摘要
针对标准微粒群优化算法(PSO)在全局优化过程中容易陷入局部极值的问题,分析了标准微粒群优化算法早熟收敛的原因,提出了一种新的基于不同进化模型的双群交换技术的改进微粒群优化算法.该方法将微粒分成两个大小相同的分群,其中第一分群采用标准PSO模型进化,第二分群采用cognition only模型进化.两个分群每迭代一次后,将第一分群的适应值最差的微粒与第二分群的适应值最优的微粒进行交换,以提高种群的多样性,改善算法的收敛性.与其它双群算法相比,该算法概念简单,程序实现容易.与标准微粒群优化算法相比,全局寻优能力更强,函数测试结果表明,提出的双群交换微粒群优化算法的收敛性能明显优于标准PSO算法.
In order to overcome such drawbacks of basic PSO as falling into local extremum, the reason of premature convergence about basic PSO is analyzed, and an improved PSO algorithm （TSE-PSO） based on two sub-swarms exchange technology is proposed. The particle swarm is divided into two identical sub-swarms ,with the first adopting the basic PSO model, and the second adopting the cognition only model. The worst fitness of the first sub-swarm exchanges with the best fitness of the second one after several iterations, which can increase the information exchange between the particles, improve the diversity of swarm and meliorate the convergence of algorithm. Compared with other PSO based on the two sub-swarms, the algorithm proposed in the paper is easy to implement, and the ability of finding global optima is better than basic PSO. The results of function test indicate that the algorithm has better convergence than the basic PSO.
出处
《南昌工程学院学报》
CAS
2008年第4期1-4,共4页
Journal of Nanchang Institute of Technology
基金
国家自然科学基金资助项目(50539020)
江西自然科学基金资助项目(2007GZS1056)
江西教育厅科技项目(赣教技字[2007]339号)
关键词
微粒群优化算法
种群多样性
优化
particle swarm optimization（PSO）
diversity of population
optimization
