TELK OMNIKA T elecommunication, Computing, Electr onics and Contr ol V ol. 19, No. 2, April 2021, pp. 583 598 ISSN: 1693-6930, accredited First Grade by K emenristekdikti, No: 21/E/KPT/2018 DOI: 10.12928/TELK OMNIKA.v19i2.18023 r 583 A modification of fuzzy arithmetic operators f or solving near -zer o fully fuzzy matrix equation W . S. W . Daud 1 , N. Ahmad 2 , G. Malkawi 3 1 Institute of Engineering Mathematics, Uni v ersiti Malaysia Perlis, Perlis, Malaysia 1,2 School of Quantitati v e Sciences, Uni v ersiti Utara Malaysia, K edah, Malaysia 3 Higher Colle ges of T echnology , Ab u Dhabi AlAin Men's Colle ge, 17155, United Arab Emirates Article Inf o Article history: Recei v ed Jun 26, 2020 Re vised Sep 8, 2020 Accepted Sep 16, 2020 K eyw ords: Control system engineering Fully fuzzy matrix equation Fully fuzzy linear system Fuzzy arithmetic operators Near -zero fuzzy numbers ABSTRA CT Matrix equations ha v e its o wn important in the field of control system engineering particularly in the stability analysis of linear control systems and the reduction of nonlinear control system models. There a re certain conditions where the classical matrix equation are not well equipped to handle the uncertainty problems such as during the process of stability analysis and reduction in control system engineering. In this study , an algorithm is de v eloped for solving fully fuzzy matrix equation particularly for ~ A ~ X ~ B ~ X = ~ C , where the coef ficients of the equation are in near -zero fuzzy numbers. By modifying the e xisting fuzzy multiplication arithmetic operators, the proposed algorithm e xceeds the positi v e restriction to allo w the near -zero fuzzy numbers as the coef ficients. Besides that, a ne w fuzzy subtraction arithmetic operator has a lso been proposed as the e xisting operator f ailed to satisfy the both sides of the near -zero fully fuzzy matri x equation. Subsequently , Kroneck er product and V ec -operator are adapted with the modified fuzzy arithmetic operator in order to transform the fully fuzzy matrix equation to a fully fuzzy linear system. On top of that, a ne w associated linear system is de v eloped to obtain the final solution. A numerical e xample and the v erification of the solution are pres ented to demonstrate the proposed algorithm. This is an open access article under the CC BY -SA license . Corresponding A uthor: W . S. W . Daud Institute of Engineering Mathematics Uni v ersiti Malaysia Perlis Perlis, Malaysia Email: wsuhana@unimap.edu.my 1. INTR ODUCTION There are man y types of matrix equations that ha v e been modelled in v arious applications [1] particularly in control system engineering [2, 3]. Basically , control system engineering is used to design the feedback loops system [4]. The e xample applications that related to the feedback loops systems are medical imaging acquisition system [5], image restoration [6], model reduction [7], signal processing [8] and stochastic control [9]. According to [10], matrix equation plays the role as an equation solv er for t he control system model. In dealing wit h an y real applications, it is possible that an y uncertainty conditions could occur , for e xample, if there e xist an y conflicting requirements and instability of the en vironmental conditions during the system process. If there is an y e xistence of noise or unnecessary elements during the process, it w ould also J ournal homepage: http://journal.uad.ac.id/inde x.php/TELK OMNIKA Evaluation Warning : The document was created with Spire.PDF for Python.
