Application of the Fuzzy Weighted Average of Fuzzy Numbers in Decision Making Models

Size: px
Start display at page:

Download "Application of the Fuzzy Weighted Average of Fuzzy Numbers in Decision Making Models"

Transcription

1 Application of the Fuzzy Weighted Average of Fuzzy Numbers in Decision Making Models Ondřej Pavlačka Department of Mathematical Analysis and Applied Mathematics, Faculty of Science, Palacký University Olomouc, tř. Svobody 6, Olomouc, Czech Republic Jana Talašová Department of Mathematical Analysis and Applied Mathematics, Faculty of Science, Palacký University Olomouc, tř. Svobody 6, Olomouc, Czech Republic Abstract The paper deals with the operation of fuzzy weighted average of fuzzy numbers. The operation can be applied to the aggregation of partial fuzzy evaluations in fuzzy models of multiple criteria decision making and to the computation of expected fuzzy evaluations of alternatives in discrete fuzzy-stochastic models of decision making under risk. Normalized fuzzy weights figuring in the operation have to form a special structure of fuzzy numbers; its properties will be studied and the practical procedures for setting them will be proposed. A fuzzy weighted average of fuzzy numbers with normalized fuzzy weights will be defined, an effective algorithm of its calculation will be described, and its uncertainty will be studied. Keywords: Multiple criteria decision making, decision making under risk, normalized fuzzy weights, fuzzy weighted average, fuzzy weights of criteria, fuzzy probabilities. In multiple criteria decision making models, the weighted average is often used for aggregating partial evaluations of alternatives. Weights of criteria are understood differently in different models. The most general definition says that the weights of criteria are non-negative real numbers whose ordering expresses the importance of criteria. According to this definition, the weights mean the measurements of criteria importance that are defined on an ordinal scale. But in most of multiple criteria decision making models, the weights of criteria represent some kind of cardinal information about the importance of criteria. For example, in the model of multiple criteria evaluation that was introduced in [8] the weights of criteria represent shares of the corresponding partial objectives of evaluation in the overall one. This concept of criteria weights will be considered in this paper. The weighted average operation is applied also in the models of decision making under risk where evaluations of alternatives depend on a finite number of states of the world whose probabilities are known. The best alternative is chosen according to expected evaluations of alternatives. They are computed by means of the weighted average operation where the probabilities of states of the world play the role of weights. The weights of criteria as well as the probabilities of states of the world are usually set subjectively, i.e. they are more or less uncertain. In this paper, it will be shown how this kind of uncertain information can be expressed by means of tools of fuzzy sets theory. A special structure of fuzzy numbers, normalized fuzzy weights, will be introduced for expressing the uncertain normalized weights and the operation of the fuzzy weighted average of fuzzy numbers with normalized fuzzy weights will be defined. In the end, application of the fuzzy weighted average operation in decision making models will be shown. 1 Introduction Applied Notions of Fuzzy Sets Theory In this paper, the following notation will be used. The symbol N m denotes the index set {1,,...,m}. A fuzzy set A on a universal set X is characterized by its membership function A : X [0,1]. Ker A denotes the kernel of A, i.e. Ker A = {x X A(x) = 1}. For α (0,1], A α denotes the α-cut of A, i.e. A α = {x X A(x) α}. Supp A denotes the support of A, i.e. Supp A = {x X A(x) > 0}. A fuzzy number is a fuzzy set C on the set of all real numbers R that fulfils the following conditions: a) KerC, b) C α are closed intervals for all α (0,1], c) Supp C is a bounded set. A fuzzy number C is

2 said to be defined on [a,b], if Supp C [a,b]. Let us denote [c 1,c 4 ] = Cl(Supp C), [c,c 3 ] = Ker C; Cl(Supp C) means the closure of Supp C. The real numbers c 1 c c 3 c 4 are called significant values of C. A fuzzy number C can be also characterized (see [3]) by the pair of functions (c,c), both from [0,1] to R, that satisfy the following property c(α) c(β) c(β) c(α), 0 α β 1. (1) Then, for the membership function C of the fuzzy number C = (c,c) the following holds C(x) = max { α x [c(α),c(α)] } () for x [c(0),c(0)], and 0 elsewhere; i.e. C α = [c(α),c(α)] for all α (0,1], Cl(SuppC) = [c(0),c(0)]. If c(α) = c = c(α), for all α [0,1], then C represents a real number; C = c, c R. A fuzzy number C is called symmetric with the central point c if the following holds c = c(α) + c(α), for α [0,1]. (3) A fuzzy number C = (c,c) is called a linear fuzzy number if both c and c are linear functions. Any linear fuzzy number C is fully characterized by its significant values c 1 c c 3 c 4 ; the functions c and c are given for all α [0,1] as follows c(α) = c 1 + α(c c 1 ), c(α) = c 4 α(c 4 c 3 ). (4) The notation C = c 1,c,c 3,c 4 is used in this paper for such a fuzzy number C. For c c 3, the linear fuzzy number C is usually referred to as a trapezoidal fuzzy number, for c = c 3 as a triangular fuzzy number. For any fuzzy number C = (c,c), the non-increasing, non-negative function σ C (α) = c(α) c(α), for α [0,1], (5) is called a width function of C. It describes in details, for any membership degree α, the uncertainty of C. The real numbers σ C (1), σ C (α) and σ C (0) are called the width of the kernel, the width of the α-cut and the width of the support, respectively. Obviously, σ C (α) = 0 for all α [0,1] if and only if C is a real number. It holds that 1 σ C (α) dα = C(x) dx, (6) 0 where the right-hand side of the equation represents the usual aggregated measure of uncertainty of the fuzzy number C. R Calculations with fuzzy numbers are based on the extension principle (see [3]). Let f : [a 1,b 1 ] [a,b ]... [a n,b n ] R be a continuous function and let f([a 1,b 1 ] [a,b ]... [a n,b n ]) = [c,d]. Let X i be fuzzy numbers defined on [a i,b i ], for i N n. Then f(x 1,X,...,X n ) is a fuzzy number Y defined on [c,d] whose membership function is given as follows Y (y) = max { min{x 1 (x 1 ),X (x ),...,X n (x n )} y = f(x 1,x,...,x n ) } (7) for y [c,d], and 0 elsewhere. Employing the (y,y) representation of the fuzzy number Y, the following holds for any α [0,1] y(α) = min{f(x 1,x,...,x n ) x i X iα, i N n }, y(α) = max{f(x 1,x,...,x n ) x i X iα, i N n }. (8) In some cases (see [5] for more details), it is necessary to consider a relation R among the variables x 1,x,...,x n. Let R be a closed and convex subset of [a 1,b 1 ] [a,b ]... [a n,b n ] and let f R : R [c,d] be the restriction of f on R. Let there exist at least one n-tuple (x 1,x,...,x n ) R such that X i (x i ) = 1 for all i N n. Then f R (X 1,X,...,X n ) is a fuzzy number Y R defined on [c,d] whose membership function is given in the following way Y R (y) = max { min{x 1 (x 1 ),X (x ),...,X n (x n )} y = f(x 1,x,...,x n ), (x 1,x,...,x n ) R } (9) for y [c,d], and 0 elsewhere. Employing the (y R,y R ) representation of the fuzzy number Y R, the following holds for any α [0,1] y R (α) = min { f(x 1,x,...,x n ) (x 1,x,...,x n ) (X 1α X α... X nα ) R }, y R (α) = max { f(x 1,x,...,x n ) (x 1,x,...,x n ) (X 1α X α... X nα ) R }. (10) 3 Normalized Fuzzy Weights In this section, a special structure of fuzzy numbers, normalized fuzzy weights, will be introduced, and its properties will be studied. Normalized fuzzy weights represent a fuzzification of crisp normalized weights that are defined as non-negative real numbers w 1,w,...,w m such that m w i = 1. Fuzzy numbers W 1,W,...,W m defined on [0,1] are called normalized fuzzy weights if for any α (0,1] and for all i N m the following holds: For any w i W iα there exist w j W jα, j N m, j i, such that w i + m,j i w j = 1. (11)

