of u , and s ( u ) is a symmetric fuzzy vector (Theorem 3.4.5).However, unlike Theorem 1.2.1 for the case of n =
1, Example 3.4.9 reveals that Equation (1.i) may not be the uniqueway of decomposing a fuzzy vector. Hence, we indeed obtain a negative answer to the possibility of establishing the n -dimensional version of Theorem 1.2.1, which is stated as Theorem 3.4.10.Finally, we investigate Mareˇs equivalent fuzzy vectors in Subsection 3.5. As we shall see, comparing with Mareˇsequivalent fuzzy numbers (see Corollary 3.5.6), the Mareˇs equivalence relation of fuzzy vectors may behave in quitedi ff erent ways. As Example 3.5.7 reveals, the smallest fuzzy vector k ( u ) of the Mareˇs equivalence class of a fuzzyvector u may not be a Mareˇs core of u , and di ff erent skew fuzzy vectors may be Mareˇs equivalent to each other.
2. Representation of fuzzy vectors via support functions
Throughout, let R n denote the n -dimensional Euclidean space. Following the terminologies of convex geometry[13, 26], by a convex body in R n we mean a nonempty, compact and convex subset of R n ; that is, A ⊆ R n is a convexbody if it is nonempty, closed, bounded and λ s + (1 − λ ) t ∈ A whenever s , t ∈ A and λ ∈ [0 , R n is denoted by C n .Let h− , −i denote the standard Euclidean inner product of R n . A hyperplane H in R n is usually denoted by H = { t ∈ R n | h t , x i = α } (2.i)for some x ∈ R n \ { o } and α ∈ R , where o is the origin of R n , and x is called a normal vector of H . Each hyperplane H given by (2.i) divides R n into two closed halfspacesH − : = { t ∈ R n | h t , x i ≤ α } and H + : = { t ∈ R n | h t , x i ≥ α } . (2.ii)Let A ⊆ R n be closed and convex. For every x ∈ R n , there exists a unique point p A ( x ) ∈ A such that || x − p A ( x ) || = d ( x , A ) : = ^ a ∈ A || x − a || , (2.iii)where || - || refers to the standard Euclidean norm on R n . Indeed, the existence of p A ( x ) is obvious by the closedness of A . For the uniqueness of p A ( x ), suppose that q A ( x ) ∈ A also satisfies (2.iii), but p A ( x ) , q A ( x ). Then p A ( x ) + q A ( x )2 ∈ A by the convexity of A , but (cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12) x − p A ( x ) + q A ( x )2 (cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12) = (cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12) x − p A ( x )2 + x − q A ( x )2 (cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12) < (cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12) x − p A ( x )2 (cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12) + (cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12) x − q A ( x )2 (cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12) = || x − p A ( x ) || = || x − q A ( x ) || ;3hat is, p A ( x ) + q A ( x )2 is strictly closer to x than p A ( x ) and q A ( x ), contradicting to the fact that p A ( x ) and q A ( x ) bothsatisfy Equation (2.iii). Thus we obtain a well-defined map p A : R n → A , called the metric projection of A . It is obvious that p A ( a ) = a for all a ∈ A . Lemma 2.1.1. (See [13, 26].) If A ⊆ R n is closed and convex, then the metric projection p A : R n → A is non-expansive in the sense that || p A ( x ) − p A ( y ) || ≤ || x − y || for all x , y ∈ R n .Proof. We only prove the case of x , y ∈ R n \ A , while the rest cases can be treated analogously. Since the conclusionholds trivially when p A ( x ) = p A ( y ), suppose that p A ( x ) , p A ( y ). In this case, the convexity of A guarantees that theline segment [ p A ( x ) , p A ( y )] ⊆ A . Considering the hyperplane H x : = { t ∈ R n | h t , p A ( x ) − p A ( y ) i = h p A ( x ) , p A ( x ) − p A ( y ) i} , it is clear that p A ( y ) ∈ H − x . We claim that x ∈ H + x . In fact, if x ∈ R n \ H + x , then h x − p A ( x ) , p A ( y ) − p A ( x ) i >
0, andconsequently the angle between x − p A ( x ) and p A ( y ) − p A ( x ) is acute. This means that the line segment [ p A ( x ) , p A ( y )]contains a point which is strictly closer to x than p A ( x ), contradicting to the definition of p A ( x ) (see Equation (2.iii)).Similarly, for the hyperplane H y : = { t ∈ R n | h t , p A ( x ) − p A ( y ) i = h p A ( y ) , p A ( x ) − p A ( y ) i} we may deduce that p A ( x ) ∈ H + y and y ∈ H − y .Since p A ( x ) ∈ H x , p A ( y ) ∈ H y and p A ( x ) − p A ( y ) is the normal vector of both the hyperplanes H x and H y , thedistance between H x and H y is precisely || p A ( x ) − p A ( y ) || . From x ∈ H + x , p A ( y ) ∈ H − x , p A ( x ) ∈ H + y and y ∈ H − y we seethat the distance between x and y is no less than the distance between H x and H y ; that is, || p A ( x ) − p A ( y ) || ≤ || x − y || .Let A ⊆ R n be a subset, and let H ⊆ R n be a hyperplane. We say that H supports A at t if t ∈ A ∩ H and either A ⊆ H + or A ⊆ H − ; in this case, t necessarily lies in the boundary bd A of A , and H is called a support hyperplane of A . If a hyperplane H given by (2.i) supports A and A ⊆ H − , then x is called an exterior normal vector of H . Lemma 2.1.2.
If A ⊆ R n is closed and convex, then for each x ∈ R n \ A, there exists a hyperplane H that supports Aat p A ( x ) , with x − p A ( x ) being its exterior normal vector.Proof. Let H = { t ∈ R n | h t , x − p A ( x ) i = h p A ( x ) , x − p A ( x ) i} . Then p A ( x ) ∈ A ∩ H , and we claim that A ⊆ H − . In fact, if there exists z ∈ A ∩ ( R n \ H − ), then the line segment[ p A ( x ) , z ] ⊆ A by the convexity of A . Note that z ∈ R n \ H − means that h z − p A ( x ) , x − p A ( x ) i >
0, and consequentlythe angle between z − p A ( x ) and x − p A ( x ) is acute. Thus, the line segment [ p A ( x ) , z ] must contain a point which isstrictly closer to x than p A ( x ), contradicting to the definition of p A ( x ).The following theorem is well known in convex geometry, and we present a proof here for the sake of self-containment: Theorem 2.1.3. (See [13, 26].) If A ⊆ R n is closed and convex, then for each t ∈ bd A, there exists a (not necessarilyunique) hyperplane H that supports A at t .Proof. Let { t k } ⊆ R n \ A be a sequence that converges to t ∈ bd A , which induces a sequence { p A ( t k ) } ⊆ bd A . Foreach positive integer k , by Lemma 2.1.2 we may find a hyperplane H k = { t ∈ R n | h t , t k − p A ( t k ) i = h p A ( t k ) , t k − p A ( t k ) i} A at p A ( t k ), with t k − p A ( t k ) being its exterior normal vector. Let x k : = t k − p A ( t k ) || t k − p A ( t k ) || . Then x k belongs to S n − , the unit sphere in R n . By the compactness of S n − , the sequence { x k } has a convergentsubsequence, and without loss generality we may suppose that { x k } itself converges to x ∈ S n − . Note that || p A ( t k ) − t || = || p A ( t k ) − p A ( t ) || ≤ || t k − t || by Lemma 2.1.1, which necessarily forces lim k →∞ p A ( t k ) = t as we already have lim k →∞ t k = t . We claim that the hyper-plane H = { t ∈ R n | h t , x i = h t , x i} supports A at t , with x being its exterior normal vector. To see this, note that t ∈ A ∩ H is obvious, and for everypositive integer k we have A ⊆ H − k , which implies that h a , t k − p A ( t k ) i ≤ h p A ( t k ) , t k − p A ( t k ) i , i.e., h a , x k i ≤ h p A ( t k ) , x k i for all a ∈ A . By letting k → ∞ in the above inequality we immediately obtain that h a , x i ≤ h t , x i for all a ∈ A , andconsequently A ⊆ H − , which completes the proof.Recall that the support function [13, 26] of a convex body A ∈ C n is given by h A : S n − → R , h A ( x ) : = _ a ∈ A h a , x i , (2.iv)where S n − is the unit sphere in R n . Obviously, h A is bounded on S n − ; indeed, | h A ( x ) | ≤ _ a ∈ A || a || for all x ∈ S n − . Remark 2.1.4.
