A stochastic method for finding the min-max optimal solution of multiple objectives

Abstract

A tabu based algorithm for finding the minmax optimal of multi-objective optimal design problems is proposed. A min-max optimal in multi-objective optimizations is that with the smallest relative increments of all the objective functions. Contrary to other methods used for multi-objective optimizations in electromagnetic device designs, the proposed one requires no scalarization techniques, thus simplifying the numerical implementations considerably. Numerical examples on both test and practical problems are presented to validate the proposed algorithm.