3 Figure 1: Example of normalized fuzzy weights. Figure 1 illustrates an example of normalized fuzzy weights. Obviously, crisp normalized weights w 1,w,...,w m represent a special case of normalized fuzzy weights. From the general point of view, normalized fuzzy weights make it possible to model mathematically an uncertain division of a unit into m fractions. In decision making models, the membership degrees W i (w i ), i N m, mean either the possibility that the share of the i-th partial objective of evaluation in the overall one is equal to w i, or the possibility that the probability of the i-th state of the world is equal to w i. In [6], fuzzy numbers defined on [0,1] fulfilling (11) were called a tuple of fuzzy probabilities. They represented the generalization of interval probabilities developed in [1], and they were applied for modelling imprecise probabilities. The structure of normalized fuzzy weights was independently introduced in [7] in order to model the uncertain weights of criteria in the method of the weighted average of partial fuzzy evaluations that was described in [8]. Later on, the normalized fuzzy weights were used to model uncertain probabilities in fuzzy-stochastic models of decision making under risk (see [10, 11]). Verifying the condition (11) directly can be very complicated. Hence, it was proved in [6] that fuzzy numbers W i = (w i,w i ), i N m, defined on [0,1] fulfil (11) if and only if for any α [0,1] and for all i N m the following two conditions hold w i (α) + w i (α) +,j i,j i w j (α) 1, (1) w j (α) 1. (13) Linear fuzzy numbers W i = w 1 i,w i,w3 i,w4 i, i N m, defined on [0,1] fulfil (11) if and only if the conditions (1) and (13) hold for α = 1 and α = 0, i.e. if for all i N m the following holds w 3 i + w 4 i +,j i,j i w j 1 and w i + w 1 j 1 and w 1 i +,j i,j i w 3 j 1, (14) w 4 j 1. (15) The linearity of the functions w i and w i, i N m, ensures that the conditions (1) and (13) are fulfilled for any α [0,1] (see [6]). For m =, we can see immediately from (1) and (13) that w (α) = 1 w 1 (α) and w (α) = 1 w 1 (α) (16) for all α [0,1]. It means that W = 1 W 1. For i N m, let us denote by σ Wi the width function of the fuzzy number W i = (w i,w i ), i.e. σ Wi (α) = w i (α) w i (α), α [0,1]. For the sake of simplicity, the conditions (1) and (13) can be, for all α [0,1], replaced by only one condition in one of the following two ways: First, subtracting m w k(α) from (1) and (13) we obtain σ Wi (α) 1 w k (α), (17),j i σ Wj (α) 1 w k (α) (18) for all i N m. Let i (α) N m be an index such that Then, j i (α) σ Wi (α) (α) = σ Wj (α) = max {σ W i (α)}. (19),,...,m min,,...,m { m j i } σ Wj (α). (0) Hence, the conditions (1) and (13) are equivalent to the following one σ Wi (α) (α) 1 w k (α) j i (α) σ Wj (α). (1) Analogously, in the case of linear fuzzy numbers W i = wi 1,w i,w3 i,w4 i, i N m, the following couple of conditions σ Wi (1) (1) 1 wk σ Wj (1), () σ Wi (0) (0) 1 wk 1,j i (1),j i (0) σ Wj (0) (3)

4 is equivalent to (14) and (15). Second, multiplying (1) and (13) by ( 1) and adding m w k(α) to these inequalities, we can see that the conditions (1) and (13) are equivalent to the following one σ Wi (α) (α) w k (α) 1 j i (α) σ Wj (α), (4) where the index i (α) N m is given by (19). In the case of linear fuzzy numbers W i = w 1 i,w i,w3 i,w4 i, i N m, the following couple of conditions σ Wi (1) (1) σ Wi (0) (0) wk 3 1 wk 4 1 is equivalent to (14) and (15).,j i (1),j i (0) σ Wj (1), (5) σ Wj (0) (6) At the end of this section, the case of symmetric normalized fuzzy weights will be studied. Let w 1,w,...,w m be normalized weights and let W i = (w i,w i ), i N m, be symmetric fuzzy numbers such that for all α [0,1] the following holds w i (α) = w i e i (α) and w i (α) = w i + e i (α), (7) where e i : [0,1] [0,min{w i,1 w i }], i N m, are non-increasing functions. Then, from (1) or (4) it follows that the fuzzy numbers W 1,W,...,W m fulfil the condition (11) if and only if for all α [0,1] the following holds where e i (α)(α) e i (α)(α) =,j i (α) e j (α), (8) max {e i(α)}. (9),,...,m Obviously, if e i are linear functions for all i N m, then (8) is fulfilled for all α [0,1] if and only if e i (1)(1) e i (0)(0),j i (1),j i (0) e j (1), (30) e j (0). (31) 4 Procedures of Setting Normalized Fuzzy Weights In practical applications, it is useful to set normalized fuzzy weights as linear fuzzy numbers W i = wi 1,w i,w3 i,w4 i, i N m. The interpretation is the following: [wi,w3 i ] represents the interval of fully possible values of the i-th weight, while outside the interval [wi 1,w4 i ] the i-th weight cannot lie. For m =, it is sufficient to set only one weight in the form of a linear fuzzy number W = w 1,w,w 3,w 4 defined on [0,1]; then it follows from (16) that the other is the linear fuzzy number 1 W = 1 w 4,1 w 3,1 w,1 w 1. For m >, it is not so easy for an expert to set the normalized fuzzy weights directly. Therefore, two procedures of setting normalized fuzzy weights will be shown here. The first one is based on the transformation of expert s estimates of fuzzy weights into the normalized fuzzy weights. The second one is a fuzzification of a crisp estimation of normalized weights. In the first procedure, an expert expresses his/her estimates of the weights by linear fuzzy numbers W i = w i 1,w i,w 3 i,w 4 i, i N m, whose significant values satisfy the following natural condition w 1 i w i 1 w 3 i w 4 i. (3) It means that the fuzzy numbers W 1,W,...,W m need not fulfil the condition (11), but there must exist at least one m-tuple of normalized weights w 1,w,...,w m such that W i (w i) = 1, i N m. The associated normalized fuzzy weights W i = wi 1,w i,w3 i,w4 i, i N m, are obtained from W 1,W,...,W m by the following transformation w k i = max { w k i,1 w l i = min { w l i,1,j i,j i w 5 k } j, w 5 l } j, (33) for k = 1,, l = 3,4. The transformation (33), that was originally described for interval probabilities in [1], eliminates the inconsistency of expert s estimates (see Figure ). For any α (0,1] and for every m-tuple of crisp normalized weights w 1,w,...,w m such that w i W iα for all i N m, it holds that w i W iα for all i N m. It means that no relevant information concerning the weights was lost by the transformation. In the second procedure, an expert sets the crisp estimation of normalized weights by real numbers w 1,w,...,w m such that w i 0, i N m, and m w i = 1. Then, he/she sets coefficients k i and s i, 0 k i s i, w i s i 0, w i + s i 1, i N m, that characterize the uncertainty of kernels and supports of the particular fuzzy weights. It follows from

5 1 W W Let W 1,W,...,W m be non-negative fuzzy numbers, i.e. Supp W i [0, ) for all i N m, and let there exist w > 0 and an index i N m such that W i (w) = 1. Then a fuzzy weighted average of fuzzy numbers U 1,U,...,U m with fuzzy weights W 1,W,...,W m was defined in [] as a fuzzy number U W whose membership function is given, for each u R, in the following way Figure : Transformation of expert s estimation of fuzzy weights into the normalized fuzzy weights. (30) and (31) that the symmetric linear fuzzy numbers W i = w i s i,w i k i,w i + k i,w i + s i, i N m, are normalized fuzzy weights if k i and s i, i N m, satisfy the following condition k i k i and s j,i i,j j s j, (34) where k i = max{k i, i N m } and s j = max{s j, j N m }. For instance, the condition (34) is satisfied in the case of symmetric triangular fuzzy numbers W i = w i s,w i,w i,w i + s, i N m, where w 1,w,...,w m are normalized weights, s 0, and, for all i N m, w i s 0 and w i + s 1. 5 Fuzzy Weighted Averages of Fuzzy Numbers A weighted average of real numbers u 1,u,...,u m with weights w 1,w,...,w m is defined by the following formula u = w 1u 1 + w u w m u m w 1 + w w m, (35) where the weights w 1,w,...,w m are generally nonnegative real numbers whose sum is different from zero. Especially, if m w i = 1, i.e. if the weights w 1,w,...,w m are normalized, the weighted average can be expressed as follows u = w 1 u 1 + w u w m u m. (36) In this paper, it will be shown that these formulas do not coincide in fuzzy case. U W (u) = max { min{u 1 (u 1 ),U (u ),...,U m (u m ), W 1 (w 1 ),W (w ),...,W m (w m )} u = w1u1+wu+...+wmum w 1+w +...+w m, m w i > 0 }. (37) The fuzzy weighted average U W represents a fuzzification according to (9) of the operation (35). An algorithm of computing the fuzzy weighted average U W was described in [4]. The fuzzy weighted average U W cannot be used, if the fuzzy weights W 1,W,...,W m have the special interpretation described in Section 1: If they express either uncertain shares of partial objectives of evaluation in the overall one in multiple criteria decision making models or uncertain probabilities of states of the world in models of decision making under risk. In such cases, the structure of normalized fuzzy weights has to be used, and a fuzzification according to (9) of the formula (36) instead of the formula (35) has to be applied for computing the fuzzy weighted average. The problem is illustrated by the following example: Example 1: Let S 1, S and S 3 be states of the world; their uncertain probabilities be, for simplicity, defined by the intervals P 1 = [0.1,0.], P = [0.4,0.6] and P 3 = [0.3,0.4] (P 1, P and P 3 express possible values of probabilities and, therefore, form the structure of normalized fuzzy weights). Let evaluations of an alternative x under the states of the world S 1, S, S 3 be given by the intervals U 1 = [0.1,0.], U = [0.4,0.5] and U 3 = [0.8,0.9]. Then, according to (37) where W i are replaced by P i for i {1,,3}, U W = [0.45,0.64]. When the minimal value of U W, p 3 p 1+p +p 3 is i.e. 0.45, is computed, the third weight 0.3 equal to = 0.7, but 0.7 P 3. Notice that this value lies outside the expertly set range of possible probabilities of S 3 ; therefore, a fuzzy expected evaluation of the alternative x cannot be computed by (37). Let fuzzy numbers W 1,W,...,W m be normalized fuzzy weights. Then the fuzzy weighted average of fuzzy numbers U 1,U,...,U m with normalized fuzzy weights W 1,W,...,W m is defined as a fuzzy number U N whose membership function is given, for each u R, as follows U N (u) = max { min{u 1 (u 1 ),U (u ),...,U m (u m ), W 1 (w 1 ),W (w ),...,W m (w m )} u = m w i u i, m w i = 1 }. (38) The fuzzy weighted average U N represents the generalization of the operation that was introduced in [1] for computing the expected value of a discrete random variable with interval probabilities. The fuzzy

6 weighted average U N was independently introduced in [7] for aggregating partial fuzzy evaluations of an alternative in the model of multiple criteria evaluation. The following denotation will be used for the fuzzy weighted average U N of fuzzy numbers U 1,U,...,U m with normalized fuzzy weights W 1,W,...,W m U N = (N) W i U i. (39) In the case of crisp normalized weights, the fuzzy weighted averages U W and U N coincide. However, this does not hold generally for normalized fuzzy weights. If W 1,W,...,W m form the structure of normalized fuzzy weights, then Uα N Uα W for all α [0,1]. It follows from the fact that even such w 1,w,...,w m whose sum is not equal to one enter the calculation of U W. For example, the fuzzy weighted average U N that expresses the fuzzy expected evaluation of the alternative x in Example 1 is equal to [0.46, 0.63]. The following algorithm of computing the fuzzy weighted average U N is based on an algorithm that was originally developed in [1] for computing the expected value of a discrete random variable with interval probabilities. The following representation of the considered fuzzy numbers will be used: U N = (u N,u N ), U i = (u i,u i ) and W i = (w i,w i ), i N m. For each α [0,1], let {i k } m be such a permutation on an index set N m that u i1 (α) u i (α)... u im (α). For k N m, let us denote k 1 w ik (α) = 1 w ij (α) j=k+1 w ij (α). (40) Let k N m be such an index that w ik (α) w ik (α) w ik (α). Then u N (α) = w ik (α) u ik (α) + k 1 j=k +1 w ij (α) u ij (α) + w ij (α) u ij (α). (41) Let {i h } m h=1 be such a permutation on an index set N m that u i1 (α) u i (α)... u im (α). For h N m, let us denote h 1 w ih (α) = 1 w ij (α) j=h+1 w ij (α). (4) Let h N m be such an index that w ih (α) w ih (α) w ih (α). Then u N (α) = h 1 w ij (α) u ij (α) + w ih (α) u ih (α) + j=h +1 w ij (α) u ij (α). (43) Let us consider now the case where the fuzzy numbers U i = (u i,u i ) and the normalized fuzzy weights W i = (w i,w i ), i N m, are linear fuzzy numbers, i.e. the functions u i, u i as well as w i, w i are linear. It follows from (41) and (43) that the fuzzy weighted average U N = (N) m W i U i is not a linear fuzzy number, in general. 6 The Fuzzy Weighted Average U N in Decision Making Models In this section, a multiple criteria decision making model and a model of decision making under risk where the fuzzy weighted average U N is applied, will be described. First, let us consider a problem of multiple criteria decision making, where the best of alternatives x 1,x,...,x n is to be chosen. The alternatives are being evaluated with respect to a given objective that is partitioned into m partial objectives associated with criteria C 1,C,...,C m. Let the uncertain information about the shares of the partial objectives in the overall one be given by normalized fuzzy weights W 1,W,...,W m. Let uncertain partial fuzzy evaluations of alternatives x i, i N n, with respect to the criteria C j, j N m, be expressed by fuzzy numbers U i,j defined on [0,1] that represent the fuzzy degrees of fulfilment of the corresponding partial objectives of evaluation. Then the overall fuzzy evaluation Ui N of the alternative x i, i N n, can be expressed as follows U N i = (N) W j U i,j. (44) The overall fuzzy evaluations U N i, i N n, express the fuzzy degrees of fulfilment of the overall objective of evaluation. The best alternative is the first alternative in an ordering of the fuzzy numbers U i, i N n, or the closest to the ideal alternative whose evaluation is equal to 1. The overall fuzzy evaluations of alternatives can also be approximated linguistically by linearly ordered elements of a proper linguistic evaluation scale defined on [0,1]. For more details on metrics, ordering of fuzzy numbers and linguistic approximation see [8]. Second, let us consider a problem of decision making under risk that is described by the following fuzzy decision matrix (see Table 1), where x 1,x,...,x n are alternatives, S 1,S,...,S r states of the world, P 1,P,...,P r their fuzzy probabilities, and U i,k, i N n, k N r, denote fuzzy degrees in which the alter-

7 natives x i, i N n, satisfy a given decision objective if the states S k, k N r, occur. Table 1: Fuzzy decision matrix. S 1 S... S r P 1 P... P r FEU N x 1 U 1,1 U 1,... U 1,r FEU N 1 x U,1 U,... U,r FEU N x n U n,1 U n,... U n,r FEU N n For i N n, the fuzzy numbers FEUi N express expected fuzzy evaluations of alternatives x i ; i.e. they are calculated according to the formula FEU N i = (N) r P k U i,k. (45) The best alternative is the alternative with the maximum expected fuzzy evaluation. Examples of practical applications of such decision making models, namely in finance and banking, are described in [10, 11]. A similar approach can be applied also to multiple criteria decision making under risk (see [9]). 7 Width Function of the Fuzzy Weighted Average U N In the last section, the dependance of the uncertainty of the fuzzy weighted average U N on given fuzzy numbers U 1,U,...,U m and on given normalized fuzzy weights W 1,W,...,W m will be studied. The uncertainty of U N = (u N,u N ) is described by its width function σ U N(α) = u N (α) u N (α), α [0,1]. Let us denote by σ Ui the width functions of the fuzzy numbers U i = (u i,u i ) for i N m, i.e. σ Ui (α) = u i (α) u i (α), α [0,1]. Let wα = (w1α,w α,...,w mα) and wα = (w1α,w α,...,w mα) be, for each α [0,1], the couple of normalized weights such that wiα,w iα W m iα, i N m, w iα = = 1, and m w iα u N (α) = wiα u i (α), u N (α) = wiα u i (α). (46) Then, for all α [0,1], the width function of U N is given by σ U N(α) = w iα u i (α) wiα u i (α). (47) From (47) it can be derived the following formula wiα σ U N(α) = + w iα σ Ui (α) + (wiα wiα) ui(α) + u i (α). (48) Let us denote ε i (α) = u i(α) + u i (α) 1 m u j (α) + u j (α), (49) then the formula (48) can be rewritten as follows wiα + w iα σ Ui (α) + σ U N(α) = (w iα w iα) ε i (α). (50) The following two examples will illustrate how the dependance characterized by the formula (50) affects the behaviour of the decision making models described above. Example : Let a pair of normalized fuzzy weights be given by the triangular fuzzy numbers W 1 = W = 0,0.5,0.5,1, and let two pairs of weighted values be given by the real numbers u 11 = 0.1, u 1 = 0.6, and u 1 = u = 0.8. Figure 3 illustrates the fact that the more different the weighted values are, the more the uncertainty of normalized fuzzy weights affects the uncertainty of the fuzzy weighted average U N. In the first case, where the weighted values are different, u 11 u 1, their fuzzy weighted average U1 N = 0.1,0.35,0.35,0.6. In the second case, where the weighted values are equal, u 1 = u, their fuzzy weighted average U N = 0.8; thus the uncertainty of the normalized weights does not affect the fuzzy weighted average at all. Figure 3: The dependance of the uncertainty of the fuzzy weighted average on the variability of the weighted values. For the multiple criteria decision making model mentioned above it means the following: If the uncertainties of the partial fuzzy evaluations are approximately

8 the same, then the overall fuzzy evaluation of an alternative whose partial fuzzy evaluations according to the particular criteria are very different is more uncertain than the overall fuzzy evaluation of an alternative whose partial fuzzy evaluations are almost uniform. Similarly, in the discrete model of decision making under risk, the expected fuzzy evaluation of an alternative that depends strongly on states of the world is more uncertain than the expected fuzzy evaluation of an alternative that is relatively stable. weights and on the uncertainty of the weighted fuzzy values but also on the variance of the weighted fuzzy values. References [1] L. M. De Campos, J. F. Heute, S. Moral, Probability inrtervals: a tool for uncertain reasoning, Int. J. of Uncertainty, Fuzziness and Knowledge- Based Systems (1994) [] W. M. Dong, F. S. Wong, Fuzzy weighted averages and implementation of the extension principle, Fuzzy Sets and Systems 1 (1987) [3] D. Dubois, E. Kerre, R. Mesiar, H. Prade, Fuzzy interval analysis, in: D. Dubois, H. Prade (Eds.), Fundamentals of fuzzy sets. Kluwer Academic Publishers, Boston-London-Dordrecht, 000, pp Figure 4: Fuzzy weighted average U N with normalized weights w 1 = 0. and w = 0.8. Example 3: Let the weighted values be given by the triangular fuzzy numbers U 1 = 0.1,0.4,0.4,0.7 and U = 0.75,0.8,0.8,0.85, and the normalized weights by the real numbers w 1 = 0. and w = 0.8. The fuzzy weighted average of U 1 and U with the normalized weights w 1 and w is U N 1 = 0.6,0.7,0.7,0.8. Figure 4 shows that the uncertainty of U N depends, apart from other things, on the size of the weights. It means that the uncertainty of the partial fuzzy evaluations with small weights (or with small probabilities) cannot affect strongly the uncertainty of the overall fuzzy evaluation (fuzzy expected evaluation). 8 Conclusion In decision making models based on the weighted average operation, weights of criteria as well as probabilities of states of the world are usually defined expertly. Therefore the models become more realistic if uncertainty of the weights and probabilities is taken into account. In the paper, the uncertain weights and probabilities were expressed by normalized fuzzy weights, and the fuzzy weighted average operation was applied to computation of the aggregated fuzzy evaluations in models of multiple criteria decision making and to computation of the fuzzy expected evaluations in models of decision making under risk. It was shown that the uncertainty of the resulting fuzzy evaluations depends not only on the uncertainty of the normalized fuzzy [4] Y. -Y. Guh, et al., Fuzzy weighted average: The linear programming approach via Charness and Cooper s rule, Fuzzy Sets and Systems 117 (001) [5] G. J. Klir, Y. Pan, Constrained fuzzy arithmetic: Basic questions and some answers, Soft Computing (1998) [6] Y. Pan, B. Yuan, Bayesian Inference of Fuzzy Probabilities, Int. J. General Systems 6 (1997) [7] O. Pavlačka, Normalized fuzzy weights (in Czech), Diploma thesis, Palacký University Olomouc, 004. [8] J. Talašová, Fuzzy Methods of Multiple Criteria Evaluation and Decision Making (in Czech), VUP, Olomouc 003. [9] J. Talašová, Fuzzy sets in decision-making under risk. (In Czech.), in: M. Kováčová (Eds.), Proceedings of the 4th Conference Aplimat. EX, Bratislava 005, pp [10] J. Talašová, Fuzzy approach to evaluation and decision making (in Czech), in: Proceedings of Aplimat 006, SUT, Bratislava 006, pp [11] J. Talašová, O. Pavlačka, Fuzzy Probability Spaces and Their Applications in Decision Making, Austrian J. of Statistics 35 (006)

Fuzzified weighted OWA (FWOWA) operator

Fuzzified weighted OWA (FWOWA) operator INTERNATIONAL JOURNAL OF MATHEMATICS AND COMPUTERS IN SIMULATION Volume 9, 25 Fuzzified weighted OWA (FWOWA) operator Pavel Holeček, Jana Talašová, Tomáš Talášek Abstract In practice, it is often necessary

More information

Probability of fuzzy events

Probability of fuzzy events Probability of fuzzy events 1 Introduction Ondřej Pavlačka 1, Pavla Rotterová 2 Abstract. In economic practice, we often deal with events that are defined only vaguely. Such indeterminate events can be

More information

Fuzzy relation equations with dual composition

Fuzzy relation equations with dual composition Fuzzy relation equations with dual composition Lenka Nosková University of Ostrava Institute for Research and Applications of Fuzzy Modeling 30. dubna 22, 701 03 Ostrava 1 Czech Republic Lenka.Noskova@osu.cz

More information

Crisp Profile Symmetric Decomposition of Fuzzy Numbers

Crisp Profile Symmetric Decomposition of Fuzzy Numbers Applied Mathematical Sciences, Vol. 10, 016, no. 8, 1373-1389 HIKARI Ltd, www.m-hikari.com http://dx.doi.org/10.1988/ams.016.59598 Crisp Profile Symmetric Decomposition of Fuzzy Numbers Maria Letizia Guerra

More information

Additive Consistency of Fuzzy Preference Relations: Characterization and Construction. Extended Abstract

Additive Consistency of Fuzzy Preference Relations: Characterization and Construction. Extended Abstract Additive Consistency of Fuzzy Preference Relations: Characterization and Construction F. Herrera a, E. Herrera-Viedma a, F. Chiclana b Dept. of Computer Science and Artificial Intelligence a University

More information

Group Decision-Making with Incomplete Fuzzy Linguistic Preference Relations

Group Decision-Making with Incomplete Fuzzy Linguistic Preference Relations Group Decision-Making with Incomplete Fuzzy Linguistic Preference Relations S. Alonso Department of Software Engineering University of Granada, 18071, Granada, Spain; salonso@decsai.ugr.es, F.J. Cabrerizo

More information

SOLVING FUZZY LINEAR SYSTEMS OF EQUATIONS

SOLVING FUZZY LINEAR SYSTEMS OF EQUATIONS ROMAI J, 4, 1(2008), 207 214 SOLVING FUZZY LINEAR SYSTEMS OF EQUATIONS A Panahi, T Allahviranloo, H Rouhparvar Depart of Math, Science and Research Branch, Islamic Azad University, Tehran, Iran panahi53@gmailcom

More information

Practical implementation of possibilistic probability mass functions

Practical implementation of possibilistic probability mass functions Soft Computing manuscript No. (will be inserted by the editor) Practical implementation of possibilistic probability mass functions Leen Gilbert, Gert de Cooman 1, Etienne E. Kerre 2 1 Universiteit Gent,

More information

Fuzzy Function: Theoretical and Practical Point of View

Fuzzy Function: Theoretical and Practical Point of View EUSFLAT-LFA 2011 July 2011 Aix-les-Bains, France Fuzzy Function: Theoretical and Practical Point of View Irina Perfilieva, University of Ostrava, Inst. for Research and Applications of Fuzzy Modeling,

More information

Numerical Method for Solving Fuzzy Nonlinear Equations

Numerical Method for Solving Fuzzy Nonlinear Equations Applied Mathematical Sciences, Vol. 2, 2008, no. 24, 1191-1203 Numerical Method for Solving Fuzzy Nonlinear Equations Javad Shokri Department of Mathematics, Urmia University P.O.Box 165, Urmia, Iran j.shokri@mail.urmia.ac.ir

More information

Integer programming: an introduction. Alessandro Astolfi

Integer programming: an introduction. Alessandro Astolfi Integer programming: an introduction Alessandro Astolfi Outline Introduction Examples Methods for solving ILP Optimization on graphs LP problems with integer solutions Summary Introduction Integer programming

More information

Approximating models based on fuzzy transforms

Approximating models based on fuzzy transforms Approximating models based on fuzzy transforms Irina Perfilieva University of Ostrava Institute for Research and Applications of Fuzzy Modeling 30. dubna 22, 701 03 Ostrava 1, Czech Republic e-mail:irina.perfilieva@osu.cz

More information

Compenzational Vagueness

Compenzational Vagueness Compenzational Vagueness Milan Mareš Institute of information Theory and Automation Academy of Sciences of the Czech Republic P. O. Box 18, 182 08 Praha 8, Czech Republic mares@utia.cas.cz Abstract Some

More information

Where are we? Operations on fuzzy sets (cont.) Fuzzy Logic. Motivation. Crisp and fuzzy sets. Examples

Where are we? Operations on fuzzy sets (cont.) Fuzzy Logic. Motivation. Crisp and fuzzy sets. Examples Operations on fuzzy sets (cont.) G. J. Klir, B. Yuan, Fuzzy Sets and Fuzzy Logic: Theory and Applications, Prentice-Hall, chapters -5 Where are we? Motivation Crisp and fuzzy sets alpha-cuts, support,

More information

Regular finite Markov chains with interval probabilities

Regular finite Markov chains with interval probabilities 5th International Symposium on Imprecise Probability: Theories and Applications, Prague, Czech Republic, 2007 Regular finite Markov chains with interval probabilities Damjan Škulj Faculty of Social Sciences

More information

S-MEASURES, T -MEASURES AND DISTINGUISHED CLASSES OF FUZZY MEASURES

S-MEASURES, T -MEASURES AND DISTINGUISHED CLASSES OF FUZZY MEASURES K Y B E R N E T I K A V O L U M E 4 2 ( 2 0 0 6 ), N U M B E R 3, P A G E S 3 6 7 3 7 8 S-MEASURES, T -MEASURES AND DISTINGUISHED CLASSES OF FUZZY MEASURES Peter Struk and Andrea Stupňanová S-measures

More information

GENERAL AGGREGATION OPERATORS ACTING ON FUZZY NUMBERS INDUCED BY ORDINARY AGGREGATION OPERATORS

GENERAL AGGREGATION OPERATORS ACTING ON FUZZY NUMBERS INDUCED BY ORDINARY AGGREGATION OPERATORS Novi Sad J. Math. Vol. 33, No. 2, 2003, 67 76 67 GENERAL AGGREGATION OPERATORS ACTING ON FUZZY NUMBERS INDUCED BY ORDINARY AGGREGATION OPERATORS Aleksandar Takači 1 Abstract. Some special general aggregation

More information

On (Weighted) k-order Fuzzy Connectives

On (Weighted) k-order Fuzzy Connectives Author manuscript, published in "IEEE Int. Conf. on Fuzzy Systems, Spain 2010" On Weighted -Order Fuzzy Connectives Hoel Le Capitaine and Carl Frélicot Mathematics, Image and Applications MIA Université

More information

Practical implementation of possibilistic probability mass functions

Practical implementation of possibilistic probability mass functions Practical implementation of possibilistic probability mass functions Leen Gilbert Gert de Cooman Etienne E. Kerre October 12, 2000 Abstract Probability assessments of events are often linguistic in nature.

More information

Chapter 1 Similarity Based Reasoning Fuzzy Systems and Universal Approximation

Chapter 1 Similarity Based Reasoning Fuzzy Systems and Universal Approximation Chapter 1 Similarity Based Reasoning Fuzzy Systems and Universal Approximation Sayantan Mandal and Balasubramaniam Jayaram Abstract In this work, we show that fuzzy inference systems based on Similarity

More information

A SECOND ORDER STOCHASTIC DOMINANCE PORTFOLIO EFFICIENCY MEASURE

A SECOND ORDER STOCHASTIC DOMINANCE PORTFOLIO EFFICIENCY MEASURE K Y B E R N E I K A V O L U M E 4 4 ( 2 0 0 8 ), N U M B E R 2, P A G E S 2 4 3 2 5 8 A SECOND ORDER SOCHASIC DOMINANCE PORFOLIO EFFICIENCY MEASURE Miloš Kopa and Petr Chovanec In this paper, we introduce

More information

Solution of Fuzzy Maximal Flow Network Problem Based on Generalized Trapezoidal Fuzzy Numbers with Rank and Mode

Solution of Fuzzy Maximal Flow Network Problem Based on Generalized Trapezoidal Fuzzy Numbers with Rank and Mode International Journal of Engineering Research and Development e-issn: 2278-067X, p-issn: 2278-800X, www.ijerd.com Volume 9, Issue 7 (January 2014), PP. 40-49 Solution of Fuzzy Maximal Flow Network Problem

More information

arxiv: v1 [math.ra] 22 Dec 2018

arxiv: v1 [math.ra] 22 Dec 2018 On generating of idempotent aggregation functions on finite lattices arxiv:1812.09529v1 [math.ra] 22 Dec 2018 Michal Botur a, Radomír Halaš a, Radko Mesiar a,b, Jozef Pócs a,c a Department of Algebra and

More information

DIFFERENCE METHODS FOR FUZZY PARTIAL DIFFERENTIAL EQUATIONS

DIFFERENCE METHODS FOR FUZZY PARTIAL DIFFERENTIAL EQUATIONS COMPUTATIONAL METHODS IN APPLIED MATHEMATICS, Vol.2(2002), No.3, pp.233 242 c Institute of Mathematics of the National Academy of Sciences of Belarus DIFFERENCE METHODS FOR FUZZY PARTIAL DIFFERENTIAL EQUATIONS

More information

Rough Approach to Fuzzification and Defuzzification in Probability Theory

Rough Approach to Fuzzification and Defuzzification in Probability Theory Rough Approach to Fuzzification and Defuzzification in Probability Theory G. Cattaneo and D. Ciucci Dipartimento di Informatica, Sistemistica e Comunicazione Università di Milano Bicocca, Via Bicocca degli

More information

Downloaded from iors.ir at 10: on Saturday May 12th 2018 Fuzzy Primal Simplex Algorithms for Solving Fuzzy Linear Programming Problems

Downloaded from iors.ir at 10: on Saturday May 12th 2018 Fuzzy Primal Simplex Algorithms for Solving Fuzzy Linear Programming Problems Iranian Journal of Operations esearch Vol 1, o 2, 2009, pp68-84 Fuzzy Primal Simplex Algorithms for Solving Fuzzy Linear Programming Problems ezam Mahdavi-Amiri 1 Seyed Hadi asseri 2 Alahbakhsh Yazdani

More information

Generalized Triangular Fuzzy Numbers In Intuitionistic Fuzzy Environment

Generalized Triangular Fuzzy Numbers In Intuitionistic Fuzzy Environment International Journal of Engineering Research Development e-issn: 2278-067X, p-issn : 2278-800X, www.ijerd.com Volume 5, Issue 1 (November 2012), PP. 08-13 Generalized Triangular Fuzzy Numbers In Intuitionistic

More information

Solving Fuzzy PERT Using Gradual Real Numbers

Solving Fuzzy PERT Using Gradual Real Numbers Solving Fuzzy PERT Using Gradual Real Numbers Jérôme FORTIN a, Didier DUBOIS a, a IRIT/UPS 8 route de Narbonne, 3062, Toulouse, cedex 4, France, e-mail: {fortin, dubois}@irit.fr Abstract. From a set of

More information

SOLVING FUZZY LINEAR SYSTEMS BY USING THE SCHUR COMPLEMENT WHEN COEFFICIENT MATRIX IS AN M-MATRIX

SOLVING FUZZY LINEAR SYSTEMS BY USING THE SCHUR COMPLEMENT WHEN COEFFICIENT MATRIX IS AN M-MATRIX Iranian Journal of Fuzzy Systems Vol 5, No 3, 2008 pp 15-29 15 SOLVING FUZZY LINEAR SYSTEMS BY USING THE SCHUR COMPLEMENT WHEN COEFFICIENT MATRIX IS AN M-MATRIX M S HASHEMI, M K MIRNIA AND S SHAHMORAD

More information

Problems in Linear Algebra and Representation Theory

Problems in Linear Algebra and Representation Theory Problems in Linear Algebra and Representation Theory (Most of these were provided by Victor Ginzburg) The problems appearing below have varying level of difficulty. They are not listed in any specific

More information

Fuzzy if-then rules for modeling interdependences in FMOP problems

Fuzzy if-then rules for modeling interdependences in FMOP problems Fuzzy if-then rules for modeling interdependences in FMOP problems Christer Carlsson Institute for Advanced Management Systems Research, Åbo Akademi University, DataCity A 3210, SF-20520 Åbo, Finland Robert

More information

Comparison of Fuzzy Operators for IF-Inference Systems of Takagi-Sugeno Type in Ozone Prediction

Comparison of Fuzzy Operators for IF-Inference Systems of Takagi-Sugeno Type in Ozone Prediction Comparison of Fuzzy Operators for IF-Inference Systems of Takagi-Sugeno Type in Ozone Prediction Vladimír Olej and Petr Hájek Institute of System Engineering and Informatics, Faculty of Economics and Administration,

More information

Interval based Uncertain Reasoning using Fuzzy and Rough Sets

Interval based Uncertain Reasoning using Fuzzy and Rough Sets Interval based Uncertain Reasoning using Fuzzy and Rough Sets Y.Y. Yao Jian Wang Department of Computer Science Lakehead University Thunder Bay, Ontario Canada P7B 5E1 Abstract This paper examines two

More information

Australian Journal of Basic and Applied Sciences, 5(9): , 2011 ISSN Fuzzy M -Matrix. S.S. Hashemi

