An improved vector evolutionary algorithm for multiobjective designs of electromagnetic devices

Abstract

Thanks to the success in the design of a simple sharing function, the introduction of a novel fitness assignment strategy, and the development of a new local search procedure, this paper proposes an improved evolutionary algorithm for optimal problems involving several, often conflicting objectives. The simulation results on solving a mathematical function and a prototype problem reveal that the proposed method is effective in sampling the entire Pareto-optimal front and in distributing the generated solutions over the trade-off surface.