Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Dynamical system wikipedia , lookup
List of important publications in mathematics wikipedia , lookup
System of polynomial equations wikipedia , lookup
Numerical continuation wikipedia , lookup
Bra–ket notation wikipedia , lookup
Mathematics of radio engineering wikipedia , lookup
Available online at www.ispacs.com/jfsva Volume 2012, Year 2012 Article ID jfsva-00117, 11 pages doi:10.5899/2012/jfsva-00117 Research Article Solving Fully Fuzzy Linear System of Equations in General Form R. Ezzati 1∗ , S. Khezerloo 1 , A. Yousefzadeh 1 (1) Department of Mathematics, Karaj Branch, Islamic Azad University, karaj, Iran. c R. Ezzati, S. Khezerloo and A. Yousefzadeh. This is an open access article distributed Copyright 2012 ⃝ under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Abstract In this work, we propose an approach for computing the positive solution of a fully fuzzy linear system where the coefficient matrix is a fuzzy n×n matrix. To do this, we use arithmetic operations on fuzzy numbers that introduced by Kaffman in [18] and convert the fully fuzzy linear system into two n × n and 2n × 2n crisp linear systems. If the solutions of these linear systems don’t satisfy in positive fuzzy solution condition, we introduce the constrained least squares problem to obtain optimal fuzzy vector solution by applying the ranking function in given fully fuzzy linear system. Using our proposed method, the fully fuzzy linear system of equations always has a solution. Finally, we illustrate the efficiency of proposed method by solving some numerical examples. Keywords : Fuzzy linear system; Fuzzy number; Ranking Function; Fuzzy number vector solution. 1 Introduction There are many linear equation systems in many areas of science and engineering. According to Moore [20] exact numerical data might be unrealistic, but there could be considered uncertain data as more aspects of a real word problem. Fuzzy data is being used as a natural way to describe uncertain data. Fuzzy concept was introduced by Zadeh [23, 24]. So, we need to solve those linear systems in which all parameters , or some of them are fuzzy numbers. Friedman et al. [13, 14] applied an embedding method for solving Ax = b, where A is a nonsingular crisp matrix. Moreover, they transformed the n × n fuzzy linear system to a 2n × 2n real linear system. As a result, they solved the 2n × 2n real linear system using the inverse matrix. There are many other numerical methods for solving ∗ Corresponding author. Email address: [email protected], Tel: +989123618518 1 2 Journal of Fuzzy Set Valued Analysis fuzzy linear systems such as Jacobi, Gauss-Seidel, Adomiam decomposition method and SOR iterative method [1, 2, 3, 4]. Cheng in [7] introduced a ranking function and Ghanbari et al. in [15] obtained fuzzy linear equation system by ranking functions so that the matrix A is non-fuzzy. Dehgan in [8, 9, 10] introduced full fuzzy