The domain of the support function of a convex body A ∈ C n is defined as R n in [13, 26]; that is, h A : R n → R , h A ( x ) : = _ a ∈ A h a , x i . (2.v)In fact, the function h A given by (2.v) is completely determined by its values on S n − , since it always holds that h A ( o ) = h A ( x ) = || x || · h A x || x || ! for all x ∈ R n \ { o } . Therefore, it does no harm to restrict the domain of h A to S n − .Conversely, to each function h : S n − → R we may associate a subset A h : = { t ∈ R n | ∀ x ∈ S n − : h t , x i ≤ h ( x ) } (2.vi)of R n . Convex bodies can be fully characterized through support functions as follows, which is a modification of [13,Theorem 4.3] and [26, Theorem 1.7.1]: Theorem 2.1.5.
A function h : S n − → R is the support function of a (unique) convex body A h if, and only if, h is sublinear in the sense that h ( x ) + h ( − x ) ≥ and h λ x + (1 − λ ) y || λ x + (1 − λ ) y || ! ≤ λ h ( x ) + (1 − λ ) h ( y ) || λ x + (1 − λ ) y || for all x , y ∈ S n − , λ ∈ [0 , with λ x + (1 − λ ) y , . f : R n → R is convex if f ( λ x + (1 − λ ) y ) ≤ λ f ( x ) + (1 − λ ) f ( y )for all x , y ∈ R n and λ ∈ [0 , Lemma 2.1.6. (See [13, 26].) A function f : R n → R is convex if, and only if, its epigraphepi f : = { ( x , α ) ∈ R n × R | f ( x ) ≤ α } ⊆ R n + is convex. In this case, f is necessarily continuous on R n and, consequently, epi f is closed.Proof. Step 1. f is convex if, and only if, epi f is convex. If f is convex, for any ( x , α ) , ( y , β ) ∈ epi f and λ ∈ [0 , f ( λ x + (1 − λ ) y ) ≤ λ f ( x ) + (1 − λ ) f ( y ) ≤ λα + (1 − λ ) β ;that is, λ ( x , α ) + (1 − λ )( y , β ) = ( λ x + (1 − λ ) y , λα + (1 − λ ) β ) ∈ epi f . Thus epi f is convex.Conversely, if epi f is convex, for any x , y ∈ R n and λ ∈ [0 ,
1] we have( λ x + (1 − λ ) y , λ f ( x ) + (1 − λ ) f ( y )) = λ ( x , f ( x )) + (1 − λ )( y , f ( y )) ∈ epi f because ( x , f ( x )) , ( y , f ( y )) ∈ epi f ; that is, f ( λ x + (1 − λ ) y ) ≤ λ f ( x ) + (1 − λ ) f ( y ), showing that f is convex. Step 2. If f : R n → R is convex, then f is continuous on R n . To this end, for any x ∈ R n we choose a simplex S = n + X i = λ i x i (cid:12)(cid:12)(cid:12)(cid:12) n + X i = λ i = λ i ≥ i = , . . . , n + with vertices x , . . . , x n + ∈ R n , such that there exists an open ball B ( x , ρ ) ⊆ S ( ρ > f is clearly boundedon S since, by Jensen’s inequality (cf. [26, Remark 1.5.1]), f ( x ) = f n + X i = λ i x i ≤ n + X i = λ i f ( x i ) ≤ c : = max { f ( x ) , . . . , f ( x n + ) } for all x = n + X i = λ i x i ∈ S . Now, for any y = x + λ t ∈ B ( x , ρ ) ( λ ∈ [0 , || t || = ρ ), the convexity of f implies that f ( y ) = f ( x + λ t ) = f ((1 − λ ) x + λ ( x + t )) ≤ (1 − λ ) f ( x ) + λ f ( x + t ) , and consequently f ( y ) − f ( x ) ≤ λ ( f ( x + t ) − f ( x )) ≤ λ ( c − f ( x )) because x + t ∈ S . Similarly, f ( x ) = f + λ y + λ + λ ( x − t ) ! ≤ + λ f ( y ) + λ + λ f ( x − t ) , and consequently f ( x ) − f ( y ) ≤ λ ( f ( x − t ) − f ( x )) ≤ λ ( c − f ( x )). It follows that | f ( y ) − f ( x ) | ≤ λ ( c − f ( x )) = c − f ( x ) ρ || y − x || for all y ∈ B ( x , ρ ), which immediately implies the continuity of f at x . Step 3. If f : R n → R is continuous, then epi f is closed. This is easy since, for any sequence { ( x k , α k ) } ⊆ epi f that converges to ( x , α ), f ( x ) ≤ α becomes an immediate consequence of f ( x k ) ≤ α k for all positive integers k inconjunction with the continuity of f , which means precisely that ( x , α ) ∈ epi f . This completes the proof.6 roof of Theorem 2.1.5. Necessity.
Let A ∈ C n be a convex body. Then it is clear that h A ( x ) + h A ( − x ) = _ a ∈ A h a , x i + _ a ∈ A h a , − x i = _ a ∈ A h a , x i − ^ a ∈ A h a , x i ≥ h A λ x + (1 − λ ) y || λ x + (1 − λ ) y || ! = _ a ∈ A * a , λ x + (1 − λ ) y || λ x + (1 − λ ) y || + ≤ λ _ a ∈ A h a , x i + (1 − λ ) _ a ∈ A h a , y i|| λ x + (1 − λ ) y || = λ h A ( x ) + (1 − λ ) h A ( y ) || λ x + (1 − λ ) y || for all x , y ∈ S n − , λ ∈ [0 ,
1] with λ x + (1 − λ ) y , Su ffi ciency. Let h : S n − → R be a sublinear function. Since A h = { t ∈ R n | ∀ x ∈ S n − : h t , x i ≤ h ( x ) } = \ x ∈ S n − { t ∈ R n | h t , x i ≤ h ( x ) } (2.vii)is the intersection of closed halfspaces which are necessarily convex, it is clearly closed and convex. Moreover,considering the standard basis { e , . . . , e n } of R n we see that each coordinate of t ∈ A h is bounded by α : = max {| h ( e i ) | , | h ( − e i ) | | ≤ i ≤ n } , and thus || t || ≤ √ n α for all t ∈ A h ; that is, A h is bounded. Next, we show that A h , ∅ and h A h = h , so that A h is aconvex body and h is the support function of A h .Firstly, A h , ∅ . To see this, let us extend h : S n − → R to ˜ h : R n → R as elaborated in Remark 2.1.4, i.e.,˜ h ( x ) = x = o , || x || · h x || x || ! elsefor all x ∈ R n . Then ˜ h ( λ x ) = λ ˜ h ( x ) and ˜ h ( x + y ) ≤ ˜ h ( x ) + ˜ h ( y ) (2.viii)for all x , y ∈ R n and λ ≥
0, where the first equality follows obviously from the definition of ˜ h , while the secondinequality holds because ˜ h ( o ) = ≤ || x || · h x || x || ! + h − x || x || !! = ˜ h ( x ) + ˜ h ( − x )for all x ∈ R n \ { o } , and˜ h ( x + y ) = || x + y || · h x + y || x + y || ! = || x + y || · h || x |||| x || + || y || · x || x || + || y |||| x || + || y || · y || y || (cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12) || x |||| x || + || y || · x || x || + || y |||| x || + || y || · y || y || (cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12) ≤ || x + y || · || x |||| x || + || y || · h x || x || ! + || y |||| x || + || y || · h y || y || !(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12) || x |||| x || + || y || · x || x || + || y |||| x || + || y || · y || y || (cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12) = ˜ h ( x ) + ˜ h ( y )for all x , y ∈ R n with x + y , o .Since ˜ h : R n → R is clearly convex by (2.viii), Lemma 2.1.6 tells us thatepi ˜ h = { ( x , α ) ∈ R n × R | ˜ h ( x ) ≤ α } ⊆ R n + is closed and convex. By Theorem 2.1.3, there exists a hyperplane H y that supports epi ˜ h at any ( y , ˜ h ( y )) ∈ bd epi ˜ h with y , o . We claim that H y passes through the origin ( o ,
