The application of a novel OBL based on lens imaging principle in PSO

Abstract

By introducing opposition-based learning (OBL) in PSO, particles are enabled to find an opposite position that is closer to the global optimization solution. However, OBL only makes a good performance on the initial phrase of the evolution, while at later stages it needs to be combined with other techniques (e. g. Cauchy mutation) to improve its ability of "exploration". In this paper, a novel OBL based on the principle of lens imaging is proposed. It uses two parameters (i. e., zoom factor and the factor of search radius), which will achieve a better balance between PSO's "exploration" and "exploitation" abilities. The simulation shows that the novel OBL possesses better convergence rate and convergence effect.