584 r ISSN: 1693-6930 distract the system [11]. In this case, the e xisting matrix equations sometimes are not well equipped to handle those conditions. Therefore, one of the approaches that can be tak en is to adapt the fuzzy numbers as the coef ficients of the matrix equation [12]. In the past fe w years, man y researchers proposed their algorithms in solving matrix equations with parameters in fuzzy numbers. This equation is kno wn as the fully fuzzy matrix equation (FFME). Otadi and Mosleh [13] are the pioneers in this field, who has applied linear programming technique to obtain a positi v e solution for arbitrary FFME, ~ A ~ X m = ~ B m . Apart from that, there is a study which has e xtended the algorithm used in solving the fully fuzzy linear system (FFLS) to solv e the FFME ~ A ~ X ~ B = ~ C [14]. Subsequently , in 2015, Shang et al. [15] proposed their algorithm in solving fully fuzzy Sylv ester matrix equation (FFSE), ~ A ~ X + ~ X ~ B = ~ C by applying the arithmetic multiplication operator , which has been pre viously proposed in Dehghan et al. [16]. On the other hand, Malka wi et al. [17] ha v e proposed an algorithm which of fers f aster computational compared to Shang et al. [15]. Whi le in 2020, Elsayed et al. [18] carried out a study in solving the FFME of ~ A ~ X + ~ X ~ B = ~ C , which considering the entries of the equation are in trapezoidal fuzzy numbers. In this paper , we are propose an algorithm to solv e the FFME of ~ A ~ X ~ B ~ X = ~ C (1) considering the fuzzy coef ficient ~ A = ( ~ a ij ) m n or ~ B = ( ~ b ij ) n n is a near -zero fuzzy number , while ~ C = ( ~ c ij ) m n is an arbitrary fuzzy coef ficient and ~ X = ( ~ x ij ) m n is the solution of the FFME. This equation has been pre viously solv ed by Daud et al. [19] in 2018. Unfortunately the algorithm proposed is only limited to non-singular and positi v e fuzzy matrices. This limitation has moti v ated us to construct an algorithm to solv e the (1) without an y restrictions. Moreo v er , in real-life applications, the coef ficients of the FFME can either be positi v e, ne g ati v e or near -zero fuzzy numbers. In de v eloping the algorithm, the e xisting fuzzy multiplication arithmetic operators are modified as the e xisting operators introduced by [17] and [20] are not applicable to perform the multiplication in v olving near -zero fuzzy numbers. Besides that, a ne w fuzzy subtraction operation is also de v eloped in solving the FFME, since the e xisting operator is inadequated to subtract a near -zero fuzzy number to a positi v e fuzzy number . Subsequently , the modified fuzzy arithmetic operator is adapted with the Kroneck er product and V ec -operator in con v erting the FFME to a simpler form of equation, which is a fully fuzzy linear system (FFLS). Later on, the solution is obtained by means of associated linear system (ALS) which has been established based on the modified fuzzy multiplication arithmetic operator . The remaining part of the paper proceeds as follo ws. In Section 2, some preliminaries on the fuzzy numbers and Kroneck er product are sho wn. Then in Section 3, the theoretical foundation which supports the de v eloped algorithm are established. In Section 4, the de v eloped algorithm for solving the FFME of (1) is sho wn. Mo ving on, a numerical e xample and v erification of the solution are illustrated in Section 5. Finally , the conclusion is dra wn in Section 6. 2. PRELIMIN ARIES 2.1. Fundamental concepts of matrix and set theory The fundamental concept of matrix theory is important in order to solv e the matrix equations. Some fundamentals of matrix theory are defined in the follo wing: Definition 1. [21] Let N be a 3 3 bloc k matrix, suc h that N = 0 @ A B C D E F G H I 1 A ; (2) then, j N j = det  A B D E C F I 1 G H det [ I ] = det A C I 1 G B C I 1 H D F I 1 G E F I 1 H det [ I ] (3) TELK OMNIKA T elecommun Comput El Control, V ol. 19, No. 2, April 2021 : 583 598 Evaluation Warning : The document was created with Spire.PDF for Python.