0) of R n × R = R n + , so that H y also supports epi ˜ h at ( o , o , < H y , then the straight line passing through ( o ,
0) and ( y , ˜ h ( y )) intersects the hyperplane H y in exactlyone point, i.e., ( y , ˜ h ( y )). It follows that ( o ,
0) and (2 y , h ( y )) are on di ff erent sides of H y . However, it is clear that ( o , y , h ( y )) are both in epi ˜ h , which contradicts to the fact that H y is a support hyperplane of epi ˜ h .Note that the definition of epi ˜ h indicates that the exterior normal vector of H y can be written as the form ( t y , − H y “points below R n ” in R n × R . Since H y also supportsepi ˜ h at ( o , H y = { ( x , α ) ∈ R n × R | h ( x , α ) , ( t y , − i = h ( o , , ( t y , − i = } . (2.ix)Since epi ˜ h ⊆ H − y , for any x ∈ R n , considering ( x , ˜ h ( x )) ∈ epi ˜ h we have h t y , x i − ˜ h ( x ) = h ( x , ˜ h ( x )) , ( t y , − i ≤ , i.e., h t y , x i ≤ ˜ h ( x ). In particular, this means that h t y , x i ≤ h ( x ) for all x ∈ S n − , and consequently t y ∈ A h (cf. (2.vii)),showing that A h , ∅ .Secondly, h A h = h . On one hand, h A h ( x ) = _ a ∈ A h h a , x i ≤ h ( x )for all x ∈ S n − is an immediate consequence of (2.iv) and (2.vii). On the other hand, for any y ∈ S n − , from( y , h ( y )) ∈ H y we have h t y , y i = h ( y ) by (2.ix). Since t y ∈ A h , it follows that h A h ( y ) = _ a ∈ A h h a , y i ≥ h t y , y i = h ( y )for all y ∈ S n − . Hence h A h = h .Finally, for the uniqueness of A h , it remains to show that A h A = A for any convex body A ∈ C n . Note that A ⊆ A h A follows easily from (2.iv) and (2.vii). For the reverse inclusion, we proceed by contradiction. Suppose that t ∈ A h A but t < A . Since A is closed and convex, by Lemma 2.1.2 we may find a hyperplane H that supports A at p A ( t ), with x : = t − p A ( t ) || t − p A ( t ) || ∈ S n − being its exterior normal vector, i.e., H = { t ∈ R n | h t , x i = h p A ( t ) , x i} . Now, it follows from A ⊆ H − that h a , x i ≤ h p A ( t ) , x i for all a ∈ A , but t ∈ R n \ H − forces h t , x i > h p A ( t ) , x i ;that is, h t , x i > h p A ( t ) , x i ≥ _ a ∈ A h a , x i = h A ( x ) , contradicting to t ∈ A h A . This completes the proof.As an immediate consequence of Theorem 2.1.5, the following characterization of elements of convex bodies willbe useful later (cf. [31, Theorem 2.2]): Proposition 2.1.7.
Let A ∈ C n be a convex body. Then t ∈ A if, and only if, h t , x i ≤ h A ( x ) for all x ∈ S n − . In particular, constant support functions correspond to closed balls centered at the origin:
Proposition 2.1.8.
A convex body A ∈ C n is a closed ball centered at the origin if, and only if, its support functionh A : S n − → R is a (nonnegative) constant function. In this case, h A is necessarily constant at the radius of A. roof. Suppose that A is a closed ball of radius λ centered at the origin. Then h A ( x ) = _ a ∈ A h a , x i = h λ x , x i = λ (2.x)for all x ∈ S n − . Conversely, if h A ( x ) = λ for all x ∈ S n − , then it follows from Proposition 2.1.7 that t ∈ A ⇐⇒ ∀ x ∈ S n − : h t , x i ≤ λ ⇐⇒ || t || = * t , t || t || + ≤ λ, which also guarantees the nonnegativity of λ .Recall that the Minkowski sum [26] of convex bodies B , C ∈ C n is given by B + C : = { b + c | b ∈ B , c ∈ C } . Note that a direct computation h B + C ( x ) = _ b ∈ B , c ∈ C h b + c , x i = _ b ∈ B , c ∈ C ( h b , x i + h c , x i ) = _ b ∈ B h b , x i + _ c ∈ C h c , x i = h B ( x ) + h C ( x )for any x ∈ S n − shows that h B + C = h B + h C (cf. [26, Theorem 1.7.5]), in combination with Theorem 2.1.5 we obtain: Proposition 2.1.9.
For convex bodies A , B , C ∈ C n , A = B + C if, and only if, h A = h B + h C .2.2. Fuzzy vectors via convex bodies and support functions Following the terminology of [19], by a fuzzy vector we mean a fuzzy subset of R n , i.e., a function u : R n → [0 , , subject to the following requirements:(V1) u is regular , i.e., there exists t ∈ R n with u ( t ) = u is compactly supported , i.e., the closure of { t ∈ R n | u ( t ) > } is compact;(V3) u is quasi-concave , i.e., u ( s ) ∧ u ( t ) ≤ u ( λ s + (1 − λ ) t ) for all s , t ∈ R n and λ ∈ [0 , u is upper semi-continuous , i.e., { t ∈ R n | u ( t ) ≥ α } is closed for all α ∈ [0 , n is denoted by F n , and a canonical embedding of R n into F n assigns toeach a ∈ R n a “crisp” fuzzy vector ˜ a : R n → [0 , , ˜ a ( t ) : = t = a , . Remark 2.2.1.
The conditions (V1)–(V4), first appeared in [16] and [24], are originated from the definition of fuzzynumbers , i.e., fuzzy vectors of dimension 1 (see [6, 8, 9, 10, 12]); so, fuzzy vectors are also called n -dimensional fuzzynumbers (see [31, 32, 28, 27]). It should be reminded that n-dimensional fuzzy vectors defined in [27] are di ff erentfrom our fuzzy vectors here.For each u ∈ F n and α ∈ [0 , α -level sets of u are defined as u α : = { t ∈ R n | u ( t ) ≥ α } if α ∈ (0 , , S α ∈ (0 , u α = { t ∈ R n | u ( t ) > } if α = . It is easy to see that u ( t ) = _ t ∈ u α α (2.xi)for each u ∈ F n and t ∈ R n . 9 emark 2.2.2. Since α ranges in the closed interval [0 , , t < u α for all α ∈ [0 , { α | t ∈ u α } = ∅ , then u ( t ) = , , Theorem 2.2.3. (See [22, 17].) Let { A α | α ∈ [0 , } be a family of subsets of R n . Then there exists a (unique) fuzzyvector u : R n → [0 , , u ( t ) = _ t ∈ A α α such that u α = A α for all α ∈ [0 , if, and only if, (L1) A α is a convex body in R n for each α ∈ [0 , ; (L2) A α ⊇ A β whenever ≤ α < β ≤ ; (L3) A α = T k ≥ A α k for each increasing sequence { α k } ⊆ [0 , that converges to α > ; (L4) A = S α ∈ (0 , A α . Since all the level sets of a fuzzy vector are convex bodies, it makes sense to define the support function [24, 3] of u ∈ F n as h u : [0 , × S n − → R , h u ( α, x ) : = _ t ∈ u α h t , x i ;that is, h u ( α, − ) : = h u α is the support function of the convex body u α for each α ∈ [0 , h u is boundedon [0 , × S n − , because h u ( α, x ) ≤ _ t ∈ u h t , x i = h u ( x )for all α ∈ [0 , x ∈ S n − , and h u is bounded on S n − .With Theorems 2.1.5 and 2.2.3 we may describe fuzzy vectors through support functions as the following repre-sentation theorem reveals, which is the main result of this paper: Theorem 2.2.4.