Australian Journal of Basic and Applied Sciences, 5(9): , 2011 ISSN Fuzzy M -Matrix. S.S. Hashemi ustralian Journal of Basic and pplied Sciences, 5(9): 2096-204, 20 ISSN 99-878 Fuzzy M -Matrix S.S. Hashemi Young researchers Club, Bonab Branch, Islamic zad University, Bonab, Iran. bstract: The theory

More information

Construction of Interval-valued Fuzzy Preference Relations from Ignorance Functions and Fuzzy Preference Relations. Application to Decision Making

Construction of Interval-valued Fuzzy Preference Relations from Ignorance Functions and Fuzzy Preference Relations. Application to Decision Making Construction of Interval-valued Fuzzy Preference Relations from Ignorance Functions and Fuzzy Preference Relations. Application to Decision Making Edurne Barrenechea a, Javier Fernandez a, Miguel Pagola

More information

Numerical Method for Fuzzy Partial Differential Equations

Numerical Method for Fuzzy Partial Differential Equations Applied Mathematical Sciences, Vol. 1, 2007, no. 27, 1299-1309 Numerical Method for Fuzzy Partial Differential Equations M. Afshar Kermani 1 and F. Saburi Department of Mathematics Science and Research

More information

Fuzzy control systems. Miklós Gerzson

Fuzzy control systems. Miklós Gerzson Fuzzy control systems Miklós Gerzson 2016.11.24. 1 Introduction The notion of fuzziness: type of car the determination is unambiguous speed of car can be measured, but the judgment is not unambiguous:

More information

Support Vector Machines

Support Vector Machines Support Vector Machines Ryan M. Rifkin Google, Inc. 2008 Plan Regularization derivation of SVMs Geometric derivation of SVMs Optimality, Duality and Large Scale SVMs The Regularization Setting (Again)

More information

The Generalized Likelihood Uncertainty Estimation methodology

The Generalized Likelihood Uncertainty Estimation methodology CHAPTER 4 The Generalized Likelihood Uncertainty Estimation methodology Calibration and uncertainty estimation based upon a statistical framework is aimed at finding an optimal set of models, parameters

More information

2 so Q[ 2] is closed under both additive and multiplicative inverses. a 2 2b 2 + b

2 so Q[ 2] is closed under both additive and multiplicative inverses. a 2 2b 2 + b . FINITE-DIMENSIONAL VECTOR SPACES.. Fields By now you ll have acquired a fair knowledge of matrices. These are a concrete embodiment of something rather more abstract. Sometimes it is easier to use matrices,

More information

On the Continuity and Convexity Analysis of the Expected Value Function of a Fuzzy Mapping

On the Continuity and Convexity Analysis of the Expected Value Function of a Fuzzy Mapping Journal of Uncertain Systems Vol.1, No.2, pp.148-160, 2007 Online at: www.jus.org.uk On the Continuity Convexity Analysis of the Expected Value Function of a Fuzzy Mapping Cheng Wang a Wansheng Tang a

More information

Research Article Deriving Weights of Criteria from Inconsistent Fuzzy Comparison Matrices by Using the Nearest Weighted Interval Approximation

Research Article Deriving Weights of Criteria from Inconsistent Fuzzy Comparison Matrices by Using the Nearest Weighted Interval Approximation Advances in Operations Research Volume 202, Article ID 57470, 7 pages doi:0.55/202/57470 Research Article Deriving Weights of Criteria from Inconsistent Fuzzy Comparison Matrices by Using the Nearest Weighted

More information

Intuitionistic Fuzzy Numbers and It s Applications in Fuzzy Optimization Problem

Intuitionistic Fuzzy Numbers and It s Applications in Fuzzy Optimization Problem Intuitionistic Fuzzy Numbers and It s Applications in Fuzzy Optimization Problem HASSAN MISHMAST NEHI a, and HAMID REZA MALEKI b a)department of Mathematics, Sistan and Baluchestan University, Zahedan,

More information

Multivariate Measures of Positive Dependence

Multivariate Measures of Positive Dependence Int. J. Contemp. Math. Sciences, Vol. 4, 2009, no. 4, 191-200 Multivariate Measures of Positive Dependence Marta Cardin Department of Applied Mathematics University of Venice, Italy mcardin@unive.it Abstract

More information

A Consistency Based Procedure to Estimate Missing Pairwise Preference Values

A Consistency Based Procedure to Estimate Missing Pairwise Preference Values A Consistency Based Procedure to Estimate Missing Pairwise Preference Values S. Alonso F. Chiclana F. Herrera E. Herrera-Viedma J. Alcalá-Fernádez C. Porcel Abstract In this paper, we present a procedure

More information

A Complete Description of Comparison Meaningful Functions

A Complete Description of Comparison Meaningful Functions A Complete Description of Comparison Meaningful Functions Jean-Luc MARICHAL Radko MESIAR Tatiana RÜCKSCHLOSSOVÁ Abstract Comparison meaningful functions acting on some real interval E are completely described

More information

An Evaluation of the Reliability of Complex Systems Using Shadowed Sets and Fuzzy Lifetime Data

An Evaluation of the Reliability of Complex Systems Using Shadowed Sets and Fuzzy Lifetime Data International Journal of Automation and Computing 2 (2006) 145-150 An Evaluation of the Reliability of Complex Systems Using Shadowed Sets and Fuzzy Lifetime Data Olgierd Hryniewicz Systems Research Institute

More information

ASSOCIATIVE n DIMENSIONAL COPULAS

ASSOCIATIVE n DIMENSIONAL COPULAS K Y BERNETIKA VOLUM E 47 ( 2011), NUMBER 1, P AGES 93 99 ASSOCIATIVE n DIMENSIONAL COPULAS Andrea Stupňanová and Anna Kolesárová The associativity of n-dimensional copulas in the sense of Post is studied.

More information

Intersection and union of type-2 fuzzy sets and connection to (α 1, α 2 )-double cuts

Intersection and union of type-2 fuzzy sets and connection to (α 1, α 2 )-double cuts EUSFLAT-LFA 2 July 2 Aix-les-Bains, France Intersection and union of type-2 fuzzy sets and connection to (α, α 2 )-double cuts Zdenko Takáč Institute of Information Engineering, Automation and Mathematics

More information

Analysis-3 lecture schemes

Analysis-3 lecture schemes Analysis-3 lecture schemes (with Homeworks) 1 Csörgő István November, 2015 1 A jegyzet az ELTE Informatikai Kar 2015. évi Jegyzetpályázatának támogatásával készült Contents 1. Lesson 1 4 1.1. The Space

More information

From Satisfiability to Linear Algebra

From Satisfiability to Linear Algebra From Satisfiability to Linear Algebra Fangzhen Lin Department of Computer Science Hong Kong University of Science and Technology Clear Water Bay, Kowloon, Hong Kong Technical Report August 2013 1 Introduction

More information

On stability in possibilistic linear equality systems with Lipschitzian fuzzy numbers

On stability in possibilistic linear equality systems with Lipschitzian fuzzy numbers On stability in possibilistic linear equality systems with Lipschitzian fuzzy numbers Robert Fullér rfuller@abo.fi Abstract Linear equality systems with fuzzy parameters and crisp variables defined by

More information

Revision of the Kosiński s Theory of Ordered Fuzzy Numbers

Revision of the Kosiński s Theory of Ordered Fuzzy Numbers axioms Article Revision of the Kosiński s Theory of Ordered Fuzzy Numbers Krzysztof Piasecki ID Department of Investment and Real Estate, Poznan University of Economics and Business, al. Niepodleglosci

More information

Fuzzy Distance Measure for Fuzzy Numbers

Fuzzy Distance Measure for Fuzzy Numbers Australian Journal of Basic Applied Sciences, 5(6): 58-65, ISSN 99-878 Fuzzy Distance Measure for Fuzzy Numbers Hamid ouhparvar, Abdorreza Panahi, Azam Noorafkan Zanjani Department of Mathematics, Saveh

More information

Fuzzy arithmetic and its application in problems of optimal choice

Fuzzy arithmetic and its application in problems of optimal choice RESEARCH ARTICLE www.iera.com OPEN ACCESS Fuzzy arithmetic and its application in problems of optimal choice Vadim N. Romanov Saint-Petersburg, Russia Corresponding Author:Vadim N. Romanov ABSTRACT The

More information

Adrian I. Ban and Delia A. Tuşe

Adrian I. Ban and Delia A. Tuşe 18 th Int. Conf. on IFSs, Sofia, 10 11 May 2014 Notes on Intuitionistic Fuzzy Sets ISSN 1310 4926 Vol. 20, 2014, No. 2, 43 51 Trapezoidal/triangular intuitionistic fuzzy numbers versus interval-valued

