作者投稿和查稿 主编审稿 专家审稿 编委审稿 远程编辑

计算机工程

• 人工智能及识别技术 • 上一篇    下一篇

基于随机扩散搜索的协同差分进化算法

张大斌1,周志刚1,叶 佳1,张文生2   

  1. (1. 华中师范大学信息管理学院,武汉 430079;2. 中国科学院自动化研究所,北京 100190)
  • 收稿日期:2013-06-28 出版日期:2014-07-15 发布日期:2014-07-14
  • 作者简介:张大斌(1969-),男,教授、博士,主研方向:商务智能,信息系统;周志刚、叶 佳,硕士研究生;张文生,教授、博士、博士生导师。
  • 基金资助:
    国家自然科学基金资助项目(70971052);中国博士后基金资助项目(2012M510607);湖北省自然科学基金创新群体基金资助项目(2011CDA116)。

Cooperation Differential Evolution Algorithm Based on Stochastic Diffusion Search

ZHANG Da-bin  1, ZHOU Zhi-gang  1, YE Jia  1, ZHANG Wen-sheng   2   

  1. (1. School of Information Management, Central China Normal University, Wuhan 430079, China; 2. Institute of Automation, Chinese Academy of Sciences, Beijing 100190, China)
  • Received:2013-06-28 Online:2014-07-15 Published:2014-07-14

摘要: 针对差分进化算法存在的收敛速度慢、稳健性差等问题,借鉴多种群并行机制和随机搜索策略,提出一种基于随机扩散搜索的协同差分进化算法。引入反向混沌搜索的初始化机制,利用随机扩散搜索策略将种群分为成功和失败2个子群并进行改进,对改进的成功和失败子群分别采用不同的差分策略,克服单一差分策略的缺陷,同时定期使子群的部分最好与最差个体实现一对一的信息交流,从而达到协同进化的目的。仿真结果证明,与粒子群优化算法及差分进化算法相比,该算法具有较好的收敛速度和寻优能力。

关键词: 差分进化, 差分策略, 反向混沌搜索, 协同机制, 正负反馈机制, 函数优化问题

Abstract: Aiming at the problem of slow convergence speed, bad robustness, reference multi populations parallel mechanism and random search strategy. A novel synergy search Differential Evolution based on Stochastic Diffusion Search(SDS-DE), which aims to accelerate convergence and improve accuracy of SDS-DE, the algorithm introduces the initialization mechanism of opposition chaos search, and stochastic diffusion search strategy is used to divided populations into successful and failed two sub-groups, and differential strategies of the success and failures sub-groups are used to overcome the shortcomings single differential strategy, meanwhile, regularly subgroup best and worst part of individuals to achieve positive and negative feedback mechanisms for information exchange, so as to achieve the purpose of co-evolution. Simulation results prove that the SDS-DE performs better convergence speed and optimization capability by the comparison with the Particle Swarm Optimization(PSO) and other DE algorithms.

Key words: Differential Evolution(DE), differential strategy, opposition chaos search, collaboration mechanism, positive and negative feedback mechanism, function optimiztion problem

中图分类号: