Solving a multiresponse simulation problem using a dual-response system and scatter search method

Abstract

This research solves a multiresponse simulation problem by using a dual-response system and scatter-search method. The proposed dual-response system constructs a response surface for each response. It then transforms the dual-response problem into a standard non-linear programming formulation. The transformation treats the secondary response as a constraint. In addition, the sample variance from simulation replications is considered simultaneously by adding search area constraints to variance. The proposed scatter-search method uses the scatter-search algorithms as an embedded mechanism in a simulation program to guide the solution-search process. This is an iterative process between the scatter-search algorithms and the simulation program. A practical case study is adopted for empirical illustration. Both the proposed methods are efficient and effective in solving the case-study problem. Finally, the paper discusses the advantages and disadvantages of the proposed methods, and suggests future research opportunities.