Fully fuzzy linear systems of triangular fuzzy numbers (a,b,c)

Abstract

PurposeThe purpose of this paper is to study a nascent theory and an emerging concept of solving a fully fuzzy linear system (FFLS) with no non negative restrictions on the triangular fuzzy numbers chosen as parameters. Two new simplified computational methods are proposed to solve a FFLS without any sign restrictions. The first method eliminates the non‐negativity constraint from the coefficient matrix while the second method eliminates the constraint of non‐negativity on the solution vector. The methods are introduced with an objective to broaden the domain of fuzzy linear systems to encompass a wide range of problems occurring in reality.Design/methodology/approachThe design of numerical methods is motivated by decomposing the fuzzy based linear system into its equivalent crisp linear form which can be further solved by variety of classical methods to solve a crisp linear system. Further the paper investigates Schur complement technique to solve the crisp equivalent of the FFLS.FindingsThe results that are obtained reveal interesting properties of a FFLS. By using the proposed methods, the authors are able to check the consistency of the fuzzy linear system as well as obtain the nature of obtained solutions, i.e. trivial, unique or infinite. Further it is also seen that an n×n FFLS may yield finitely many solutions which may not be entirely feasible (strong). Also the methods successfully remove the non‐negativity restriction on the coefficient matrix and the solution vector, respectively.Research limitations/implicationsEvolving methods with better computational complexity and that which remove the non‐negativity restriction jointly on all the parameters are left as an open problem.Originality/valueThe proposed methods are new and conceptually simple to understand and apply in several scientific areas where fuzziness persists. The methods successfully remove several constraints that have been employed exhaustively by researchers and thus eventually tend to widen the breadth of applicability and usability of fuzzy linear models in real life situations. Heretofore, the usability of FFLS is largely dormant.

Application of the Fuzzy Delphi Method and the Fuzzy Analytic Hierarchy Process for the Managerial Competence of Multinational Corporation Executives

“…Therefore, this study employed the fuzzy mean and spread method [8], [9], transforming the result into a crisp number S j . (5) The crisp numbers of indicators were ranged in order from the largest to the smallest.…”

A new solution of pair matrix equations with arbitrary triangular fuzzy numbers

“…1. Our constructed algorithm is able to solve the PFFME and PFME since the FME is a subset of FFME and the FLS is also a subset of FFLS [21].…”

Krylov subspace method for fuzzy eigenvalue problem

“…Recently, several researchers considered the solution of (1) when the coefficient matrix A is fuzzy in nature, denoted by A. Muzzioli and Reynaerts [42] studied the problem in terms of fully fuzzy linear systems [2,5,7,10,11,23,25,26,35,36,40] (that is, the coefficient matrix of the fuzzy linear system [1, 3, 6, 8, 9, 12-15, 17, 22, 30, 43] is also fuzzy), in which 2 n(n+1) crisp system need to be solved. Subsequently, making the scheme impractical for large scale fuzzy eigenvalue problem (FEVP).…”