华南理工大学学报(自然科学版) ›› 2008, Vol. 36 ›› Issue (9): 6-10.

• 计算机科学与技术 • 上一篇    下一篇

一种改进的自适应粒子群优化算法

徐刚 瞿金平 杨智韬   

  1. 华南理工大学 聚合物新型成型装备国家工程研究中心、聚合物成型加工工程教育部重点实验室, 广东 广州 510640
  • 收稿日期:2007-08-30 修回日期:2008-01-07 出版日期:2008-09-25 发布日期:2008-09-25
  • 通信作者: 徐刚(1974-),男,博士生,南昌大学讲师,主要从事聚合物加工过程建模优化和智能控制研究. E-mail:xgang_csu@163.com
  • 作者简介:徐刚(1974-),男,博士生,南昌大学讲师,主要从事聚合物加工过程建模优化和智能控制研究.
  • 基金资助:

    国家自然科学基金重大项目(10472034,10590351)

An Improved Adaptive Particle Swarm Optimization Algorithm

Xu Gang  Qu Jin-ping  Yang Zhi-tao   

  1. National Engineering Research Center of Novel Equipment for Polymer Processing, Key Laboratory of Polymer Processing Engineering of the Ministry of Education, South China University of Technology, Guangzhou 510640, Guangdong, China
  • Received:2007-08-30 Revised:2008-01-07 Online:2008-09-25 Published:2008-09-25
  • Contact: 徐刚(1974-),男,博士生,南昌大学讲师,主要从事聚合物加工过程建模优化和智能控制研究. E-mail:xgang_csu@163.com
  • About author:徐刚(1974-),男,博士生,南昌大学讲师,主要从事聚合物加工过程建模优化和智能控制研究.
  • Supported by:

    国家自然科学基金重大项目(10472034,10590351)

摘要: 针对粒子群优化算法中出现的对大规模问题的搜索失败,分析了粒子群优化算法的收敛性,指出了粒子速度与搜索失败的关系,提出了一种根据速度信息自适应调整参数的粒子群优化算法.在满足收敛性的条件下,该算法能使粒子根据理想速度自适应调整参数进行搜索.实验结果表明,该算法能解决基本粒子群算法在求解高维、多峰等复杂非线性优化问题时出现的易陷入局部最优和不收敛等搜索失败问题.

关键词: 粒子群优化算法, 自适应性, 平均速度

Abstract:

According to the search failure for large-scale problem via the particle swarm optimization algorithm, the convergence of particle swarm optimization algorithm is analyzed and the relationship between the particle velocity and the search failure is pointed out. Then, an adaptive parameter-adjusting particle swarm optimization algorithm according to the velocity information is put forward. Under the convergent conditions, the proposed algorithm can perform the search by adaptively adjusting the parameters according to the ideal velocity. Experimental results indicate that the proposed algorithm avoids the local optimization and divergence commonly occurred in the conventional particle swarm optimization algorithm in multi-dimension and multi-peak conditions.

Key words: particle swarm optimization algorithm, adaptability, average velocity