system in which b and A are fuzzy vector and fuzzy matrix, respectively. Then Kumar in [19] obtained exact solution of fully fuzzy linear system by solving a linear programming. In this article, we will obtain positive exact solution or positive approximate solution of fully fuzzy linear system by using ranking function. This paper is organized as follows: In Section 2, the basic concept of fuzzy number operation is brought. In Section 3, the main Section of the paper, a new approach based on ranking functions for solving fully fuzzy linear system, is suggested. The proposed idea is illustrated by solving some numerical examples in the Section 4. Finally conclusion is drawn in Section 5. 2 Preliminaries In this section, we give some basic definitions of fuzzy numbers. Definition 2.1. A fuzzy number is a fuzzy set µÃ : R −→ I = [0, 1] which satisfies: 1. µÃ is upper semi continuous. 2. µÃ (x) = 0outside some interval [c, d]. 3. There are real numbers a, b : c ≤ a ≤ b ≤ d for which a. µÃ (x) is monotonic increasing on [c, a], b. µÃ (x) is monotonic decreasing on [b, d], c. µÃ (x) = 1, a ≤ x ≤ b. The set of all fuzzy numbers (as given by Definition (2.1)) is denoted by F (ℜ)1 . An alternative definition or parametric form of a fuzzy number which yields the same F (ℜ)1 is given by Kaleva [16]. Definition 2.2. A fuzzy number à is LR−type if there exit L (for left) and R (for right) and scalars α > 0 , β > 0 with L( a−x α ), x ≤ α, µÃ (x) = (2.1) R( x−a ), x ≥ a β where L and R are strictly decreasing functions defined on [0, 1] and satisfy the conditions: L(0) = R(0) = 1, L(1) = R(1) = 0, 0 < L(x) < 1, (2.2) 0 < R(x) < 1, x ̸= 0, 1. The mean value of Ã, m, is a real number, and α, β are called the left and right spreads, respectively. à is denoted by (m, α, β)LR . Journal of Fuzzy Set Valued Analysis 3 Definition 2.3. M̃ = (m, α, β)LR is a triangular fuzzy number if L = R = max(0, 1 − x). 1 We denote the set of triangular fuzzy numbers √ by F (ℜ)T . In this paper, we use ranking function R(e x) = x20 + y02 introduced by Cheng [7] which is based on centroid point where for any triangular fuzzy number x e = (m, α, β), x0 and y0 are as follows: x0 = m + 31 (β − α), ) ( y0 = 13 6m+(β−α) 4m+(β−α) . Remark 2.1. According to Eq. (2.1) and definition (2.3), throughout the paper, we assume that all the fuzzy numbers are triangular in the form (m − α, m, m + β). Arithmetic operations between two triangular fuzzy numbers, defined on universal set of real numbers ℜ, are reviewed [18]. Theorem 2.1. [17], Let M̃ = (a, b, c) and Ñ = (x, y, z) are two arbitrary triangular fuzzy numbers and λ > 0 is a real number. Then (1) M̃ ⊕ Ñ = (a + x, b + y, c + z), (2) −M̃ = (−c, −b, −a), (3) M̃ ⊖ Ñ = (a − z, b − y, c − x), (4) Let M̃ = (a, b, c) be any triangular fuzzy number and Ñ = (x, y, z) be a non-negative triangular fuzzy number, then (ax, by, cz), a ≥ 0, (az, by, cz), a < 0, c ≥ 0, M̃ ⊗ Ñ = (az, by, cx), c < 0. Definition 2.4. A matrix