A function h : [0 , × S n − → R is the support function of a (unique) fuzzy vectoru : R n → [0 , given by u ( t ) = _ t ∈ A h ( α, − ) α = _ { α | ∀ x ∈ S n − : h t , x i ≤ h ( α, x ) } if, and only if, (VS1) h ( α, − ) : S n − → R is sublinear for each α ∈ [0 , , i.e.,h ( α, x ) + h ( α, − x ) ≥ and h α, λ x + (1 − λ ) y || λ x + (1 − λ ) y || ! ≤ λ h ( α, x ) + (1 − λ ) h ( α, y ) || λ x + (1 − λ ) y || for all α ∈ [0 , , x , y ∈ S n − , λ ∈ [0 , with λ x + (1 − λ ) y , , h ( − , x ) : [0 , → R is non-increasing, left-continuous on (0 , and right-continuous at for each x ∈ S n − .Proof. Necessity.
Let u ∈ F n be a fuzzy vector. Then h u clearly satisfies (VS1) by Theorem 2.1.5. For (VS2), let usfix x ∈ S n − . Then h u ( − , x ) : [0 , → R is non-increasing because of (L2).To see that h u ( − , x ) is left-continuous at each α ∈ (0 , { α k } ⊆ (0 ,
1] be an increasing sequence that convergesto α . Then u α = T k ≥ u α k by (L3). For each ǫ >
0, we claim that there exists a positive integer k such that for all t ∈ u α k , there exists r t ∈ u α with || t − r t || < ǫ. (2.xii)Indeed, suppose that we find an ǫ > k , there exists t k ∈ u α k such that d ( t k , u α ) : = ^ r ∈ u α || t k − r || ≥ ǫ . Then the sequence { t k } is contained in the compact set u α , and thus it has a convergent subsequence. Without loss ofgenerality we may suppose that lim k →∞ t k = t . Then t ∈ u α k for all k ≥ { t m | m ≥ k } ⊆ u α k , and consequently t ∈ T k ≥ u α k = u α . But the construction of t k forces d ( t , u α ) ≥ ǫ , which is a contradiction.Note that (2.xii) actually means that u α k ⊆ u α + B ǫ , where B ǫ refers to the closed ball of radius ǫ centered at the origin, since each t ∈ u α k may be written as t = r t + ( t − r t )with r t ∈ u α and t − r t ∈ B ǫ . Then it follows from Propositions 2.1.8 and 2.1.9 that h u ( α k , x ) = h u α k ( x ) ≤ h u α ( x ) + h B ǫ ( x ) = h u ( α , x ) + ǫ. Hence, together with the monotonicity of h u ( − , x ) : [0 , → R we conclude that lim k →∞ h u ( α k , x ) = h u ( α , x ), whichproves the left-continuity of h u ( − , x ) at α .To see that h u ( − , x ) is right-continuous at 0, let ǫ >
0. The compactness of u allows us to find q , . . . , q k ∈ u such that u is covered by finitely many open balls B (cid:18) q , ǫ (cid:19) , . . . , B (cid:18) q k , ǫ (cid:19) centered at q , . . . , q k , respectively, with radii ǫ t ∈ u = S α ∈ (0 , u α , there exists α t ∈ (0 ,
1] and s t ∈ u α t such that || t − s t || < ǫ α q : = min { α q , . . . , α q k } > , and let B (cid:18) q t , ǫ (cid:19) ( q t ∈ { q , . . . , q k } ) be the open ball containing t . Then s q t ∈ u α qt ⊆ u α q , and || t − s q t || ≤ || t − q t || + || q t − s q t || < ǫ + ǫ = ǫ. It follows that u ⊆ u α q + B ǫ , because t = s q t + ( t − s q t ) for all t ∈ u , where s q t ∈ u α q and t − s q t ∈ B ǫ . By Propositions 2.1.8 and 2.1.9 this meansthat h u (0 , x ) = h u ( x ) ≤ h u α q ( x ) + h B ǫ ( x ) = h u ( α q , x ) + ǫ. Hence, together with the monotonicity of h u ( − , x ) : [0 , → R we conclude that lim α → + h u ( α, x ) = h u (0 , x ), whichproves the right-continuity of h u ( − , x ) at 0. 11 u ffi ciency. It su ffi ces to show that { A h ( α, − ) | α ∈ [0 , } satisfies the conditions (L1)–(L4).Firstly, (L1) is a direct consequence of Theorem 2.1.5.Secondly, (L2) holds since h ( − , x ) is non-increasing for all x ∈ S n − .Thirdly, for (L3), let { α k } ⊆ (0 ,
1] be an increasing sequence that converges to α ∈ (0 , \ k ≥ A h ( α k , − ) ⊆ A h ( α , − ) , since the reverse inclusion is trivial by (L2). Suppose that t ∈ A h ( α k , − ) for all k ≥
1. Then, by Proposition 2.1.7, h t , x i ≤ h ( α k , x ) for all x ∈ S n − . Thus the left-continuity of h ( − , x ) at α implies that h t , x i ≤ lim k →∞ h ( α k , x ) = h (cid:18) lim k →∞ α k , x (cid:19) = h ( α , x ) , and consequently t ∈ A h ( α , − ) .Finally, we prove (L4) by showing that A h (0 , − ) ⊆ A : = [ α ∈ (0 , A h ( α, − ) as the reverse inclusion is trivial by (L2). We proceed by contradiction. Suppose that t ∈ A h (0 , − ) but t < A . Notethat A is also a convex body, since A ⊆ A h (0 , − ) and A is the closure of the union of a family of convex bodies linearlyordered by inclusion. Thus, by Lemma 2.1.2 we may find a hyperplane H that supports A at p A ( t ), with x : = t − p A ( t ) || t − p A ( t ) || ∈ S n − as its exterior normal vector, i.e., H = { t ∈ R n | h t , x i = h p A ( t ) , x i} , which necessarily satisfies t ∈ R n \ H − and A ⊆ H − . It follows that h ( α, x ) = _ t ∈ A h ( α, − ) h t , x i (Theorem 2.1.5) ≤ h p A ( t ) , x i ( A h ( α, − ) ⊆ A ⊆ H − ) = h t , x i − h t − p A ( t ) , x i = h t , x i − || t − p A ( t ) || x = t − p A ( t ) || t − p A ( t ) || ! ≤ h (0 , x ) − || t − p A ( t ) || ( t ∈ A h (0 , − ) )for all α ∈ (0 , h ( − , x ) at 0. The proof is thus completed.
3. Mareˇs cores of fuzzy vectors
With the results of Section 2 we are now able to characterize the addition ⊕ of fuzzy vectors through their levelsets and support functions. Explicitly, the sum u ⊕ v ∈ F n of fuzzy vectors u , v ∈ F n is defined by Zadeh’s extension principle (cf. [30, 6]), i.e.,( u ⊕ v )( t ) : = _ r + s = t u ( r ) ∧ v ( s )for all t ∈ R n . 12 heorem 3.1.1. For fuzzy vectors u , v , w ∈ F n , the following statements are equivalent: (i) u = v ⊕ w. (ii) u α = v α + w α for all α ∈ [0 , . (iii) h u = h v + h w .Proof. (ii) ⇐⇒ (iii) is an immediate consequence of Proposition 2.1.9. For (i) ⇐⇒ (ii), by Theorem 2.2.3 it su ffi cesto observe that ( v ⊕ w ) α = v α + w α (3.i)for all α ∈ [0 , v α + w α ⊆ ( v ⊕ w ) α . Let r ∈ v α , s ∈ w α . Then v ( r ) ≥ α and w ( s ) ≥ α , and consequently ( v ⊕ w )( r + s ) ≥ v ( r ) ∧ w ( s ) ≥ α, showing that r + s ∈ ( v ⊕ w ) α . On the other hand, in order to verify the reverse inclusion, let t ∈ ( v ⊕ w ) α . Then( v ⊕ w )( t ) = _ r + s = t v ( r ) ∧ w ( s ) ≥ α. Consequently, there exists a sequence { r k } in R n such that v ( r k ) ∧ w ( t − r k ) ≥ − k ! α ≥ α k . In particular, this means that { r k } ⊆ u α . Since u α is compact, { r k } has a convergentsubsequence, and without loss of generality we may assume that lim k →∞ r k = r . Note that for any positive integers k , l with l ≥ k , v ( r l ) ∧ w ( t − r l ) ≥ − l ! α ≥ − k ! α. Letting l → ∞ in the above inequality, the upper semi-continuity of v , w (see (V4)) then implies that v ( r ) ∧ w ( t − r ) ≥ − k ! α. Since k is arbitrary, the above inequality forces v ( r ) ∧ w ( t − r ) ≥ α . It follows that r ∈ v α and t − r ∈ w α , andtherefore t = r + ( t − r ) ∈ v α + w α . Let O ( n ) denote the orthogonal group of dimension n , i.e., the group of n × n orthogonal matrices. Definition 3.2.1.