TELK OMNIKA T elecommun Comput El Control r 585 Remark 1. F or the bloc k matrix 3 3 suc h that P = 0 @ A B C D E F G H I 1 A : (4) Clearly that based on Definition 1, the determinant P is given as follows: j P j = det  E F H I D G A 1 B C det [ A ] = det E D A 1 B F D A 1 C H GA 1 B I GA 1 C det [ A ] (5) Definition 2. [22] Let A and B be sets. The union of A and B is the set of A [ B = f x : x 2 A or x 2 B g . 2.2. Theory of fuzzy numbers The follo wing definition describing the theory of fuzzy numbers has been introduced since 1965 by Zadeh [23]. Definition 3. Let X be a nonempty set. A fuzzy set ~ A in X is c har acterized by its member ship function ~ A : X ! [0 ; 1] (6) and ~ A ( x ) r epr esents the de gr ee of member ship of element x in fuzzy set ~ A for eac h x 2 X . In this study , the representation of fuzzy numbers is based on the triangular fuzzy numbers. 2.2.1. T riangular fuzzy number Definition 4. A fuzzy number ~ M = ( m; ; ) is said to be a triangular fuzzy number (TFN), if its member ship function is given by: ~ M ( x ) = 8 > < > : 1 m x ; m x m; > 0 ; 1 x m ; m x m + ; > 0 ; 0 ; otherwise : (7) In this case, m is the mean v alue of ~ M , whereas and are the right and left spreads, respecti v ely . Definition 5. A fuzzy number ~ M = ( m; ; ) is called as an arbitr ary fuzzy number wher e it may be positive , ne gative or near zer o whic h can be classified as follows: ~ M is a positive(ne gative) fuzzy number if f m 0 ( + m 0) . ~ M is a zer o fuzzy number if ( m = 0 ; ; = 0) . ~ M is a near zer o fuzzy number if f m 0 + m . The follo wing definitions describe some important arithmetic operations of TFN [20]. Definition 6. The arithmetic oper ations of two TFN, ~ M = ( m; ; ) and ~ N = ( n; ; ) , ar e as follows: i. Addition: ~ M ~ N = ( m; ; ) ( n; ; ) = ( m + n; + ; + ) : (8) ii. Opposite: ~ M = ( m; ; ) = ( m; ; ) : (9) iii. Subtr action: ( m; ; ) ( n; ; ) = ( m; ; ) ( n; ; ) = ( m; ; ) ( n; ; ) = ( m n; + ; + ) : (10) A modification of fuzzy arithmetic oper ator s for solving near -zer o... (W . S. W . Daud) Evaluation Warning : The document was created with Spire.PDF for Python.
586 r ISSN: 1693-6930 iv . Multiplication: If ~ M > 0 and ~ N > 0 , then ~ M ~ N = ( m; ; ) ( n; ; ) = ( mn; m + n ; m + n ) (11) If ~ M < 0 and ~ N > 0 , then ~ M ~ N = ( m; ; ) ( n; ; ) = ( mn; n m ; n m ) (12) If ~ M < 0 and ~ N < 0 , then ~ M ~ N = ( m; ; ) ( n; ; ) = ( mn; n m ; n m ) (13) Based on the multiplication arithmetic operator in (11) to (13), there is no operator applicable for a near -zero fuzzy number . This is because a near -zero fuzzy number cannot be defined in the form of ( m; ; ) , unlik e a positi v e or ne g ati v e fuzzy number could. Therefore, a ne w form of multiplication arithmetic operator has been introduced by [24] which adapted the system of min-max function. Definition 7. [24] The pr oduct of two fuzzy number s ~ M = ( m; ; ) and ~ N = ( n; ; ) , can be defined as ~ M ~ N = ( mn; f 1 ; f 2 ) (14) wher e f 1 = mn Min (( m )( n ) ; ( m )( n + )) , f 2 = Max (( m + )( n ) ; ( m + )( n + )) mn: The operator as gi v en in (14) is basically has been init iated based on [25] and [26]. In implementing the multiplication, fe w times multiplication and comparison are needed, to obtain the minimum and maximum v alues. Besi des that, the opera tor is only compatible for positi v e fuzzy number ~ N as stated in the follo wing Theorem 1. Theor em 1. [24] Consider an arbitr ary fuzzy number ~ M = ( m; ; ) and a positive fuzzy number ~ N = ( n; ; ) , i. If ~ M is positive , then the following inequalities ar e satisfied: 0 ( m )( n ) ( m )( n + ) ; (15) 0 ( m + )( n ) ( m + )( n + ) (16) ii. If ~ M is ne gative , then the following inequalities ar e satisfied: 0 ( m )( n ) ( m )( n + ) ; (17) 0 ( m + )( n ) ( m + )( n + ) (18) iii. If ~ M is near zer o, then the inequalities in (16) and (17) ar e satisfied. 