A method for solving reliability optimization problem by fuzzy multiobjective 0–1 linear programming

Abstract

AbstractOptimization problems related to system reliability such as unit selection and redundant allocation are generally considered to be 0–1 linear programming (0–1 LP) problems. Recently an efficient method which considers the particular structure called generalized upper bounding (GUB) was proposed. On the other hand, it is increasingly necessary to deal with such reliability optimization problem as multiple objective decision making (MODM) problems with multiple conflicting functions. But the above method (GUB) is limited to solving single‐objective or bicriteria 0–1 LP problems. In this paper we propose a new fuzzy multiple objective 0–1 LP (0–1 FMLP) method which can adapt to changes in the utilization of system resources and the economic environment by making use of tolerances in the system control parameters. At the same time, the method considers the variability which arises from the subjective judgment of the system decisionmaker (DM) by introducing fuzzy logic into the goal specification. This method is an extension to the more general MODM problem of the efficient algorithm which takes into account the GUB structure that arises in the solution of a system reliability optimization problem. Further, in order to show the effectiveness of the method proposed here, we consider the problem of optimally planning unit selection and redundant distribution in order to maximize system availability.