A fuzzy vector u ∈ F n is symmetric (around the origin) if it is O ( n )-invariant; that is, if u ( t ) = u ( Qt )for all t ∈ R n and Q ∈ O ( n ).We denote by F n s the set of all symmetric fuzzy vectors of dimension n . Remark 3.2.2.
In the case of n =
1, since O (1) = {− , } , u ∈ F is symmetric if u ( t ) = u ( − t ) for all t ∈ R ; that is, u is a symmetric fuzzy number in the sense of Mareˇs [20, 21]. Hence, the symmetry of fuzzy numbers is a special caseof Definition 3.2.1. In fact, as indicated by [25, Remark 2.1], a symmetric fuzzy number actually refers to a fuzzynumber that is symmetric around zero . 13 heorem 3.2.3. For each fuzzy vector u ∈ F n , the following statements are equivalent: (i) u is symmetric. (ii) For each α ∈ [0 , , u α is invariant under the action of O ( n ) ; that is, Qt ∈ u α for all t ∈ u α and Q ∈ O ( n ) . (iii) For each α ∈ [0 , , u α is a closed ball centered at the origin. (iv) For each α ∈ [0 , , h u ( α, − ) : S n − → R is a (nonnegative) constant function.Proof. Since (iii) ⇐⇒ (iv) is an immediate consequence of Proposition 2.1.8, it remains to prove that (i) = ⇒ (iv) and(iii) = ⇒ (ii) = ⇒ (i).(i) = ⇒ (iv): Let α ∈ (0 ,
1] and x ∈ S n − . For each Q ∈ O ( n ) and t ∈ u α , the O ( n )-invariance of u implies that Q − t ∈ u α since u ( Q − t ) = u ( t ) ≥ α , and consequently h t , Qx i = h Q − t , x i ≤ h u ( α, x ) . Thus h u ( α, Qx ) = _ t ∈ u α h t , Qx i ≤ h u ( α, x ) . Since Q is arbitrary, it also holds that h u ( α, Q − x ) ≤ h u ( α, x ), and consequently h u ( α, x ) ≤ h u ( α, Qx ). Hence h u ( α, Qx ) = h u ( α, x )for all Q ∈ O ( n ). Note that the function O ( n ) → S n − , Q Qx is surjective, and thus h u ( α, x ) = h u ( α, y )for all x , y ∈ S n − ; that is, h u ( α, − ) : S n − → R is constant.In this case, in order to see that the value of h ( α, − ) is nonnegative, just note that for any t ∈ u α with t , o , from t || t || ∈ S n − we deduce that h u ( α, − ) = h u α, t || t || ! ≥ * t , t || t || + = || t || ≥ . (iii) = ⇒ (ii): Let α ∈ [0 , u α is a closed ball of radius λ α centered at the origin, then Qt ∈ u α whenever t ∈ u α ,since || t || ≤ λ α obviously implies that || Qt || ≤ λ α .(ii) = ⇒ (i): Let t ∈ R n and Q ∈ O ( n ). If u ( t ) >
0, then t ∈ u u ( t ) , and consequently Qt ∈ u u ( t ) , i.e., u ( Qt ) ≥ u ( t ). As Q is arbitrary, from u ( Q − t ) ≥ u ( t ) we immediately deduce that u ( t ) ≥ u ( Qt ), and thus u ( t ) = u ( Qt ).If u ( t ) =
0, then t < u α for all α ∈ (0 , Qt < u α for all α ∈ (0 , u ( Qt ) = h u : [0 , × S n − → R of a symmetric fuzzy vector u ∈ F n s actually reduces to a single-variable function h u : [0 , → R , and conversely: Corollary 3.2.4.
A function h : [0 , → R is the support function of a symmetric fuzzy vector u ∈ F n s if, and only if,h is nonnegative, non-increasing, left-continuous on (0 , and right-continuous at .Proof. Follows immediately from Theorems 2.2.4 and 3.2.3.As a direct application of Proposition 2.1.8 and Theorem 3.2.3, let us point out the following easy but useful facts:
Corollary 3.2.5.
Let A , B ∈ C n be convex bodies and u , v ∈ F n be fuzzy vectors. If A and B are both closed balls centered at the origin, then so is A + B, and λ A + B = λ A + λ B , where λ A , λ B and λ A + B are the radii of A, B and A + B, respectively. (ii)
If u and v are both symmetric, then so is u ⊕ v.Proof. For (i), just note that the support function of A + B satisfies h A + B = h A + h B by Proposition 2.1.9, and thus h A + B is a (nonnegative) constant function since so are h A and h B . The conclusion thenfollows from Proposition 2.1.8.For (ii), since u and v are both symmetric, for each α ∈ [0 , h u ( α, − ) and h v ( α, − )are both constant on S n − which, in conjunction with Theorem 3.1.1, implies that h u ⊕ v ( α, − ) = h u ( α, − ) + h v ( α, − )is constant on S n − ; that is, u ⊕ v is also symmetric. Let B λ denote the closed ball of radius λ ≥ inner parallel body (see[13, 26]) of a convex body A ∈ C n at distance λ is given by A − λ = { t ∈ R n | t + B λ ⊆ A } , which is also a convex body as long as A − λ , ∅ . Lemma 3.3.1.
For convex bodies A , B ∈ C n and λ ≥ ,A = B + B λ ⇐⇒ h A = h B + λ = ⇒ B = A − λ . Proof.
The equivalence of A = B + B λ and h A = h B + λ follows immediately from Propositions 2.1.8 and 2.1.9. Inthis case, from A = B + B λ and the definition of A − λ we soon see that B ⊆ A − λ . For the reverse inclusion, suppose that a ∈ A − λ . For each x ∈ S n − , note that λ x ∈ B λ , and consequently a + λ x ∈ A . (3.ii)It follows that h a , x i + λ = h a , x i + h λ x , x i = h a + λ x , x i ≤ h A ( x ) , where the last inequality is obtained by applying Proposition 2.1.7 to (3.ii). Therefore, h a , x i ≤ h A ( x ) − λ = h B ( x )for all x ∈ S n − , and thus Proposition 2.1.7 guarantees that a ∈ B .In general, the last implication of Lemma 3.3.1 is proper; that is, B = A − λ does not imply A = B + B λ . For example,let A and B be the hypercubes of side lengths 4 and 3, respectively, both centered at the origin. Then B = A − , but A ) B + B .We say that a nonempty inner parallel body A − λ of A ∈ C n is regular if A = A − λ + B λ , or equivalently (see Lemma 3.3.1), if h A = h A − λ + λ. For each convex body A ∈ C n , we write Λ A : = { λ ≥ | A − λ is a regular inner parallel body of A } . roposition 3.3.2. Λ A is a closed interval, given by Λ A = [0 , λ A ] , where λ A : = W Λ A .Proof. Step 1. A − λ A is a regular inner parallel body of A .In order to obtain A − λ A + B λ A = A , it su ffi ces to show that every t ∈ A lies in A − λ A + B λ A . Since A = A − λ + B λ forall λ ∈ Λ A , for an increasing sequence { λ k } ⊆ Λ A that converges to λ A we may find a k ∈ A − λ k and b k ∈ B λ k with t = a k + b k for all positive integers k . Note that both the sequences { a k } and { b k } are bounded, and thus they have convergentsubsequences. Without loss of generality we assume that lim k →∞ a k = a and lim k →∞ b k = b . Then it is clear that t = a + b and b ∈ B λ A , and it remains to prove that a ∈ A − λ A . To this end, we need to show that a + y ∈ A for all y ∈ B λ A .Indeed, let { y k } ⊆ B λ A be a sequence with lim k →∞ y k = y and y k ∈ B λ k for all positive integers k . Then from a k ∈ A − λ k wededuce that a k + y k ∈ A , and consequently a + y ∈ A , as desired. Step 2. If λ ∈ Λ A and 0 ≤ λ ′ < λ , then λ ′ ∈ Λ A .In order to obtain A − λ ′ + B λ ′ = A , it su ffi ces to show that every t ∈ A lies in A − λ ′ + B λ ′ . Since A = A − λ + B λ , wemay find a ∈ A − λ and b ∈ B λ with t = a + b . Then it is clear that t = " a + − λ ′ λ ! b + λ ′ λ b and λ ′ λ b ∈ B λ ′ , and it remains to prove that a + − λ ′ λ ! b ∈ A − λ ′ . To this end, we need to show that a + − λ ′ λ ! b + y ′ ∈ A for all y ′ ∈ B λ ′ . Indeed, − λ ′ λ ! b + y ′ ∈ B λ since (cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12) − λ ′ λ ! b + y ′ (cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12)(cid:12) ≤ − λ ′ λ ! || b || + || y ′ || ≤ − λ ′ λ ! λ + λ ′ = λ, and together with a ∈ A − λ we deduce that a + − λ ′ λ ! b + y ′ ∈ A , which completes the proof.As an immediate consequence of Proposition 3.3.2, we have the following: Corollary 3.3.3.
