| 基于Metropolis准则的微粒群算法 |
| |
| 引用本文： | 王丽芳,GUO Xiao-dong,曾建潮.基于Metropolis准则的微粒群算法[J].系统仿真学报,2008,20(14). |
| |
| 作者姓名： | 王丽芳 GUO Xiao-dong 曾建潮 |
| |
| 作者单位： | 1. 兰州理工大学电气工程与信息工程学院,甘肃兰州,730050;太原科技大学系统仿真与计算机应用研究所,太原,030024
2. 太原科技大学系统仿真与计算机应用研究所,太原,030024 |
| |
| 基金项目： | 国家自然科学基金，太原科技大学校科研和教改项目 |
| |
| 摘 要： | 通过对微粒群算法的分析,指出其早熟收敛的原因,并提出利用Metropolis准则更新微粒的个体经验位置,从而增强了算法的全局探索能力。该算法也可以认为是模拟退火算法中利用微粒群算法的进化公式作为一种新的状态产生函数。通过理论分析阐明了该算法以概率1收敛于全局最优解。实例仿真验证了其有效性。
|
| 关 键 词： | 微粒群算法 模拟退火算法 全局收敛性 全局探索能力 |
Modified Particle Swarm Optimization Based on Metropolis Rule |
| |
| WANG Li-fang,GUO Xiao-dong,ZENG Jian-chao.Modified Particle Swarm Optimization Based on Metropolis Rule[J].Journal of System Simulation,2008,20(14). |
| |
| Authors: | WANG Li-fang GUO Xiao-dong ZENG Jian-chao |
| |
| Abstract: | Through analysis to particle swarm optimization, the reason of premature convergence was pointed. A new particle swarm optimization was introduced which improved the global explorative ability by updating the individual best position according to the Metropolis Rule. The algorithm could also be considered as a modified SA that utilized particle swarm optimization evolutionary equation as a new state-generator. It was discussed that the algorithm could converge to the global optimum in probability. The simulation results also confirm the validity of the new algorithm. |
| |
| Keywords: | particle swarm optimization simulated annealing algorithm global convergence global exploration |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
