A fast evolutionary method for identifying non-inferior solutions in multicriteria shape optimization of a shielded reactor

When tackling the multicriteria optimization of an electromagnetic device, the accurate sampling of the Pareto optimal front implies the use of complex and time-consuming algorithms. In several cases, however, the identification of a few non-inferior solutions is often sufficient for the design purposes; as a consequence, the computational cost may be reduced A simple evolutionary methodology is proposed to obtain a fast approximation of non-inferior solutions; the optimal shape design of a shielded reactor is presented as a case study.