Abstract
Since Particle Swarm Optimization (PSO) does not require gradient information of an objective function and its algorithm is simple, it is applied to various engineering fields. However, it is well known that in the case that the objective function is multimodal or that design parameters are high-dimensional, the search performance of PSO falls. In this article, the authors propose mPSO (modified PSO), for which the algorithm of PSO is slightly revised, and present that these problems are greatly improved in numerical experiment of benchmark problems. Furthermore, design parameters are digitized and multilevel optimization process is added to mPSO, which is applied to a three-dimensional truss structure and three-dimensional rigid frame structure, to present effectiveness of this method.