2.3. Fundamental concepts of fuzzy Kr oneck er pr oducts and fuzzy V ec -operator Kroneck er products and V ec -operator are the important tools in solving matrix equations. The definitions and theorems of the fuzzy Kroneck er products and fuzzy V ec -operator , are pro vided as follo ws: Definition 8. [17] Let ~ A = ( ~ a ij ) m n and ~ B = ( ~ b ij ) p q be fuzzy matrices. Fuzzy Kr onec k er pr oduct is r epr esented as ~ A k ~ B , wher e ~ A k ~ B = 0 B B B @ ~ a 11 ~ B ~ a 12 ~ B : : : ~ a 1 n ~ B ~ a 21 ~ B ~ a 22 ~ B : : : ~ a 2 n ~ B . . . . . . . . . . . . ~ a m 1 ~ B ~ a m 2 ~ B : : : ~ a mn ~ B 1 C C C A = [ ~ a ij ~ B ] ( mp ) ( nq ) (19) TELK OMNIKA T elecommun Comput El Control, V ol. 19, No. 2, April 2021 : 583 598 Evaluation Warning : The document was created with Spire.PDF for Python.
TELK OMNIKA T elecommun Comput El Control r 587 Definition 9. [17] V ec -oper ator of a fuzzy matrix is a linear tr ansformation that con verts the fuzzy matrix of ~ C = ( ~ c 1 ; ~ c 2 ; :::; ~ c n ) into a column vector as V ec ( ~ C ) = 0 B B B @ ~ c 1 ~ c 2 . . . ~ c n 1 C C C A : (20) Theor em 2. [17] If ~ A = ( ~ a ij ) m m is a fuzzy matrix, and ~ U = ( ~ u ij ) p p is a unitary fuzzy matrix defined as ~ U = 0 B B B @ (1 ; 0 ; 0) (0 ; 0 ; 0) : : : (0 ; 0 ; 0) (0 ; 0 ; 0) (1 ; 0 ; 0) : : : (0 ; 0 ; 0) . . . . . . . . . . . . (0 ; 0 ; 0) (0 ; 0 ; 0) : : : (1 ; 0 ; 0) 1 C C C A ; (21) then i. ~ A ~ U = ~ U ~ A = ~ A ii. ~ U T = ~ U . Definition 10. [17] Let A = ( a ij ) m m , B = ( b ij ) n n and X = ( x ij ) m n , then i. V ec [ ~ A ~ X ] = [ ~ U n k ~ A ] V ec ( ~ X ) ii. V ec [ ~ X ~ B ] = [ ~ B T k ~ U m ] V ec ( ~ X ) iii. V ec [ ~ A ~ X ~ B ] = [ ~ B T k ~ A ] V ec ( ~ X ) iv . V ec ( ~ X ) = [ ~ U ] V ec ( ~ X ) 3. THEORETICAL DEVELOPMENT This section demonstrates the establishment of the theoretical foundations which in v olv ed some theorems, definitions and corallaries. There are four sections presented, consist of the introduction of a ne w near -zero positi v e subtraction operator , a modification of arithmetic multiplication operator , some related properties of FFME ~ A ~ X ~ B ~ X = ~ C and also the construction of an associated linear systems. 3.1. Near -zer o positi v e subtraction operator Theor em 3. Let ~ M = ( m; ; ) be a near -zer o fuzzy number and ~ N = ( n; ; ) is a positive fuzzy number . The subtr action of ~ M and ~ N is given by ~ M np ~ N = ( m n; + ; ) : (22) wher e > . Pr oof. Let < , then < 0 , which means that the spread v alue of is ne g ati v e. This is violated since it is al w ays positi v e, as mentioned in Definition 4 . Thus, > . This ne w operator is kno wn as a Near -zero positi v e subtraction operator , denoted as np . 3.2. Modification of multiplication arithmetic operators In this study , fuzzy arithmetic multiplication operator as stated in Definition (7) is modified. The modified multiplication operator pro vides simpler and direct computation as compared to the pre vious operators. A modification of fuzzy arithmetic oper ator s for solving near -zer o... (W . S. W . Daud) Evaluation Warning : The document was created with Spire.PDF for Python.
