电子学报 ›› 2014, Vol. 42 ›› Issue (8): 1522-1530.DOI: 10.3969/j.issn.0372-2112.2014.08.010

• 学术论文 • 上一篇    下一篇

基于精英区域学习的动态差分进化算法

彭虎1,2, 吴志健1,2, 周新宇1,2, 邓长寿3   

  1. 1. 武汉大学软件工程国家重点实验室, 湖北武汉 430072;
    2. 武汉大学计算机学院, 湖北武汉 430072;
    3. 九江学院信息科学与技术学院, 江西九江 332005
  • 收稿日期:2013-11-07 修回日期:2013-12-23 出版日期:2014-08-25
    • 作者简介:
    • 彭虎男,1981年生,武汉大学计算机学院博士研究生,研究方向:智能计算及其在软件工程中的应用.E-mail:hu_peng@whu.edu.cn;吴志健男,1963年生,教授,博士生导师,武汉大学软件工程国家重点实验室副主任,研究方向:智能计算、并行计算和智能信息处理.E-mail:zhijianwu@whu.edu.cn;周新宇男,1987年生,武汉大学计算机学院博士研究生,研究方向:智能计算、并行计算.;邓长寿男,1972年生,博士,教授,九江学院信息科学与技术学院副院长,研究方向:智能计算和数据挖掘.
    • 基金资助:
    • 国家自然科学基金 (No.61364025,No.61070008); 中央高校科研业务费专项资金 (No.2012211020205); 武汉大学软件工程国家重点实验室开放基金 (No.SKLSE2012-09-39); 江西省教育厅科学技术项目 (No.JJ13729)

Dynamic Differential Evolution Algorithm Based on Elite Local Learning

PENG Hu1,2, WU Zhi-jian1,2, ZHOU Xin-yu1,2, DENG Chang-shou3   

  1. 1. State Key Lab of Software Engineering, Wuhan University, Wuhan, Hubei 430072, China;
    2. Computer School, Wuhan University, Wuhan, Hubei 430072, China;
    3. School of Information Science and Technology, Jiujiang University, Jiujiang, Jiangxi 332005, China
  • Received:2013-11-07 Revised:2013-12-23 Online:2014-08-25 Published:2014-08-25
    • Supported by:
    • National Natural Science Foundation of China (No.61364025, No.61070008); Fundamental Research Funds for the Central Universities (No.2012211020205); Open Fund of State Key Laboratory of Software Engineering,  Wuhan University,  China (No.SKLSE2012-09-39); Science and Technique Research Program of Education Department of Jiangxi Province (No.JJ13729)

摘要:

DE算法简单高效,但对复杂问题也存在收敛效率较低的问题,为提高DE算法的全局勘探能力和收敛精度,提出了一种新的精英区域学习动态差分进化算法,算法首先将历史精英保存在精英池中,然后采用正弦函数对精英池中的精英进行区域学习,最后利用动态DE模式有效提高收敛的速度,并从理论上证明了算法的收敛性.通过对包括单峰函数、多峰函数和偏移函数的20个基准测试函数的仿真实验和分析,验证了新算法的有效性和适用性,其能在保持较高的收敛速度的同时也能保持较好的收敛精度,经与多种知名的DE算法在统计学上的分析比较,证明了该算法是一种具有竞争力的新算法.

关键词: 差分进化, 精英池, 精英区域学习, 动态差分进化

Abstract:

DE algorithm is simple and efficient,but for complex problems also exist the problem of low efficiency of convergence.In order to improve the global exploration ability and convergence precision,this paper proposes a novel elite local learning dynamic differential evolution algorithm.Firstly the history elites are preserved in the elite pool,and then the elites in the pool conduct local learning by sine functions,finally dynamic DE model is used to effectively improve the speed of convergence,and the convergence of the algorithm is proved in theory.Algorithm has been tested on 20 benchmark functions including unimodal functions and multimodal functions and shift functions,experiments result verified the effectiveness and applicability,and the new algorithm can maintain higher convergence speed while maintaining better convergence accuracy.Comparison with the state-of-the-art DE in statistical analysis proves that the algorithm is a kind of new competitive algorithm.

Key words: differential evolution, elite pool, elite local learning, dynamic differential evolution

中图分类号: