A tabu based vector optimal algorithm applied to multiobjective designs of electromagnetic devices

Abstract

A tabu based vector optimal algorithm to seek the Pareto front of multiobjective designs of electromagnetic devices is introduced in this paper. As a general rule, an efficient solver for multiobjective optimization problems should have the following features: the solutions obtained are Pareto optimal, and the solutions are uniformly sampled from the Pareto optimal set. To achieve these two goals, different approaches; including the introduction of the parameter and objective strings, the ranking concept, and an intensification phase; the new termination criterion; as well as the new neighborhood generation scheme; are proposed. Computer simulation results are reported to demonstrate the robustness of the proposed algorithm for both obtaining and uniformly sampling the Pareto optimal front of the multiobjective optimal design problems.