On the nonlinear programming problems subject to a system of generalized bipolar fuzzy relational equalities defined with continuous t-norms

Abstract: This paper, in the first step, develops the system of bipolar fuzzy relational equations (FRE) to the most general case where the bipolar FREs are defined by an arbitrary continuous t-norm. Also, since fuzzy relational equations are special cases of the bipolar FREs, the proposed system can be also interpreted as a generalization of traditional FREs where the fuzzy compositions are generally defined by any continuous t-norm. The consistency of the continuous bipolar FREs is initially investigated and some necessary and sufficient conditions are derived for determining the feasibility of the proposed system. Subsequently, the feasible solutions set of the problem is completely characterized. It is shown that unlike FREs and those bipolar FREs defined by continuous Archimedean t-norms, the feasible solutions set of the generalized bipolar FREs is formed as the union of a finite number of compact sets that are not necessarily connected. Moreover, five techniques have also been introduced with the aim of simplifying the current problem, and then an algorithm is accordingly presented to find the feasible region of the problem. In the second step, a new class of optimization models is studied where the constraints are defined by the continuous bipolar FREs and the objective function covers a wide range of (non)linear functions such as maximum function, geometric mean function, log-sum-exp function, maximum eigenvalue of a symmetric matrix, support function of a set, etc. It is proved that the problem has a finite number of local optimal solutions and a global optimal solution can always be obtained by choosing a point having the minimum objective value compared to all the local optimal solutions. Finally, to illustrate the discussed study, a step-by-step example is described in several sections, whose constraints are a system of the bipolar FRES defined by Dubois-Prade t-norm.