588 r ISSN: 1693-6930 Theor em 4. Let ~ M = ( m; ; ) be a positive , ne gative or near -zer o fuzzy number , and ~ N = ( n; ; ) be a positive fuzzy number . Then, the min and max in (14) ar e given by: Min [( m )( n ) ; ( m )( n + )] = ( ( m )( n ) if ~ M 0 ( m )( n + ) if otherwise (23) Max [( m + )( n ) ; ( m + )( n + )] = ( ( m + )( n ) if ~ M < 0 ( m + )( n + ) if otherwise (24) Pr oof. Based on Theorem 1 and realize that ( n ) < ( n + ) , then ob viously: If ~ M is positi v e which is ( m ) 0 , both multiplications of ( m )( n ) and ( m + )( n ) are minimum compared to the multiplications of ( m )( n + ) and ( m + )( n + ) respecti v ely . Ho we v er , if ( m ) < ( m + ) , then ( m )( n ) is minimum. On the other hand, since b ot h multiplications of ( m )( n + ) and ( m + )( n + ) are maximum compared to the multiplication of ( m )( n ) and ( m + )( n ) respecti v ely , b ut since ( m + ) > ( m ) , thus the maximum v alue is ( m + )( n + ) . If ~ M is ne g ati v e which is ( m ) < 0 , both multiplications of ( m )( n + ) and ( m + )( n + ) are minimum compared to the multiplication of ( m )( n ) and ( m + )( n ) respecti v ely . From that, since ( m ) < ( m + ) , then ( m )( n + ) is minimum. On the other hand, since both ( m )( n ) and ( m + )( n ) are maximum compared to the multiplication of ( m )( n + ) and ( m + )( n + ) respecti v ely , b ut ( m + ) > ( m ) thus the maximum v alue is ( m + )( n ) . If ~ M is near -zero which is ( m ) 0 ( + m ) , based on the inequilities in (16) and (17), then ob viously ( m )( n + ) is minimum, whereas ( m + )( n + ) is maximum. From Theorem 4 and (14), the modified multiplication arithmetic operators are defined in the follo wing theorem. Theor em 5. Let ~ M = ( m; ; ) be a positive , ne gative or near -zer o fuzzy number , and ~ N = ( n; ; ) be a positive fuzzy number , then the multiplication of ~ M ~ N is defined as follows: 1. If ~ M is positive , then ~ M ~ N = ( m; ; ) ( n; ; ) = ( mn; n + ( m ) ; n + ( m + ) ) (25) 2. If ~ M is ne gative , then ~ M ~ N = ( m; ; ) ( n; ; ) = ( mn; n ( m ) ; n ( m + ) ) (26) 3. If ~ M is near -zer o, then ~ M ~ N = ( m; ; ) ( n; ; ) = ( mn; n ( m ) ; n + ( m + ) ) (27) Pr oof. By considering the Corollary 4, and applying it to (14), thus: 1. F or ~ M is positi v e, ~ M ~ N = ( mn; mn ( m )( n ) ; ( m + )( n + ) mn ) = ( mn; mn mn + m + n ; mn + m + n + mn ) = ( mn; m + n ; m + n + ) = ( mn; n + ( m ) ; n + ( m + ) ) (28) TELK OMNIKA T elecommun Comput El Control, V ol. 19, No. 2, April 2021 : 583 598 Evaluation Warning : The document was created with Spire.PDF for Python.
