Parameter tuning of particle swarm optimization by using Taguchi method and its application to motor design

Abstract

Particle swarm optimization (PSO) has made significant progress and has been widely applied to computer science and engineering. Since its introduction, the parameter tuning of PSO has always been a hot topic. As a robust design method, the Taguchi method is known as a good tool in designing parameters. Thus the Taguchi method is adopted to analyze the effect of inertia weight, acceleration coefficients, population size, fitness evaluations, and population topology on PSO, and to identify the best settings of them for different optimization problems. The results of benchmark functions show that the optimum parameter settings depend on the benchmarks, and all the functions obtain their optimum solutions after parameter tuning. Good result are also achieved when dealing with the optimization design of a Halbach permanent magnet motor, which indicates that the PSO with best parameter settings identified by the Taguchi method is more suitable to such actual engineering problem.