| 引用本文: | 莫思敏,曾建潮,谢丽萍.扩展的微粒群算法[J].控制理论与应用,2012,29(6):811~816.[点击复制] |
| MO Si-min,ZENG Jian-chao,XIE Li-ping.Extended particle-swarm optimization algorithm[J].Control Theory and Technology,2012,29(6):811~816.[点击复制] |
|
| 扩展的微粒群算法 |
| Extended particle-swarm optimization algorithm |
| 摘要点击 2543 全文点击 1717 投稿时间：2011-02-11 修订日期：2011-10-25 |
| 查看全文 查看/发表评论 下载PDF阅读器 |
| DOI编号 10.7641/j.issn.1000-8152.2012.6.CCTA110136 |
| 2012,29(6):811-816 |
| 中文关键词 微粒群算法 拟态物理学 引斥力规则 |
| 英文关键词 particle-swarm optimization algorithm artificial physics attraction and repulsion force rule |
| 基金项目 国家自然科学基金资助项目(60975074); 山西省青年基金资助项目(2011021019-3). |
|
| 中文摘要 |
| 微粒间的作用方式是影响微粒群算法的关键因素. 为克服微粒群算法的早熟问题, 提出一种扩展的微粒群算法(EPSO). 基于拟态物理学中的引斥力思想, 重新构建微粒间的作用方式. 通过微粒间适应值的比较定义微粒间作用的引斥力规则, 使微粒在所有微粒对其产生的引斥力的合力方向上随机地移动寻找最优解. 扩展的微粒群算法与相关算法进行比较, 仿真结果表明: 它能够有效提高微粒群算法的全局优化性能. |
| 英文摘要 |
| The interaction among particles is a key factor affecting the performance of particle swarm optimization (PSO) algorithm. To overcome the premature convergence, an extended particle swarm optimization (EPSO) algorithm is proposed, in which the interaction mechanism among particles is redefined based on the idea of attraction and repulsion forces in Artificial Physics. Furthermore, the rule of attraction and repulsion among particles is defined by comparing particle fitness values. To look for the global optimum, each particle randomly moves along the direction of the resultant force produced by all particles. Simulation results show that EPSO algorithm effectively improves the global performances of other related algorithms. |