à = [ãij ]ni,j=1 is called a fuzzy matrix if for all i and j, ãij ∈ e will be positive (negative) and denoted by A e > 0 (A e < 0) if for all i and j, F (ℜ)1T . A e e aij > 0 (e aij < 0). Clearly, N = (a, b, c) is positive (negative), if and only if a > 0 (c < 0). Non-negative and non-positive fuzzy matrices will be defined similarly. Definition 2.5. A vector X̃ = (x̃1 , . . . , x̃n )T , denoted by X̃ ∈ F (ℜ)nT , is called a fuzzy numbers vector, where x̃i ∈ F (ℜ)1T , i = 1, . . . , n. 3 Solutions of fully fuzzy linear system by ranking function Definition 3.1. The n × n linear system (e a11 ⊗ x e1 ) ⊕ (e a12 ⊗ x e2 ) ⊕ · · · ⊕ (e a1n ⊗ x en ) = eb1 (e a21 ⊗ x e1 ) ⊕ (e a22 ⊗ x e2 ) ⊕ · · · ⊕ (e a2n ⊗ x en ) = eb2 .. . (e an1 ⊗ x e1 ) ⊕ (e an2 ⊗ x e2 ) ⊕ · · · ⊕ (e ann ⊗ x en ) = ebn , (3.3) 4 Journal of Fuzzy Set Valued Analysis or in its matrix form, e⊗x A e = eb, (3.4) is called a fully fuzzy linear system of equations (FFLSE) where the coefficient matrix à = [ãij ]ni,j=1 is a fuzzy matrix and b̃ = [b̃1 , . . . , b̃n ]T is a fuzzy number vector and the fuzzy number vector x e is the unknown to be found. Definition 3.2. We call x̃ ∈ F (ℜ)n a solution of à ⊗ x̃ = b̃ with respect to the ranking function R if and only if we have, { By = b2 , R(ãTi x̃) = R(b̃i ), i = 1, . . . , n where the crisp linear system By = b2 is the 1−cut or mean value of system à ⊗ x̃ = b̃ and ãTi is the i−th row of A. Notation 3.1. Set e aij = (aij , bij , cij ), ebi = (di1 , di2 , di3 ), x ei = (xi , yi , zi ), A= d1 = ( ( aij di1 ) n×n , B= n×1 , d2 = ) ( ( bij di2 ) n×n ) n×1 , C= ( , d3 = ( cij di3 ) n×n , ) n×1 . Notation 3.2. We break ( + up ) the matrix A (into− two ) n × n matrices such that their + − addition is A. Let A = aij and A = aij , where, n×n { + aij = + aij 0 aij ≥ 0 n×n { , aij < 0 − aij = 0 aij ≥ 0 aij aij < 0 . − Then A + up the matrix C into two n × n matrices, similarly. (A + =) A. We also break ( − ) + − c c ij ij Let C = , and C = be n × n matrices, where, { + cij = cij 0 cij ≥ 0 cij < 0 { , − cij = 0 cij ≥ 0 cij cij < 0 , Theorem 3.1. The system (3.4) with the multiplication defined in Theorem 2.1 is equivalent to )( ) ( ) ( + A A− x d1 = , C− C+ z d3 (3.5) By = d2 , where x ej = (xj , yj , zj ), j = 1, . . . , n, are nonnegative and x = (x1 , . . . , xn )T , y = (y1 , . . . , yn )T , z = (z1 , . . . , zn )T . 5 Journal of Fuzzy Set Valued Analysis Proof. Using the multiplication defined in Theorem 2.1, we have e⊗x (A e)i = = = + = ∑n aij j=1 (e ⊗x ej ) ∑n ∑ ∑ ⊗ (xj , yj , zj ) ∑ j,aij ≥0 (aij xj , bij yj , cij zj ) + j,aij <0,cij ≥0 (aij zj , bij yj , cij zj ) j=1 (aij , bij , cij ) j,cij <0 (aij zj , bij yj , cij xj ) ( ∑ j,aij ≥0 aij xj + ∑ j,aij <0 aij zj , ∑n j=1 bij yj , ∑ j,cij ≥0 cij zj ( A+ A− + ∑ j,cij <0 cij xj ) ) are invertiable, C− C+ and (e x1 , . . . , x en )T given by x ej = (xj , yj , zj ), j = 1, . . . , n, be the solution of (3.5). Then this solution is a nonnegative fuzzy exact solution of (3.4) if it satisfies 0 ≤ xi ≤ yi ≤ zi , i = 1, . . . , n. Proposition 3.1. Suppose that the matrices B and M = Proof. Using Theorem 3.1, the proof is clear. e⊗x According to Theorem 3.1, we know that the system (3.4), A e = eb, is equivalent to (A+ x + A− z, By, C − x + C + z) = eb = (d1 , d2 , d3 ). (3.6) Now, suppose that one of the matrices B and M (or both) be singular or the solution of (3.6) doesn’t hold in the condition 0 ≤ xi ≤ yi ≤ zi , i = 1, . . . , n. In this case, we may try to find approximate solution for (3.4). For this purpose, we consider the following constrained least squares problem : min s.t. ∑n ∑ + A− z)i , nj=1 bij yj , (C − x + C − z)i ) − R(b̃i ))2 ∑ ∑ ∑n i = 1, . . . , n, j=1 bij yj − j,aij ≥0 aij xj − j,aij <0 aij zj ≥ 0, ∑ ∑ ∑n i = 1, . . . , n, j,cij ≥0 cij zj + j,cij <0 cij xj − j=1 bij yj ≥ 0, ∑n i = 1, . . . , n, j=1 bij yj = di2 , + i=1 (R((A x xi ≥ 0, i = 1, . . . , n, yi − xi ≥ 0, i = 1, . . . , n, zi − yi ≥ 0, i = 1, . . . , n. (3.7) Remark 3.1. Let for all i and j, 0 ∈ supp(ãij ), i.e., for all i and j, aij < 0 and cij ≥ 0. Using Theorem (3.1), the system (3.4) is transformed to the following form: n n n ∑ ∑ ∑ aij zj , bij yj , cij zj = (di1 , di2 , di3 ), i = 1, . . . , n. (3.8) j=1 j=1 j=1 6 Journal of Fuzzy Set Valued Analysis Also, we can rewrite (3.8) in the matrix form as: (Az, By, Cz) = (d1 , d2 , d3 ). (3.9) If A, B and C are non-singular matrices, then we obtain: (z, y, z) = (A−1 d1 , B −1 d2 , C −1 d3 ). (3.10) It is clear that, the vector y is unique, but there are two vectors for z, i.e., z1 = A−1 d1 and z2 = C −1 d3 . In order to find the fuzzy number vector solution, we have the following cases: (I) Let one of the two vector z1 or z2 satisfy in the conditions 0 ≤ y ≤ z1 or 0 ≤ y ≤ z2 , respectively. For example without the loss generality, suppose z1 satisfies in the condition 0 ≤ y ≤ z1 , so we set x = y, and therefore we obtain (y, y, z1 ) as a positive fuzzy number vector solution of FFLSE. (II) Let both of the two vector z1 and z2 satisfy in the conditions 0 ≤ y ≤ z1 and 0 ≤ y ≤ z2 , respectively. So we set x = y and therefore FFLSE hasn’t unique solution and two positive fuzzy number vector Se1 = (y, y, z1 ) and Se2 = (y, y, z2 ) are eSe1 and A eSe2 , we use a distance function it’s solution. For a comparison between A e e e to measure the closeness of the vectors AS1 and ASe2 to eb. We will use the Ming et al. [19] metric proposed for triangular fuzzy numbers. If x̃ = (x, αx , βx ) and ỹ = (y, αy , βy ) are two triangular fuzzy numbers, then Ming et al. [19] introduced the distance function, ) 1( 4(x − y)2 + (αy − αx )2 + (βy − βx )2 + (x − y)(βy + αy − βx − αx ), 2 (3.11) and for two LR fuzzy vectors x̃ = (x̃1 , . . . , x̃n ) and ỹ = (ỹ1 , . . . , ỹn ) defined the distance between x̃ and ỹ to be, D22 (x̃, ỹ) = Dn2 (x̃, ỹ) = n ∑ D22 (x̃i , ỹi ). (3.12) i=1 eSe1 , eb) and Dn (A eSe2 , eb) and choose the nearest solution Using (3.12), we obtain Dn (A e to b. (III) Let vectors y or z1 and z2 don’t hold in condition y ≥ 0 or y ≤ z1 and y ≤ z2 , respectively. In this case, we solve the constrained least squares problem (3.7) to obtain the approximation of positive fuzzy number vector solution of (3.3). 