TELK OMNIKA T elecommun Comput El Control r 589 2. F or ~ M is ne g ati v e, ~ M ~ N = ( mn; mn ( m )( n + ) ; ( m + )( n ) mn ) = ( mn; mn mn m + n + ; mn m + n mn ) = ( mn; m + n + ; m + n ) = ( mn; n ( m ) ; n ( m + ) ) (29) 3. F or ~ M is near -zero, ~ M ~ N = ( mn; mn ( m )( n + ) ; ( m + )( n + ) mn ) = ( mn; mn mn m + n + ; mn + m + n + mn ) = ( mn; m + n + ; m + n + ) = ( mn; n ( m ) ; n + ( m + ) ) (30) Since (25) to (27) are sho wn, hence the theorem is pro v ed. Cor ollary 1. Let ~ M = ( m; ; ) be a positive , ne gative or near -zer o fuzzy number , and ~ N = ( n; ; ) be a positive fuzzy number: 1. If ~ M is positive , then the multiplication of ~ M ~ N is positive , suc h that mn ( n + ( m ) ) > 0 . 2. If ~ M is ne gative , then the multiplication of ~ M ~ N is ne gative , suc h that ( n ( m + ) ) + mn < 0 . 3. If ~ M is near -zer o, then the multiplication of ~ M ~ N is near -zer o, suc h that mn ( n ( m ) ) < 0 < ( n + ( m + ) ) + mn . Pr oof. The multiplication of ~ M ~ N in Theorem 5 must satisfy the Definition 5, where 1. If ~ M is positi v e, then mn ( n + ( m ) ) > 0 which is mn ( n + ( m ) ) = mn n m + = ( m ) n ( m ) = ( m )( n ) (31) Since both ( m ) and ( n ) are > 0 , then mn ( n + ( m ) ) > 0 . 2. If ~ M is ne g ati v e, then ( n ( m + ) ) + mn < 0 which is ( n ( m + ) ) + mn = n m + mn = ( m + ) n ( m + ) = ( m + )( n ) (32) Since ( m + ) < 0 and ( n ) > 0 , then mn + ( n ( m + ) ) < 0 . 3. If ~ M is near -zero, then mn ( n ( m ) ) < 0 < ( n + ( m + ) ) + mn which is mn ( n ( m ) ) = mn n + m = ( m ) n + ( m ) = ( m )( n + ) (33) Since ( m ) < 0 and ( n + ) > 0 , then mn ( n ( m ) ) < 0 . On the other hand, ( n + ( m + ) ) + mn = n + m + + mn = ( m + ) n + ( m + ) = ( m + )( n + ) (34) Since ( m + ) > 0 and ( n + ) > 0 , then ( n + ( m + ) ) + mn > 0 . After all the conditions are satisfied, then the corollary is pro v ed. A modification of fuzzy arithmetic oper ator s for solving near -zer o... (W . S. W . Daud) Evaluation Warning : The document was created with Spire.PDF for Python.
