A Novel Multimodal Optimization Algorithm Using a Subgroup Concept for the Design of Electric Machines

The design of electric machines is a multivariable and multimodal problem which requires much time to find the optimal design results. To address this problem, we propose a novel multimodal optimization algorithm using a subgroup concept (MS), which uses a grey wolf optimizer, a subgroup, a Kriging surrogate model, a scout, and an adaptive coefficient. In multivariable and multimodal design problems, the proposed novel algorithm is more rapid, exact, and reliable compared to the conventional algorithms such as niching genetic algorithm (NGA) and auto tuning NGA. The validity of the proposed method was confirmed via test functions. For the verification of the application to electric machines, interior permanent-magnet synchronous motors for electric vehicles was designed using the proposed algorithm.