7 Journal of Fuzzy Set Valued Analysis 4 Numerical examples Example 4.1. Consider the following system: (1, 2, 5) ⊗ (x1 , y1 , z1 ) ⊕ (3, 4, 4) ⊗ (x2 , y2 , z2 ) (2, 3, 5) ⊗ (x1 , y1 , z1 ) ⊕ (0, 1, 11) ⊗ (x2 , y2 , z2 ) ⊕(0, 1, 2) ⊗ (x3 , y3 , z3 ) = (19, 68, 115) ⊕(4, 5, 6) ⊗ (x3 , y3 , z3 ) (2, 5, 7) ⊗ (x1 , y1 , z1 ) ⊕ (4, 6, 6) ⊗ (x2 , y2 , z2 ) So, we must solve two following 1 3 0 2 0 4 2 4 5 0 0 0 0 0 0 0 0 0 (4.13) = (30, 77, 261) ⊕(5, 7, 10) ⊗ (x3 , y3 , z3 ) = (61, 167, 253) systems: 0 0 0 0 0 0 0 0 0 5 4 2 5 11 6 7 6 10 x1 x2 x3 z1 z2 z3 = 19 30 61 115 261 253 2 4 1 y1 68 3 1 5 y2 = 77 5 6 7 y3 167 and Using (3.6), we have (x1 , y1 , z1 ) (x2 , y2 , z2 ) (x3 , y3 , z3 ) (1, 2, 7) = (6, 12, 14) (7, 10, 12) as a positive fuzzy number vector solution of (4.13). Example 4.2. Consider the following system: (5.6621, 6.2945, 7.2520) ⊗ (x1 , y1 , z1 ) (8.0183, 8.1185, 9.0807) ⊗ (x1 , y1 , z1 ) ⊕(−7.7388, −7.4603, −7.3027) ⊗ (x2 , y2 , z2 ) = (6.5421, 24.0383, 35.5858) = (100.8751, 117.8106, 134.2153) So, we must solve two following systems: 5.6621 0 8.0183 7.7206 0 −7.7388 0 0 0 −7.3027 0 0 7.2520 0 9.0807 8.2675 and ( 6.2945 −7.4603 8.1185 8.2675 (4.14) ⊕(7.7206, 8.2675, 8.2675) ⊗ (x2 , y2 , z2 ) )( y1 y2 6.5421 x1 x2 100.8751 z1 = 35.5858 134.2153 z2 ) ( = 24.0383 117.8106 ) 8 Journal of Fuzzy Set Valued Analysis Using (3.7), we have [ (x1 , y1 , z1 ) ] [ (0.1002069, 9.569978, 16.56075) ] = (x2 , y2 , z2 ) (0, 4.852342, 10.14861) as a positive fuzzy number vector solution of (4.14) but the proposed method by kumar [17] has not solution. Also, the objective value of (3.7) is 0.6812963 × 10−20 . Example 4.3. Consider the following system: (−10, −7, −4) ⊗ (x1 , y1 , z1 ) ⊕ (−7, −5, −3) ⊗ (x2 , y2 , z2 ) (−6, −4, −2) ⊗ (x1 , y1 , z1 ) ⊕ (−4, −3, −1) ⊗ (x2 , y2 , z2 ) (−8, −7, −3) ⊗ (x1 , y1 , z1 ) ⊕ (−11, −1, −1) ⊗ (x2 , y2 , z2 ) ⊕(−5, −2, −1) ⊗ (x3 , y3 , z3 ) = (−36, −36, −33) ⊕(−8, −5, −5) ⊗ (x3 , y3 , z3 ) = (−26, −25, −23) ⊕(−3, −2, −1) ⊗ (x3 , y3 , z3 ) = (−44, −21, −19) (4.15) Using (3.6), we have (x1 , y1 , z1 ) (x2 , y2 , z2 ) (x3 , y3 , z3 ) (1, 2, 4) = (3, 4, 5) (1, 1, 2) as a positive fuzzy number vector solution of (4.15). Example 4.4. Consider the following system: (4.9950, 5.8441, 6.5873) ⊗ (x1 , y1 , z1 ) (8.2559, 9.1898, 9.5821) ⊗ (x1 , y1 , z1 ) ⊕(2.4361, 3.1148, 3.7703) ⊗ (x2 , y2 , z2 ) = (34.0822, 42.2359, 48.4583) ⊕(−10.1400, −9.2500, −9.1180) ⊗ (x2 , y2 , z2 ) (4.16) = (50.2106, 61.9286, 64.9654) Using (3.7), we have [ (x1 , y1 , z1 ) ] [ = (x2 , y2 , z2 ) (0.2578402, 7.058067, 12.14679) ] (0.5868982 × 10−2 , 0.3171481, 0.3245293) as a positive fuzzy number vector solution of (4.16) but the proposed method by kumar [17] has not solution. Also, the