590 r ISSN: 1693-6930 3.3. Related pr operties of FFME ~ A ~ X ~ B ~ X = ~ C The definition of FFME ~ A ~ X ~ B ~ X = ~ C is gi v en as follo ws: Definition 11. The matrix equation 0 B B B @ ~ a 11 ~ a 12 : : : ~ a 1 m ~ a 21 ~ a 22 : : : ~ a 2 m . . . . . . . . . . . . ~ a m 1 ~ a m 2 : : : ~ a mm 1 C C C A 0 B B B @ ~ x 11 ~ x 12 : : : ~ x 1 n ~ x 21 ~ x 22 : : : ~ x 2 n . . . . . . . . . . . . ~ x m 1 ~ x m 2 : : : ~ x mn 1 C C C A 0 B B B @ ~ b 11 ~ b 12 : : : ~ b 1 n ~ b 21 ~ b 22 : : : ~ b 2 n . . . . . . . . . . . . ~ b n 1 ~ b n 2 : : : ~ b nm 1 C C C A 0 B B B @ ~ x 11 ~ x 12 : : : ~ x 1 n ~ x 21 ~ x 22 : : : ~ x 2 n . . . . . . . . . . . . ~ x m 1 ~ x m 2 : : : ~ x mn 1 C C C A = 0 B B B @ ~ c 11 ~ c 12 : : : ~ c 1 n ~ c 21 ~ c 22 : : : ~ c 2 n . . . . . . . . . . . . ~ c m 1 ~ c m 2 : : : ~ c mn 1 C C C A (35) can also be represented as ~ A ~ X ~ B ~ X = ~ C (36) where ~ A = ( a ij ) , 1 i; j n , ~ B = ( b ij ) , 1 i; j m , the right hand side matrix ~ C = ( c ij ) , 1 i n; 1 j m is the fuzzy matrices, and ~ X = ( x ij ) , 1 i n; 1 j m is an unkno wn fuzzy matrix. There is a special criterion related to the order of matrix coef ficients for FFME ~ A ~ X ~ B ~ X = ~ C . Remark 2. Let ~ A ~ X ~ B ~ X = ~ C be an FFME, wher e the fuzzy coef ficient of ~ A and ~ B must be any squar e matrices. Example 1. If ~ A and ~ B ar e non-squar e matrices with any appr opriate or der s of ~ A r p and ~ B q s , and the solution is ~ X p q , then ~ A r p ~ X p q ~ B q s ~ X p q = ~ A ~ X ~ B r s ~ X p q : Howe ver , the subtr action of ~ A ~ X ~ B r s and ~ X p q is not possible due to the dif fer ent or der . Thus, in all cases, ~ A and ~ B in FFME ~ A ~ X ~ B ~ X = ~ C must be squar e matrices. 3.4. Construction of an associated linear system Definition 12. Consider a fully fuzzy linear system (FFLS) in the form of ~ S ~ X = ~ C (37) wher e ~ S = ( m; ; ) , ~ X = ( n; ; ) and ~ C = ( C ; G; H ) , whic h is equivalent to n X j =1 ;:::;n ( m ij ; ij ; ij ) ( n j ; j ; j ) = ( C i ; G i ; H i ) : (38) According to the ne w multiplication arithmetic operators stated in Theorem 5, the FFLS can be transformed in a form of a crisp linear system, called as the ALS. Definition 13. Let ~ S = ( m; ; ) be a positive , ne gative or near -zer o fuzzy number , ~ X = ( n; ; ) be a positive fuzzy number and ~ C = ( C ; G; H ) be any form of fuzzy number s, based on the multiplication arithmetic oper ator s in Theor em 5. Then, thr ee forms of ALS ar e obtained, suc h that: If ~ S is positive , 8 > < > : mn = C n + ( m ) = G n + ( m + ) = H 0 @ m 0 0 ( m ) 0 0 ( m + ) 1 A 0 @ n 1 A = 0 @ C G H 1 A (39) TELK OMNIKA T elecommun Comput El Control, V ol. 19, No. 2, April 2021 : 583 598 Evaluation Warning : The document was created with Spire.PDF for Python.