For each nonempty subset Λ ⊆ Λ A , let λ = W Λ . Then A − λ is a regular inner parallel body of A,which satisfies A − λ = \ λ ∈ Λ A − λ and h A − λ = ^ λ ∈ Λ h A − λ . Proof.
Firstly, with Lemma 3.3.1 we obtain that h A − λ = h A − λ = h A − _ Λ = ^ λ ∈ Λ ( h A − λ ) = ^ λ ∈ Λ h A − λ . t ∈ A − λ for all λ ∈ Λ , then Proposition 2.1.7 implies that h t , x i ≤ h A − λ ( x ) = h A ( x ) − λ for all x ∈ S n − and λ ∈ Λ , and thus h t , x i ≤ ^ λ ∈ Λ ( h A ( x ) − λ ) = h A − λ ( x )for all x ∈ S n − , which means that t ∈ A − λ . Hence T λ ∈ Λ A − λ ⊆ A − λ , which in fact becomes an identity since thereverse inclusion is trivial.In particular, A − λ A = \ λ ∈ Λ A A − λ (3.iii)is the smallest regular inner parallel body of A . We say that a convex body A ∈ C n is irreducible if A = A − λ A ;that is, if A does not have any non-trivial regular inner parallel body.Let C n i denote the set of irreducible convex bodies in R n , and let B n denote the set of closed balls in R n centered atthe origin. For each convex body A ∈ C n , the decomposition A = A − λ A + B λ A is unique in the sense of the following: Theorem 3.3.4.
For each convex body A ∈ C n , there exist a unique B ∈ C n i and a unique B λ ∈ B n such that A = B + B λ .Moreover, the correspondence A ( A − λ A , B λ A ) establishes a bijection C n ∼ ←→ C n i × B n , whose inverse is given by ( B , B λ ) B + B λ .3.4. Skew fuzzy vectors and Mareˇs cores Motivated by the notion of skew fuzzy number in the sense of Chai-Zhang [2], we introduce skew fuzzy vectors:
Definition 3.4.1.
A fuzzy vector u ∈ F n is skew if it cannot be written as the sum of a fuzzy vector and a non-trivialsymmetric fuzzy vector; that is, if u = v ⊕ w for some v ∈ F n and w ∈ F n s , then w = ˜ o .Following the terminology from fuzzy numbers [21, 15, 25], Mareˇs cores of fuzzy vectors are defined as follows: Definition 3.4.2.
A fuzzy vector v ∈ F n is a Mareˇs core of a fuzzy vector u ∈ F n if v is skew and u = v ⊕ w for some symmetric fuzzy vector w ∈ F n s .These concepts are closely related to inner parallel bodies introduced in Subsection 3.3: Lemma 3.4.3.
For fuzzy vectors u , v ∈ F n , if u = v ⊕ w for some symmetric fuzzy vector w ∈ F n s , then for each α ∈ [0 , , (i) v α is a regular inner parallel body of u α ; h u ( α, − ) − h v ( α, − ) : S n − → R is a (nonnegative) constant function.Proof. Since u = v ⊕ w and w is symmetric, Theorem 3.1.1 and Corollary 3.2.4 ensure that h u ( α, − ) = h v ( α, − ) + h w ( α ) , (3.iv)and thus (ii) holds. For (i), by setting λ = h w ( α ) and rewriting (3.iv) as h u α = h v α + λ it follows soon that v α = ( u α ) − λ and u α = v α + B λ by Lemma 3.3.1, and hence v α is a regular inner parallel body of u α .In order to construct a Mareˇs core of each fuzzy vector u ∈ F n , we start with the following proposition, in which Υ u : = { v ∈ F n | u = v ⊕ w , w ∈ F n s } is clearly a non-empty set as u ∈ Υ u : Proposition 3.4.4.
There is a fuzzy vector c ( u ) ∈ F n whose level sets are given by c ( u ) α : = T v ∈ Υ u v α if α ∈ (0 , , S β ∈ (0 , c ( u ) β if α = , and whose support function h c ( u ) : [0 , × S n − → R is given byh c ( u ) ( α, x ) = V v ∈ Υ u h v ( α, x ) if α ∈ (0 , , lim β → + h c ( u ) ( β, x ) if α = . Proof.
For the existence of c ( u ), we show that { c ( u ) α | α ∈ [0 , } satisfies the conditions (L1)–(L3) in Theorem 2.2.3,as (L4) trivially holds.Firstly, (L2) holds since v α ⊇ v β for all v ∈ Υ u whenever 0 ≤ α < β ≤ α ∈ (0 , c ( u ) α is the intersection of a family ofregular inner parallel bodies of u α , and thus c ( u ) α is itself a regular inner parallel body of u α (see Corollary 3.3.3); inparticular, c ( u ) α is a convex body. It remains to show that c ( u ) is a convex body. Indeed, c ( u ) is convex since it isthe closure of the union of a family of convex bodies linearly ordered by inclusion, and its boundedness follows from c ( u ) ⊆ u .Thirdly, in order to obtain (L3), let { α k } ⊆ (0 ,
1] be an increasing sequence that converges to α >
0. Then v α = T k ≥ v α k for each v ∈ Υ u , and thus c ( u ) α = \ v ∈ Υ u v α = \ v ∈ Υ u \ k ≥ v α k = \ k ≥ \ v ∈ Υ u v α k = \ k ≥ c ( u ) α k . For the support function of c ( u ), let α ∈ (0 , c ( u ) α is the intersection of a family of regular inner parallelbodies of u α , it follows from Corollary 3.3.3 that h c ( u ) ( α, − ) = h c ( u ) α = ^ v ∈ Υ u h v α = ^ v ∈ Υ u h v ( α, − ) . Finally, the value of h c ( u ) (0 , − ) follows from the right-continuity of h c ( u ) ( − , x ) at 0 (see Theorem 2.2.4).In fact, c ( u ) also lies in Υ u , and it is a Mareˇs core of each fuzzy vector u ∈ F n : Theorem 3.4.5. c ( u ) is skew, and there exists a symmetric fuzzy vector s ( u ) ∈ F n s such that u = c ( u ) ⊕ s ( u ) . Hence, c ( u ) is a Mareˇs core of u. roof. Step 1.