objective value is of (3.7) 0.4089455 × 10−26 . Example 4.5. Consider the following system: { (−2, 3, 4) ⊗ (x1 , y1 , z1 ) ⊕ (−3, −2, −1) ⊗ (x2 , y2 , z2 ) = (5, 21, 43) (−1, 1, 2) ⊗ (x1 , y1 , z1 ) ⊕ (1, 3, 4) ⊗ (x2 , y2 , z2 ) = (−6, 14, 34) (4.17) 9 Journal of Fuzzy Set Valued Analysis Using (3.7), we have [ (x1 , y1 , z1 ) ] [ (0, 8.272727, 46.22885) ] = (x2 , y2 , z2 ) (0, 1.909091, 14.81923) as a positive fuzzy number vector solution of (4.17) but the proposed method by kumar [17] has not solution. Example 4.6. Consider the following system: (−1, 0, 3) ⊗ (x1 , y1 , z1 ) ⊕ (−3, 1, 2) ⊗ (x2 , y2 , z2 ) ⊕(−2, 0, 0) ⊗ (x3 , y3 , z3 ) = (−31, 6, 24) (−2, −1, 2) ⊗ (x1 , y1 , z1 ) ⊕ (−1, 2, 5) ⊗ (x2 , y2 , z2 ) ⊕(−5, 3, 4) ⊗ (x3 , y3 , z3 ) (−4, 1, 2) ⊗ (x1 , y1 , z1 ) ⊕ (−2, 5, 7) ⊗ (x2 , y2 , z2 ) Using (3.6), we have (x1 , y1 , z1 ) (x2 , y2 , z2 ) (x3 , y3 , z3 ) (4.18) = (−43, 38, 75) ⊕(−1, 2, 3) ⊗ (x3 , y3 , z3 ) = (−32, 59, 69) (7, 7, 8) = (6, 6, 10) (11, 11, 12) as a positive fuzzy number vector solution of (4.18). 5 Conclusion In this paper, we used arithmetic operations on fuzzy numbers that introduced by Kaffman [18] and found the positive fuzzy number vector solution for the fully fuzzy linear system of equations. First, we converted the FFLSE into two crisp systems and if both coefficient matrices be non-singular then we solved the crisp systems. By allocating the vector solution of 2n × 2n system to the vector solution of n × n system, we obtained the vector solution of the FFLSE. If the vector solutions were acceptable (Solution be positive and satisfies the fuzzy number condition), we set that the fuzzy numbers vector solution of FFLSE. Else we solved the constrained least squares problem to found the optimal fuzzy numbers vector solution. In comparison with the proposed method by kumar [17], our proposed method is very useful and the FFLSE always has positive fuzzy numbers vector solution and we showed this capability by solving some numerical examples. References [1] T. Allahviranloo, Numerical methods for fuzzy system of linear equations, Applied Mathematics and Computation, 155 (2004) 493-502. http://dx.doi.org/10.1016/S0096-3003(03)00793-8. 10 Journal of Fuzzy Set Valued Analysis [2] T. Allahviranloo, Successive over relaxation iterative method for fuzzy system of linear equations, Applied Mathematics and Computation, 162 (2005) 189-196. http://dx.doi.org/10.1016/j.amc.2003.12.085. [3] T. Allahviranloo, The Adomian decomposition method for fuzzy system of linear equations, Applied Mathematics and Computation, 163 (2005) 553-563. http://dx.doi.org/10.1016/j.amc.2004.02.020. [4] T. Allahviranloo, E. Ahmady, N. Ahmady and Kh. Shams Alketaby, Block Jacobi twostage method with Gauss-Sidel inner iterations for fuzzy system of linear equations, Applied Mathematics and Computation, 175 (2006) 1217-1228. http://dx.doi.org/10.1016/j.amc.2005.08.047. [5] T. Allahviranloo, N. Mikaeilvand, Non Zero Solutions