TELK OMNIKA T elecommun Comput El Control r 591 If ~ S is ne gative , 8 > < > : mn = C n ( m ) = G n ( m + ) = H whic h can be r epr esented as 0 @ m 0 0 0 ( m ) ( m + ) 0 1 A 0 @ n 1 A = 0 @ C G H 1 A (40) If ~ S is near -zer o, 8 > < > : mn = C n ( m ) = G n + ( m + ) = H whic h can be r epr esented as 0 @ m 0 0 0 ( m ) 0 ( m + ) 1 A 0 @ n 1 A = 0 @ C G H 1 A (41) By applying the concept of union sets as stated in Definition 2, these three ALS block matrices in (39), (40) and (41) can be combined into a single ALS as illustrated in Definition 14. Definition 14. Let ~ S ~ X = ~ C be a FFLS, wher e the fuzzy coef ficients ~ S and ~ C ar e arbitr ary fuzzy number s and ~ X be a positive fuzzy solution. ALS is r epr esented as 8 > < > : mn = C n + ( m ) ( m ) = G n ( m + ) + ( m + ) = H (42) whic h can be written in the matrix form of 0 @ m 0 0 ( m ) ( m ) ( m + ) ( m + ) 1 A 0 @ n 1 A = 0 @ C G H 1 A (43) wher e m = ( m ij ) m n = 0 B @ m 11 ::: m 1 n . . . . . . . . . m m 1 ::: m mn 1 C A ; = ( ij ) m n = 0 B @ 11 ::: 1 n . . . . . . . . . m 1 ::: mn 1 C A ; = ( ij ) m n = 0 B @ 11 ::: 1 n . . . . . . . . . m 1 ::: mn 1 C A ; n = 0 B @ n 1 . . . n n 1 C A ; = 0 B @ 1 . . . n 1 C A ; = 0 B @ 1 . . . n 1 C A ; A modification of fuzzy arithmetic oper ator s for solving near -zer o... (W . S. W . Daud) Evaluation Warning : The document was created with Spire.PDF for Python.
592 r ISSN: 1693-6930 and C = 0 B @ C 1 . . . C m 1 C A ; G = 0 B @ G 1 . . . G m 1 C A ; H = 0 B @ H 1 . . . H m 1 C A : This ALS can be denoted as S X = C : (44) Ho we v er , the matrix S in (43) is al w ays inconsistent since j S j = 0 , which is pro v ed in the follo wing theorem: Theor em 6. Let S be a coef ficient of an ALS. The matrix S is singular or j S j = 0 , when j m j = 0 or ( m ) ( m ) ( m + ) ( m + ) = 0 : Pr oof. Let S = 0 @ m 0 0 ( m ) ( m ) ( m + ) ( m + ) 1 A The singularity of S can be determined from the follo wing procedure, which is based on Remark 1. j S j = det ( m ) ( m ) 1 (0) ( m ) ( m ) 1 (0) ( m + ) ( m ) 1 (0) ( m + ) ( m ) 1 (0) det [ m ] = det ( m ) ( m ) ( m + ) ( m + ) det [ m ] From this, if j m j = 0 , then ob viously matrix S is singular . On the other hand, if j m j 6 = 0 , b ut ( m ) ( m ) ( m + ) ( m + ) = 0 , hence, matrix S is singular . Remark 3. Ther e ar e two possibilities that mak e the determinant of ( m ) ( m ) ( m + ) ( m + ) = 0 , whic h ar e: i. At least one bloc k matrix in both dia gonal and anti-dia gonal have all zer oes in a r ow , suc h that: 0 B B @ 0 0 a b 0 0 c d 0 0 e f 0 0 g h 1 C C A ii. The i th r ow or j th column of a matrix is a multiple of another r ow or column, suc h that: 0 B B @ a b a b c d c d e f e f g h g h 1 C C A In order to a v oid the inconsistenc y of the solution, the ALS in (43) has been impro vised to be in the follo wing form as stated in the ne xt theorem. Definition 15. Let ~ S ~ X = ~ C be a FFLS suc h that ~ S = ( m; ; ) , ~ X = ( n; ; ) and ~ C = ( C ; G; H ) , with solution ~ X as a positive fuzzy number . Then the ALS of S X = C is written as: 0 @ m 0 0 ( m ) + ( m ) ( m + ) ( m + ) + 1 A 0 @ n 1 A = 0 @ C G H 1 A (45) wher e ( m ) + and ( m + ) + contain the positive elements of ( m ) and ( m + ) r espectively , while the ne gati ve elements ar e r eplaced by zer o values. Similarly , ( m ) and ( m + ) contain the ne gative elements of ( m ) and ( m + ) r espectively , while the positive elements ar e r eplaced by zer o values. TELK OMNIKA T elecommun Comput El Control, V ol. 19, No. 2, April 2021 : 583 598 Evaluation Warning : The document was created with Spire.PDF for Python.