There exists a symmetric fuzzy vector s ( u ) ∈ F n s such that u = c ( u ) ⊕ s ( u ).Let h : = h u − h c ( u ) : [0 , × S n − → R . Note that for each x ∈ S n − , h ( − , x ) is left-continuous on (0 ,
1] andright-continuous at 0 since so are h u ( − , x ) and h c ( u ) ( − , x ) by Theorem 2.2.4. Moreover, as there exists a symmetricfuzzy vector w v ∈ F n s such that u = v ⊕ w v for all v ∈ Υ u , it follows from Theorem 3.1.1 and Proposition 3.4.4 that h ( α, x ) = h u ( α, x ) − h c ( u ) ( α, x ) = h u ( α, x ) − ^ v ∈ Υ u h v ( α, x ) = _ v ∈ Υ u ( h u ( α, x ) − h v ( α, x )) = _ v ∈ Υ u h w v ( α )for all α ∈ (0 , x ∈ S n − . Hence, h is nonnegative, independent of x ∈ S n − and non-increasing on α ∈ [0 ,
1] becauseso is each h w v ( v ∈ Υ u ); that is, h satisfies all the conditions of Corollary 3.2.4. Therefore, h is the support function ofa symmetric fuzzy vector s ( u ) ∈ F n s , which clearly satisfies u = c ( u ) ⊕ s ( u ) by Theorem 3.1.1. Step 2. c ( u ) is skew.Suppose that c ( u ) = v ⊕ w and w is symmetric. Then h v ≤ h c ( u ) by Lemma 3.4.3.Conversely, since w and s ( u ) are both symmetric, so is w ⊕ s ( u ) by Corollary 3.2.5. Thus, together with u = c ( u ) ⊕ s ( u ) = v ⊕ w ⊕ s ( u ) = v ⊕ ( w ⊕ s ( u ))we obtain that v ∈ Υ u , which implies that h c ( u ) ≤ h v by Proposition 3.4.4.Therefore, h c ( u ) = h v , and it forces w = ˜ o , which shows that c ( u ) is skew.An obvious application of Theorem 3.4.5 is to determine whether a fuzzy vector is skew: Corollary 3.4.6.
A fuzzy vector u ∈ F n is skew if, and only if, u = c ( u ) .Proof. The “if” part is already obtained in Theorem 3.4.5. For the “only if” part, just note that Υ u = { u } if u is skew,and thus u = c ( u ) necessarily follows.Let F n k denote the set of skew fuzzy vectors of dimension n . Theorem 3.4.5 actually induces a surjective map asfollows: Corollary 3.4.7.
The assignment ( v , w ) v ⊕ w establishes a surjective map F n k × F n s → F n . Unfortunately, as the following Example 3.4.9 reveals, unlike Theorem 1.2.1 for the case of n = not be injective. In other words, a fuzzyvector may have many Mareˇs cores, so that there may be many ways to decompose a fuzzy vector as the sum of askew fuzzy vector and a symmetric fuzzy vector!As a preparation, let us present a su ffi cient condition for a fuzzy vector to be skew that is easy to verify: Lemma 3.4.8.
Let u ∈ F n be a fuzzy vector. If the -level set u of u is an irreducible convex body, then u is skew.Proof. Suppose that u = v ⊕ w and w is symmetric. Then u α = v α + w α for all α ∈ [0 , u = v + w .Since u is irreducible, w must be trivial, i.e., w = { o } , where o is the origin of R n . The condition (L2) of Theorem2.2.3 then forces w α = { o } for all α ∈ [0 , w = ˜ o . Example 3.4.9.
Suppose that n ≥
2. For each α ∈ [0 , A α = n Y i = [ α − , − α ]be the hypercube in R n centered at the origin whose edge length is 2 − α .For every λ ∈ [0 , v λ ∈ F n whose level sets are given by( v λ ) α = A α + B λα , w λ ∈ F n s whose level sets are given by( w λ ) α = B − λα . Then there exists a fuzzy vector u ∈ F n with u α = A α + B = A α + B λα + B − λα = ( v λ ) α + ( w λ ) α for all λ ∈ [0 , u = v λ ⊕ w λ for all λ ∈ [0 , v λ ) = A = n Y i = [ − ,
1] is irreducible, from Lemma 3.4.8 we know that every v λ is skew, andtherefore every v λ (0 ≤ λ ≤
1) is a Mareˇs core of u .With Theorem 1.2.1 and Example 3.4.9 we can now conclude: Theorem 3.4.10.
Every fuzzy vector u ∈ F n has a unique Mareˇs core if, and only if, the dimension n = .3.5. Mareˇs equivalent fuzzy vectors The following definition is also originated from fuzzy numbers (see [21, 25, 2]):
Definition 3.5.1.
Fuzzy vectors u , v ∈ F n are Mareˇs equivalent , denoted by u ∼ M v , if there exist symmetric fuzzyvectors w , w ′ ∈ F n s such that u ⊕ w = v ⊕ w ′ . The relation ∼ M is clearly an equivalence relation on F n , and we denote by [ u ] M the equivalence class of each u ∈ F n . Proposition 3.5.2.
For fuzzy vectors u , v ∈ F n , the following statements are equivalent: (i) u ∼ M v. (ii) For each α ∈ [0 , , either u α is a regular inner parallel body of v α , or v α is a regular inner parallel body of u α . (iii) For each α ∈ [0 , , the function h u ( α, − ) − h v ( α, − ) is constant on S n − .Proof. (ii) ⇐⇒ (iii) is an immediate consequence of Theorem 3.1.1 and Lemma 3.3.1, and (i) = ⇒ (iii) follows soonfrom Theorems 3.1.1 and 3.2.3. For (iii) = ⇒ (i), let us fix x ∈ S n − , and let η be a common upper bound of h u and h v on [0 , × S n − . Then the functions [0 , → R α η + h u ( α, x )and [0 , → R α η + h v ( α, x )clearly satisfy the conditions of Corollary 3.2.4, and thus they are support functions of symmetric fuzzy vectors w , w ′ ∈ F n s , respectively.Since h u ( α, − ) − h v ( α, − ) is constant on S n − for each α ∈ [0 , h u ( α, x ) − h v ( α, x ) = h u ( α, x ) − h v ( α, x ) = h w ( α ) − h w ′ ( α )for all α ∈ [0 , x ∈ S n − ; that is, h u + h w ′ = h v + h w , and therefore u ⊕ w ′ = v ⊕ w by Theorem 3.1.1, showing that u ∼ M v .20 emark 3.5.3. For fuzzy vectors u , v ∈ F n , u ∼ M v does not necessarily imply that u = v ⊕ w or v = u ⊕ w for some w ∈ F n s . For example, if u ( t ) = t ∈ B , t < B and v ( t ) = − || t || t ∈ B , t < B , then u , v ∈ F n s and it trivially holds that u ⊕ v = v ⊕ u ; hence u ∼ M v . However, there is no w ∈ F n s such that u = v ⊕ w or v = u ⊕ w . Indeed, the level sets of u , v are given by u α ≡ B and v α = B − α for all α ∈ [0 , u α is a regular inner parallel body of v α when 0 ≤ α ≤
12 , and v α is a regular inner parallelbody of u α when 12 ≤ α ≤ c ( u ) ∼ M u for all u ∈ F n . In fact, every Mareˇs core of u is Mareˇs equivalent to u . In what followswe construct another skew fuzzy vector k ( u ) that is Mareˇs equivalent to u but may not be a Mareˇs core of u : Proposition 3.5.4.
There is a skew fuzzy vector k ( u ) ∈ F n whose level sets are given by k ( u ) α : = T v ∈ [ u ] M v α if α ∈ (0 , , S β ∈ (0 , k ( u ) β if α = , and whose support function h k ( u ) : [0 , × S n − → R is given byh k ( u ) ( α, x ) = V v ∈ [ u ] M h v ( α, x ) if α ∈ (0 , , lim β → + h k ( u ) ( β, x ) if α = . In particular, k ( u ) ∼ M u.Proof. The verification of k ( u ) being a fuzzy vector is similar to Proposition 3.4.4 under the help of Proposition 3.5.2,and thus we leave it to the readers. In particular, k ( u ) ∼ M u is an immediate consequence of Proposition 3.5.2 and thefact that each k ( u ) α is a regular inner parallel body of u α .To see that k ( u ) is skew, suppose that k ( u ) = v ⊕ w and w is symmetric. Then h v ≤ h k ( u ) by Lemma 3.4.3.Conversely, since k ( u ) ∼ M v , it holds that u ∼ M v , and thus h k ( u ) ≤ h v . Hence h k ( u ) = h v , which forces w = ˜ o andcompletes the proof. Remark 3.5.5.
