A Novel Two-Subpopulation Particle Swarm Optimization

Yan Zhe-ping,Deng Chao,Zhou Jia-jia,Chi Dong-nan
DOI: https://doi.org/10.1109/wcica.2012.6359164
2012-01-01
Abstract:The performance of the particle swarm is mainly influenced by individual particles experience and group experience in the period of evolution for particle swarm optimization. To make full use of the two factors and effectively improve the particle swarm optimization performance, Introduced a novel Two-subpopulation Particle Swarm Optimization, The proportion of individual experience and group experiences is different in each subpopulation swarm. If the proportion of individual experience greater than the group experience, the particle swarm search space abroad, whereas, the proportion of group experience greater than individual experience, the particle swarm search the local area fully. The proposed Two-subpopulation particle swarm optimization combines both advantages, make the search more fully and not easily into the local minimum points. Finally simulations were carried out and the results showed that the proposed Two-subpopulation particle swarm optimization, obviously better than the basic particle swarm algorithm in search precision and stability.
What problem does this paper attempt to address?