计算机应用 ›› 2011, Vol. 31 ›› Issue (04): 1094-1098.DOI: 10.3724/SP.J.1087.2011.01094

• 人工智能 • 上一篇    下一篇

求解高维多模优化问题的正交小生境自适应差分演化算法

拓守恒1,汪文勇2   

  1. 1. 陕西理工学院 计算机科学与技术系，陕西 汉中 723000
    2. 电子科技大学 计算机科学与工程学院，成都610054
  • 收稿日期:2010-09-07 修回日期:2010-11-01 发布日期:2011-04-08 出版日期:2011-04-01
  • 通讯作者: 拓守恒
  • 作者简介:拓守恒(1978-)，男，宁夏中卫人，讲师，硕士，主要研究方向：进化计算、人工智能、神经网络；
    汪文勇(1967-)，男，四川简阳人，教授，主要研究方向：下一代互联网管理与测量、下一代互联网应用、无线传感器网络。
  • 基金资助:
    国家863计划项目(2008AA01A303);陕西省教育厅科研基金资助项目();陕西理工学院青年科研基金资助项目(SLG0818)

Self-adaptive differential evolution algorithm based on orthogonal and niche elite for high-dimensional multi-modal optimization

Shou-heng TUO1,Wen-yong WANG2   

  1. 1. Department of Computer Science and Technology, Shaanxi University of Technology, Hanzhong Shaanxi 723000, China
    2. School of Computer Science and Engineering, University of Electronic Science and Technology of China, Chengdu Sichuan 610054, China
  • Received:2010-09-07 Revised:2010-11-01 Online:2011-04-08 Published:2011-04-01
  • Contact: Shou-heng TUO

PDF

摘要： 针对传统优化算法在求解高维多模态优化问题时存在收敛速度慢、求解精度低的问题，提出一种基于正交设计与小生境精英策略的自适应差分进化算法ONDE。首先利用正交表产生初始种群，然后采用小生境精英策略来产生小生境种群(NP)，并用小生境种群更新精英个体；接着应用拥挤裁剪避免种群陷入局部搜索，最后利用自适应差分变异算子改进了差分进化(DE)算法。通过对7个benchmark函数仿真验证，实验结果表明，算法在收敛速度、求解精度和稳定性方面都有较大优势。

关键词: 高维多模态, 正交设计, 小生境识别, 自适应, 差分演化算法

Abstract: Traditional Differential Evolution (DE) algorithm has shortcomings, such as being trapped into local optimum easily, low convergence speed and solution precision. An Orthogonal Niche Differential Evolution (ONDE) algorithm was proposed to resolve these problems. Firstly, the orthogonal table was used to generate initial population; secondly, the niche elite selection strategy was utilized to produce Niche Population (NP), and update Elite Population (EP) with niche population; thirdly, trapping into local search was prevented by crowded cutting; finally, differential evolution operator was improved by using self-adaptive mutation operators. Simulations on seven benchmark functions were used to test the proposed algorithm. The experimental results illustrate that ONDE algorithm has some advantages in convergence velocity, solution precision and stability.

Key words: high-dimensional multi-modal, orthogonal design, niche recognition, self-adapting, Differential Evolution (DE) algorithm

中图分类号: 

版权所有 © 2021 四川计算机应用杂志社有限公司
蜀ICP备 05010208 号-3　新出网证(川)字026号
技术支持：北京玛格泰克科技发展有限公司
访问总数：
今日访问：