For each α ∈ (0 , α -level set u α = { t ∈ R n | u ( t ) ≥ α } of a fuzzy vector u ∈ F n has a smallest regular inner parallel body given by Equation (3.iii) below Corollary 3.3.3. Itis then tempting to ask whether k ( u ) could be determined by the α -level sets k ( u ) α = ( u α ) − λ u α . Unfortunately, this is not true since, in general, { ( u α ) − λ u α | α ∈ [0 , } does not satisfy the conditions of Theorem 2.2.3even when n =
1. For example, the level sets of the fuzzy number u : R → [0 , , u ( t ) = − t t ∈ [0 , , u α = [0 , − α ]for all α ∈ [0 , { ( u α ) − λ u α | α ∈ [0 , } = {{ − α } | α ∈ [0 , } consists of non-identical singleton sets which obviously violate the condition (L2).From the definition it is clear that k ( u ) is the smallest fuzzy vector in the Mareˇs equivalence class of u ∈ F n ; thatis, h k ( u ) ≤ h v for all v ∈ [ u ] M . In fact, if the dimension n = k ( u ) is precisely the (unique) Mareˇs core of a fuzzy number u constructed in [2, Proposition 4.2] (cf. [2, Remark 4.4]); that is: Corollary 3.5.6.
For each fuzzy number u, it holds that k ( u ) = c ( u ) . Moreover, k ( u ) is the unique Mareˇs core of u and the unique skew fuzzy number in the Mareˇs equivalence class [ u ] M . However, if the dimension n ≥
2, the following continuation of Example 3.4.9 shows that k ( u ) may not be a Mareˇscore of u , and skew fuzzy vectors may be Mareˇs equivalent to each other: Example 3.5.7.
For the fuzzy vectors u , v λ , w λ (0 ≤ λ ≤
1) considered in Example 3.4.9, since v λ ∈ [ u ] M for all λ ∈ [0 , v λ are Mareˇs equivalent to each other. Moreover, although v = k ( u ) = k ( v λ ) , v is not a Mareˇs core of v λ whenever λ ∈ (0 , v λ is skew, its Mareˇs core must be itself.
4. Concluding remarks
The well-known Theorem 2.2.3 builds up a bridge from convex geometry to fuzzy vectors. From the viewpointof a convex geometer, the theory of fuzzy vectors is a theory about sequences of convex bodies . Our main result,the representation of fuzzy vectors via support functions (Theorem 2.2.4), can never be achieved without the aid ofconvex geometry. We believe that the power of convex geometry in the study of fuzzy vectors is still to be unveiled,which is worth further exploration.We end this paper with two questions about Mareˇs cores of fuzzy vectors that remain unsolved:(1) Is it possible to find all the Mareˇs cores of a given fuzzy vector?(2) Is it possible to find a necessary and su ffi cient condition to characterize those fuzzy vectors with exactly oneMareˇs core? Acknowledgment
The authors acknowledge the support of National Natural Science Foundation of China (No. 11501393 and No.11701396), and the first named author also acknowledges the support of Sichuan Science and Technology Program(No. 2019YJ0509) and a joint research project of Laurent Mathematics Research Center of Sichuan Normal Universityand V. C. & V. R. Key Lab of Sichuan Province.The authors are grateful for helpful comments received from Professor Hongliang Lai and Professor BaochengZhu. 22 eferences [1] T. Bonnesen and W. Fenchel.
Theorie der Konvexen K¨orper . Springer, Berlin–Heidelberg, 1934.[2] Y. Chai and D. Zhang. A representation of fuzzy numbers.
Fuzzy Sets and Systems , 295:1–18, 2016.[3] P. Diamond and P. Kloeden. Characterization of compact subsets of fuzzy sets.
Fuzzy Sets and Systems , 29(3):341–348, 1989.[4] P. Diamond and P. Kloeden.
Metric Spaces of Fuzzy Sets: Theory and Applications . World Scientific, Singapore, 1994.[5] J. G. Dijkman, H. van Haeringen, and S. J. de Lange. Fuzzy numbers.
Journal of Mathematical Analysis and Applications , 92(2):301–341,1983.[6] D. Dubois and H. Prade. Operations on fuzzy numbers.
International Journal of Systems Science , 9(6):613–626, 1978.[7] D. Dubois and H. Prade.
Fuzzy Sets and Systems: Theory and Applications , volume 144 of
Mathematics in Science and Engineering .Academic Press, 1980.[8] D. Dubois and H. Prade. Towards fuzzy di ff erential calculus part 1: Integration of fuzzy mappings. Fuzzy Sets and Systems , 8(1):1–17, 1982.[9] D. Dubois and H. Prade. Towards fuzzy di ff erential calculus part 2: Integration on fuzzy intervals. Fuzzy Sets and Systems , 8(2):105–116,1982.[10] D. Dubois and H. Prade. Towards fuzzy di ff erential calculus part 3: Di ff erentiation. Fuzzy Sets and Systems , 8(3):225–233, 1982.[11] D. Dubois and H. Prade. Fuzzy numbers: an overview. In D. Dubois, H. Prade, and R. R. Yager, editors,
Readings in Fuzzy Sets for IntelligentSystems , pages 112–148. Morgan Kaufmann, 1993.[12] R. Goetschel Jr. and W. Voxman. Topological properties of fuzzy numbers.
Fuzzy Sets and Systems , 10(1):87–99, 1983.[13] P. M. Gruber.
Convex and Discrete Geometry , volume 336 of
A Series of Comprehensive Studies in Mathematics . Springer, Berlin–Heidelberg,2007.[14] P. M. Gruber and J. M. Wills, editors.
Handbook of Convex Geometry . North Holland, Amsterdam, 1993.[15] D. H. Hong. An additive decomposition of fuzzy numbers.
Kybernetika , 39(3):289–294, 2003.[16] O. Kaleva. On the convergence of fuzzy sets.
Fuzzy Sets and Systems , 17(1):53–65, 1985.[17] O. Kaleva. Fuzzy di ff erential equations. Fuzzy Sets and Systems , 24(3):301–317, 1987.[18] A. Koldobsky and V. Yaskin.
The Interface Between Convex Geometry and Harmonic Analysis , volume 108 of
CBMS Regional ConferenceSeries in Mathematics . American Mathematical Society, Providence, 2008.[19] T. Maeda. On characterization of fuzzy vectors and its applications to fuzzy mathematical programming problems.
Fuzzy Sets and Systems ,159(24):3333–3346, 2008.[20] M. Mareˇs. Addition of fuzzy quantities: disjunction-conjunction approach.
Kybernetika , 25(2):104–116, 1989.[21] M. Mareˇs. Additive decomposition of fuzzy quantities with finite supports.
Fuzzy Sets and Systems , 47(3):341–346, 1992.[22] C. V. NegoiT¸ ˘a and D. A. Ralescu.
Applications of Fuzzy Sets to Systems Analysis , volume 11 of
Interdisciplinary Systems Research / Interdisziplin¨are Systemforschung . Birkh¨auser, Basel, 1975.[23] H. T. Nguyen. A note on the extension principle for fuzzy sets.
Journal of Mathematical Analysis and Applications , 64(2):369–380, 1978.[24] M. L. Puri and D. A. Ralescu. The concept of normality for fuzzy random variables.
The Annals of Probability , 13(4):1373–1379, 1985.[25] D. Qiu, C. Lu, W. Zhang, and Y. Lan. Algebraic properties and topological properties of the quotient space of fuzzy numbers based on mareÅequivalence relation.
Fuzzy Sets and Systems , 245:63–82, 2014.[26] R. Schneider.
Convex bodies: the Brunn-Minkowski theory , volume 151 of
Encyclopedia of Mathematics and its Applications . CambridgeUniversity Press, Cambridge, second edition, 2013.[27] G. Wang, Y. Li, and C. Wen. On fuzzy n-cell numbers and n-dimension fuzzy vectors.
Fuzzy Sets and Systems , 158(1):71–84, 2007.[28] G. Wang and C. Wu. Fuzzy n-cell numbers and the di ff erential of fuzzy n-cell number value mappings. Fuzzy Sets and Systems , 130(3):367–381, 2002.[29] L. A. Zadeh. Fuzzy sets.
Information and Control , 8(3):338–353, 1965.[30] L. A. Zadeh. The concept of a linguistic variable and its application to approximate reasoning–I.
Information Sciences , 8(3):199–249, 1975.[31] B. Zhang. On measurability of fuzzy-number-valued functions.
Fuzzy Sets and Systems , 120(3):505–509, 2001.[32] B. Zhang and C. Wu. On the representation of n-dimensional fuzzy numbers and their informational content.
Fuzzy Sets and Systems ,128(2):227–235, 2002.,128(2):227–235, 2002.