计算机应用 ›› 2009, Vol. 29 ›› Issue (12): 3273-3276.

• 人工智能 • 上一篇    下一篇

基于非线性共轭梯度法的混沌微粒群优化算法

陈红安   

  1. 湖南大学
  • 收稿日期:2009-06-11 修回日期:2009-08-04 发布日期:2009-12-10 出版日期:2009-12-01
  • 通讯作者: 陈红安

Chaotic particle swarm optimization algorithm based on nonlinear conjugate gradient algorithm

  • Received:2009-06-11 Revised:2009-08-04 Online:2009-12-10 Published:2009-12-01

摘要: 为了寻找多峰函数的全部极值点,提出一种基于非线性共轭梯度法的混沌微粒群算法。该算法引入混沌序列设置微粒群位置以提高种群的多样性;然后使用改进的微粒群认知模型对可行域内的所有极值点进行全局搜索;最后利用非线性共轭梯度法对混沌微粒群算法搜索到的较优解进行局部搜索以提高解的精度。仿真实验表明,该算法能准确、快速地找到连续可微多峰函数的全部极值点。

关键词: 微粒群算法, 非线性共轭梯度法, 混沌, 多峰函数, 极值点

Abstract: For searching for all local optimization of the multi-modal function, a chaos-PSO algorithm based on nonlinear conjugate gradient algorithm was proposed. This algorithm employed chaos sequence to initialize particle swarm location in order to enhance the diversity of population, and then utilized improved PSO cognitive model to search all local optimization in the feasible region, and then used nonlinear conjugate gradient algorithm to improve the accuracy of the sub-optimal solution which chaos-PSO has found. The experiments manifest that the hybrid algorithm can properly and quickly find all local optimization of the continuous-differential function.

Key words: nonlinear conjugate gradient algorithm, chaos, multi-modal function, extreme point