More information

A Generalized Decision Logic in Interval-set-valued Information Tables

A Generalized Decision Logic in Interval-set-valued Information Tables A Generalized Decision Logic in Interval-set-valued Information Tables Y.Y. Yao 1 and Qing Liu 2 1 Department of Computer Science, University of Regina Regina, Saskatchewan, Canada S4S 0A2 E-mail: yyao@cs.uregina.ca

More information

Fuzzy efficiency: Multiplier and enveloping CCR models

Fuzzy efficiency: Multiplier and enveloping CCR models Available online at http://ijim.srbiau.ac.ir/ Int. J. Industrial Mathematics (ISSN 28-5621) Vol. 8, No. 1, 216 Article ID IJIM-484, 8 pages Research Article Fuzzy efficiency: Multiplier and enveloping

More information

Key Renewal Theory for T -iid Random Fuzzy Variables

Key Renewal Theory for T -iid Random Fuzzy Variables Applied Mathematical Sciences, Vol. 3, 29, no. 7, 35-329 HIKARI Ltd, www.m-hikari.com https://doi.org/.2988/ams.29.9236 Key Renewal Theory for T -iid Random Fuzzy Variables Dug Hun Hong Department of Mathematics,

More information

EEE 8005 Student Directed Learning (SDL) Industrial Automation Fuzzy Logic

EEE 8005 Student Directed Learning (SDL) Industrial Automation Fuzzy Logic EEE 8005 Student Directed Learning (SDL) Industrial utomation Fuzzy Logic Desire location z 0 Rot ( y, φ ) Nail cos( φ) 0 = sin( φ) 0 0 0 0 sin( φ) 0 cos( φ) 0 0 0 0 z 0 y n (0,a,0) y 0 y 0 z n End effector

More information

ALGEBRAIC OPERATIONS ON FUZZY NUMBERS USING OF LINEAR FUNCTIONS. Jae Deuk Myung

ALGEBRAIC OPERATIONS ON FUZZY NUMBERS USING OF LINEAR FUNCTIONS. Jae Deuk Myung Kangweon-Kyungki Math. Jour. 11 (2003), No. 1, pp. 1 7 ALGEBRAIC OPERATIONS ON FUZZY NUMBERS USING OF LINEAR FUNCTIONS Jae Deuk Myung Abstract. In this paper, we introduce two types of algebraic operations

More information

type-2 fuzzy sets, α-plane, intersection of type-2 fuzzy sets, union of type-2 fuzzy sets, fuzzy sets

type-2 fuzzy sets, α-plane, intersection of type-2 fuzzy sets, union of type-2 fuzzy sets, fuzzy sets K Y B E R N E T I K A V O L U M E 4 9 ( 2 3 ), N U M B E R, P A G E S 4 9 6 3 ON SOME PROPERTIES OF -PLANES OF TYPE-2 FUZZY SETS Zdenko Takáč Some basic properties of -planes of type-2 fuzzy sets are investigated

More information

Developing a Data Envelopment Analysis Methodology for Supplier Selection in the Presence of Fuzzy Undesirable Factors

Developing a Data Envelopment Analysis Methodology for Supplier Selection in the Presence of Fuzzy Undesirable Factors Available online at http://ijim.srbiau.ac.ir Int. J. Industrial Mathematics (ISSN 2008-5621) Vol. 4, No. 3, Year 2012 Article ID IJIM-00225, 10 pages Research Article Developing a Data Envelopment Analysis

More information

New approach to solve symmetric fully fuzzy linear systems

New approach to solve symmetric fully fuzzy linear systems Sādhanā Vol. 36, Part 6, December 2011, pp. 933 940. c Indian Academy of Sciences New approach to solve symmetric fully fuzzy linear systems 1. Introduction P SENTHILKUMAR and G RAJENDRAN Department of

More information

Lecture 1: Introduction & Fuzzy Control I

Lecture 1: Introduction & Fuzzy Control I Lecture 1: Introduction & Fuzzy Control I Jens Kober Robert Babuška Knowledge-Based Control Systems (SC42050) Cognitive Robotics 3mE, Delft University of Technology, The Netherlands 12-02-2018 Lecture

More information

Finitely Valued Indistinguishability Operators

Finitely Valued Indistinguishability Operators Finitely Valued Indistinguishability Operators Gaspar Mayor 1 and Jordi Recasens 2 1 Department of Mathematics and Computer Science, Universitat de les Illes Balears, 07122 Palma de Mallorca, Illes Balears,

More information

Transformation-Based Constraint-Guided Generalised Modus Ponens

Transformation-Based Constraint-Guided Generalised Modus Ponens Transformation-Based Constraint-Guided Generalised Modus Ponens Michaël Blot, Marie-Jeanne Lesot, Marcin Detyniecki Sorbonne Universités, LIP6, UPMC Univ Paris 06, CNRS, UMR 7606 F-75005, Paris, France

More information

Inverse arithmetic operators for fuzzy intervals

Inverse arithmetic operators for fuzzy intervals Inverse arithmetic operators for fuzzy intervals Reda Boukezzoula Sylvie Galichet Laurent Foulloy LISTIC, Université de Savoie LISTIC, Université de Savoie LISTIC, Université de Savoie Domaine Universitaire,

More information

MATHEMATICS OF DATA FUSION

MATHEMATICS OF DATA FUSION MATHEMATICS OF DATA FUSION by I. R. GOODMAN NCCOSC RDTE DTV, San Diego, California, U.S.A. RONALD P. S. MAHLER Lockheed Martin Tactical Defences Systems, Saint Paul, Minnesota, U.S.A. and HUNG T. NGUYEN

More information

FURTHER DEVELOPMENT OF CHEBYSHEV TYPE INEQUALITIES FOR SUGENO INTEGRALS AND T-(S-)EVALUATORS

FURTHER DEVELOPMENT OF CHEBYSHEV TYPE INEQUALITIES FOR SUGENO INTEGRALS AND T-(S-)EVALUATORS K Y BERNETIK VOLUM E 46 21, NUMBER 1, P GES 83 95 FURTHER DEVELOPMENT OF CHEBYSHEV TYPE INEQULITIES FOR SUGENO INTEGRLS ND T-S-EVLUTORS Hamzeh gahi, Radko Mesiar and Yao Ouyang In this paper further development

More information

From Crisp to Fuzzy Constraint Networks

From Crisp to Fuzzy Constraint Networks From Crisp to Fuzzy Constraint Networks Massimiliano Giacomin Università di Brescia Dipartimento di Elettronica per l Automazione Via Branze 38, I-25123 Brescia, Italy giacomin@ing.unibs.it Abstract. Several

More information

Sup-t-norm and inf-residuum are a single type of relational equations

Sup-t-norm and inf-residuum are a single type of relational equations International Journal of General Systems Vol. 00, No. 00, February 2011, 1 12 Sup-t-norm and inf-residuum are a single type of relational equations Eduard Bartl a, Radim Belohlavek b Department of Computer

More information

Application of Fuzzy Measure and Fuzzy Integral in Students Failure Decision Making

Application of Fuzzy Measure and Fuzzy Integral in Students Failure Decision Making IOSR Journal of Mathematics (IOSR-JM) e-issn: 2278-5728, p-issn: 2319-765X. Volume 10, Issue 6 Ver. III (Nov - Dec. 2014), PP 47-53 Application of Fuzzy Measure and Fuzzy Integral in Students Failure Decision

More information

A new Approach to Drawing Conclusions from Data A Rough Set Perspective

A new Approach to Drawing Conclusions from Data A Rough Set Perspective Motto: Let the data speak for themselves R.A. Fisher A new Approach to Drawing Conclusions from Data A Rough et Perspective Zdzisław Pawlak Institute for Theoretical and Applied Informatics Polish Academy

More information

Using ranking functions in multiobjective fuzzy linear programming 1

Using ranking functions in multiobjective fuzzy linear programming 1 Fuzzy Sets and Systems 111 (2000) 47 53 www.elsevier.com/locate/fss Using ranking functions in multiobective fuzzy linear programming 1 J.M. Cadenas a;, J.L. Verdegay b a Departamento de Informatica, Inteligencia

More information

Fuzzy Difference Equations in Finance

Fuzzy Difference Equations in Finance International Journal of Scientific and Innovative Mathematical Research (IJSIMR) Volume 2, Issue 8, August 2014, PP 729-735 ISSN 2347-307X (Print) & ISSN 2347-3142 (Online) www.arcjournals.org Fuzzy Difference

More information

Intuitionistic Fuzzy Sets - An Alternative Look

Intuitionistic Fuzzy Sets - An Alternative Look Intuitionistic Fuzzy Sets - An Alternative Look Anna Pankowska and Maciej Wygralak Faculty of Mathematics and Computer Science Adam Mickiewicz University Umultowska 87, 61-614 Poznań, Poland e-mail: wygralak@math.amu.edu.pl

More information

1.5 F15 O Brien. 1.5: Linear Equations and Inequalities

1.5 F15 O Brien. 1.5: Linear Equations and Inequalities 1.5: Linear Equations and Inequalities I. Basic Terminology A. An equation is a statement that two expressions are equal. B. To solve an equation means to find all of the values of the variable that make

More information

Solving Fuzzy Nonlinear Equations by a General Iterative Method

Solving Fuzzy Nonlinear Equations by a General Iterative Method 2062062062062060 Journal of Uncertain Systems Vol.4, No.3, pp.206-25, 200 Online at: www.jus.org.uk Solving Fuzzy Nonlinear Equations by a General Iterative Method Anjeli Garg, S.R. Singh * Department

More information

Uncertain Systems are Universal Approximators

Uncertain Systems are Universal Approximators Uncertain Systems are Universal Approximators Zixiong Peng 1 and Xiaowei Chen 2 1 School of Applied Mathematics, Central University of Finance and Economics, Beijing 100081, China 2 epartment of Risk Management

More information

Vector Spaces. Addition : R n R n R n Scalar multiplication : R R n R n.

