COTS products selection using fuzzy chance-constrained multiobjective programming

Abstract

In this paper, we study the commercial off-theshelf (COTS) products selection problem in COTS-based modular software systems with fuzzy parameters using chance-constrained multiobjective programming. The criteria used in this work are total cost, size, execution time, software reliability, delivery time, and compatibility issues among available COTS products. The COTS selection model presented herein simultaneously minimizes the total cost, size, and execution time of the modular software system at a credibility that is no less than pre-determined confidence levels. We propose a hybrid approach that combines the technique for order preference by similarity to ideal solution (TOPSIS) and the compensatory fuzzy approach based on the "fuzzy and" aggregation operator. TOPSIS is used to reduce the multiobjective problem with conflicting and non-commensurable objective functions to a bi-objective problem with conflicting but commensurable objective functions. The resulting bi-objective problem is solved using the compensatory fuzzy approach. The obtained compromise solutions are both compensatory and pareto-optimal for the multiobjective COTS products selection problem. A small-scale real-life case study of modular software development is presented to demonstrate the applicability of the proposed model and solution approach in real-life applications of COTS selection. A detailed performance analysis and comparison with some existing related fuzzy programming approaches are performed using distance measures to illustrate the superiority of the proposed hybrid interactive fuzzy-programming approach over existing approaches.

Risk response incorporating risk preferences in international construction projects

“…The model formulated here is a multi-objective integer linear programming (MOILP) model. Of the different kinds of methods used to solve multi-objective models, the fuzzy interactive method is regarded as one of the most efficient because of its advantage in finding the optimization solution by measuring and adjusting the satisfaction degree of each objective function based on the decision makers' preferences (relative importance) for each objective (Mehlawat and Gupta, 2015; Pishvaee et al , 2012). A fuzzy interactive method, which uses an aggregation function proposed by Torabi and Hassini (2008), is applied to convert the original MOILP model into its equivalent single objective one in this research.…”