Multiobjective optimum design in mixed integer and discrete design variable problems

Abstract

A min-max variant of the global criterion approach is proposed to obtain solutions to multiobjective optimum design problems involving a mix of continuous, discrete, and integer design variables. This modified global criterion approach is used in conjunction with a branch-and-bound algorithm, where the latter was configured especially to accommodate a mix of integer and discrete design variables. The use of a weighting strategy with the proposed method allows the generation of a set of Pareto optimal solutions for both convex and nonconvex problems