Vector Spaces. Addition : R n R n R n Scalar multiplication : R R n R n. Vector Spaces Definition: The usual addition and scalar multiplication of n-tuples x = (x 1,..., x n ) R n (also called vectors) are the addition and scalar multiplication operations defined component-wise:

More information

A new approach to solve fuzzy system of linear equations by Homotopy perturbation method

A new approach to solve fuzzy system of linear equations by Homotopy perturbation method Journal of Linear and Topological Algebra Vol. 02, No. 02, 2013, 105-115 A new approach to solve fuzzy system of linear equations by Homotopy perturbation method M. Paripour a,, J. Saeidian b and A. Sadeghi

More information

Estimation of the Muskingum routing coefficients by using fuzzy regression

Estimation of the Muskingum routing coefficients by using fuzzy regression European Water 57: 33-40, 207. 207 E.W. Publications Estimation of the Muskingum routing coefficients by using fuzzy regression M. Spiliotis * and L. Garrote 2 Democritus University of Thrace, School of

More information

Aumann-Shapley Values on a Class of Cooperative Fuzzy Games

Aumann-Shapley Values on a Class of Cooperative Fuzzy Games Journal of Uncertain Systems Vol.6, No.4, pp.27-277, 212 Online at: www.jus.org.uk Aumann-Shapley Values on a Class of Cooperative Fuzzy Games Fengye Wang, Youlin Shang, Zhiyong Huang School of Mathematics

More information

APPLYING SIGNED DISTANCE METHOD FOR FUZZY INVENTORY WITHOUT BACKORDER. Huey-Ming Lee 1 and Lily Lin 2 1 Department of Information Management

APPLYING SIGNED DISTANCE METHOD FOR FUZZY INVENTORY WITHOUT BACKORDER. Huey-Ming Lee 1 and Lily Lin 2 1 Department of Information Management International Journal of Innovative Computing, Information and Control ICIC International c 2011 ISSN 1349-4198 Volume 7, Number 6, June 2011 pp. 3523 3531 APPLYING SIGNED DISTANCE METHOD FOR FUZZY INVENTORY

More information

Fuzzy Eigenvectors of Real Matrix

Fuzzy Eigenvectors of Real Matrix Fuzzy Eigenvectors of Real Matrix Zengfeng Tian (Corresponding author Composite Section, Junior College, Zhejiang Wanli University Ningbo 315101, Zhejiang, China Tel: 86-574-8835-7771 E-mail: bbtianbb@126.com

More information

CS , Fall 2011 Assignment 2 Solutions

CS , Fall 2011 Assignment 2 Solutions CS 94-0, Fall 20 Assignment 2 Solutions (8 pts) In this question we briefly review the expressiveness of kernels (a) Construct a support vector machine that computes the XOR function Use values of + and

More information

Fuzzy Sets, Fuzzy Logic, and Fuzzy Systems II

Fuzzy Sets, Fuzzy Logic, and Fuzzy Systems II Fuzzy Sets, Fuzzy Logic, and Fuzzy Systems II SSIE 617 Fall 2008 Radim BELOHLAVEK Dept. Systems Sci. & Industrial Eng. Watson School of Eng. and Applied Sci. Binghamton University SUNY Radim Belohlavek

More information

Research Article Generalized Fuzzy Soft Expert Set

Research Article Generalized Fuzzy Soft Expert Set Journal of Applied Mathematics Volume 2012 Article ID 328195 22 pages doi:1155/2012/328195 Research Article Generalized Fuzzy Soft Expert Set Ayman A. Hazaymeh Ismail B. Abdullah Zaid T. Balkhi and Rose

More information

Statistical Hypotheses Testing in the Fuzzy Environment

Statistical Hypotheses Testing in the Fuzzy Environment Journal of Uncertain Systems Vol.6, No.3, pp.86-99, 22 Online at: www.jus.org.uk Statistical Hypotheses Testing in the Fuzzy Environment Mohammad Ghasem Akbari Department of Statistics, Faculty of Sciences,

More information

Copulas with given diagonal section: some new results

Copulas with given diagonal section: some new results Copulas with given diagonal section: some new results Fabrizio Durante Dipartimento di Matematica Ennio De Giorgi Università di Lecce Lecce, Italy 73100 fabrizio.durante@unile.it Radko Mesiar STU Bratislava,

More information

Directional Monotonicity of Fuzzy Implications

Directional Monotonicity of Fuzzy Implications Acta Polytechnica Hungarica Vol. 14, No. 5, 2017 Directional Monotonicity of Fuzzy Implications Katarzyna Miś Institute of Mathematics, University of Silesia in Katowice Bankowa 14, 40-007 Katowice, Poland,

More information

Support Vector Machines

Support Vector Machines Support Vector Machines Sridhar Mahadevan mahadeva@cs.umass.edu University of Massachusetts Sridhar Mahadevan: CMPSCI 689 p. 1/32 Margin Classifiers margin b = 0 Sridhar Mahadevan: CMPSCI 689 p.

More information

The evacuation plan design under uncertain times of transportation

The evacuation plan design under uncertain times of transportation Urban Transport XVI 83 The evacuation plan design under uncertain times of transportation J. Janáček Faculty of Management Science and Informatics, University of Žilina, Slovak Republic Abstract When the

More information

AN EFFICIENT APPROACH FOR SOLVING A WIDE CLASS OF FUZZY LINEAR PROGRAMMING PROBLEMS

AN EFFICIENT APPROACH FOR SOLVING A WIDE CLASS OF FUZZY LINEAR PROGRAMMING PROBLEMS AN EFFICIENT APPROACH FOR SOLVING A WIDE CLASS OF FUZZY LINEAR PROGRAMMING PROBLEMS A. V. KAMYAD N. HASSANZADEH J. CHAJI Abstract. In this paper we are going to introduce an efficient approach for solving

More information

Fuzzy Sets. Mirko Navara navara/fl/fset printe.pdf February 28, 2019

Fuzzy Sets. Mirko Navara   navara/fl/fset printe.pdf February 28, 2019 The notion of fuzzy set. Minimum about classical) sets Fuzzy ets Mirko Navara http://cmp.felk.cvut.cz/ navara/fl/fset printe.pdf February 8, 09 To aviod problems of the set theory, we restrict ourselves

More information

Fuzzy Order Statistics based on α pessimistic

Fuzzy Order Statistics based on α pessimistic Journal of Uncertain Systems Vol.10, No.4, pp.282-291, 2016 Online at: www.jus.org.uk Fuzzy Order Statistics based on α pessimistic M. GH. Akbari, H. Alizadeh Noughabi Department of Statistics, University

More information

Fuzzy Modal Like Approximation Operations Based on Residuated Lattices

Fuzzy Modal Like Approximation Operations Based on Residuated Lattices Fuzzy Modal Like Approximation Operations Based on Residuated Lattices Anna Maria Radzikowska Faculty of Mathematics and Information Science Warsaw University of Technology Plac Politechniki 1, 00 661

More information

Normalized priority vectors for fuzzy preference relations

Normalized priority vectors for fuzzy preference relations Normalized priority vectors for fuzzy preference relations Michele Fedrizzi, Matteo Brunelli Dipartimento di Informatica e Studi Aziendali Università di Trento, Via Inama 5, TN 38100 Trento, Italy e mail:

More information

Decomposition and Intersection of Two Fuzzy Numbers for Fuzzy Preference Relations

Decomposition and Intersection of Two Fuzzy Numbers for Fuzzy Preference Relations S S symmetry Article Decomposition and Intersection of Two Fuzzy Numbers for Fuzzy Preference Relations Hui-Chin Tang ID Department of Industrial Engineering and Management, National Kaohsiung University

More information