Of The Fully Fuzzy Linear Systems, Appl. Comput. Math, 10 (2) 271-282. [6] J.J. Buckley and Y. Qu, Solving system of linear fuzzy equations, Fuzzy Sets and Systems, 43 (1991) 33-43. http://dx.doi.org/10.1016/0165-0114(91)90019-M. [7] C. Cheng, A new approach for ranking fuzzy numbers by distance method, Fuzzy Sets Syst. 95 (1998) 307-317. http://dx.doi.org/10.1016/S0165-0114(96)00272-2. [8] M. Dehghan, B. Hashemi and M. Ghatee, Solution of the fully fuzzy linear systems using iterative techniques, Chaos Solutions and Fractals 34 (2007) 316-336. http://dx.doi.org/10.1016/j.chaos.2006.03.085. [9] M. Dehghan and B. Hashemi, Solution of the fully fuzzy linear systems using the decomposition procedure, Applied Mathematics and Computation, 182 (2006) 15681580. http://dx.doi.org/10.1016/j.amc.2006.05.043. [10] M.Dehghan, B.Hashemi, M.Ghati, Computational methods for solving fully fuzzy linear systems, Appl Math and Comput 179 (2006) 328-343. http://dx.doi.org/10.1016/j.amc.2005.11.124. [11] M. Dehghan, B. Hashemi, Iterative solution of fuzzy linear systems, Applied Mathematics and Computation, in press. http://dx.doi.org/10.1016/j.amc.2005.07.03. [12] D. Dubois, H. Prade, Fuzzy Sets and Systems; Theory and Applications, Academic Press, New York, 1980. [13] M. Friedman, M. Ma, A. Kandel, Fuzzy linear systems, Proc. IEEE Int. Conf. Syst., Man, Cybernet. 1 (1996) 14-17. October 336-338. [14] M. Friedman, M. Ma, A. Kandel, Fuzzy linear systems, Fuzzy Sets Syst. 96 (1998) 201-209. http://dx.doi.org/10.1016/S0165-0114(96)00270-9. Journal of Fuzzy Set Valued Analysis 11 [15] R. Ghanbari , N.Mahdavi-Amiri, New solution of linear systems using ranking functions and ABS algorithms, Appl. Math. Comput. 34 (2010) 3363-3375. [16] O. Kaleva, Fuzzy differential equations, Fuzzy Sets and Systems 24 (1987) 301-317. http://dx.doi.org/10.1016/0165-0114(87)90029-7. [17] A. Kumar, J.Kaur, P.Singh, A new method for solving fully fuzzy linear programming problems, Appl. Math. Comput. 35 (2011) 817-823. [18] A. Kaufmann, M.M. Gupta, Introduction to Fuzzy Arithmetic Theory and Applications, Van Nostrand Reinhold, New York,1985. [19] M. Ming, M. Friedman, A. Kandel, General fuzzy least squares, Fuzzy Sets and Systems 88 (1997) 107118. http://dx.doi.org/10.1016/S0165-0114(96)00051-6. [20] R.E. Moore, Methods and Applications of Interval Analysis, SIAM, Philadelphia, 1979. http://dx.doi.org/10.1137/1.9781611970906. [21] H. T. Nguyen, E.A. Wallker, A First Course in Fuzzy Logic, Chapman , Hall, 2000. [22] H.J. Zimmermann, Fuzzy Set Theory and its Applications, third ed., Kluwer Academic , Norwell, 1996. [23] L.A. Zadeh, A fuzzy-set-theoretic interpretation of linguistic hedges, Journal of Cybernetics 2 (1972) 4-34. http://dx.doi.org/10.1080/01969727208542910. [24] L.A. Zadeh, The concept of the linguistic variable and its application to approximate reasoning, Information Sciences 8 (1975) 199-249. http://dx.doi.org/10.1016/0020-0255(75)90036-5.