Improved Particle Swarm Optimization and Its Application in Photonic Crystal

ZENG Yuan,LI Yuan,XU Jia-dong
DOI: https://doi.org/10.3969/j.issn.1006-9348.2008.03.054
2008-01-01
Abstract:In order to balance the relationship of globe exploration and local exploitation efficiently, and avoid the premature convergence and slowdown convergence phenomenon, the improved Particle Swarm Optimization (IPSO), based on balance mechanism of attraction and repulsion, is proposed. Optimization process is divided into three stages and each stage has its own goals. The paper uses it as guidance to control the value of attraction and repulsion and propagation velocity of excellent value, and focuses on global search or local development in order to enhance the efficiency of algorithm. The experiment results of four benchmark function and application in photonic crystal optimization indicate that IPSO can well avoid the phenomenon of premature convergence and slowdown convergence, compared to PSO, so the IPSO is a highly reliable, stable and efficient algorithm.
What problem does this paper attempt to address?