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 ›› 2011, Vol. 37 ›› Issue (13): 150-152. doi: 10.3969/j.issn.1000-3428.2011.13.048

• 人工智能及识别技术 • 上一篇    下一篇

一种求解数值优化问题的进化规划算法

谢 波,余永权   

  1. (广东工业大学计算机学院,广州 510006)
  • 收稿日期:2010-11-22 出版日期:2011-07-05 发布日期:2011-07-05
  • 作者简介:谢 波(1982-),男,硕士研究生,主研方向:计算智能;余永权,教授、博士生导师
  • 基金资助:
    国家自然科学基金资助项目(60272089);广东省自然科学基金资助项目(980406)

Evolutionary Programming Algorithm for Solving Numerical Optimization Problems

XIE Bo, YU Yong-quan   

  1. (Faculty of Computer, Guangdong University of Technology, Guangzhou 510006, China)
  • Received:2010-11-22 Online:2011-07-05 Published:2011-07-05

摘要: 针对进化规划算法收敛速度慢、容易早熟收敛等问题,提出一种基于探测变异的进化规划算法。该算法通过降维得到多个探测变异量,对个体进行探测变异,使个体始终向适应度好的方向进化,并利用自适应高斯变异标准差伸缩搜索空间,使个体跳出局部最优解。通过3个经典算例对其性能进行测试,实验结果证明该算法收敛速度快,求解质量高,可以解决早熟收敛等问题。

关键词: 进化规划, 探测变异, 降维, 数值优化, 高斯变异

Abstract: Aiming at the problems of Evolutionary Programming(EP) such as premature convergence, slow convergence, this paper proposes an EP algorithm based on probe mutation. It can obtain a few mutational variables by reducing the dimensionality while fixating the others, and makes these individuals mutate which makes the individual always move forward to the direction with high fitness. By applying self-adaptive Gauss standard deviation to recover the search space, the individual has opportunity for jumping out the local optimum solutions. Experimental results show that the improved algorithm converges quickly, has high-quality solutions, and overcomes premature convergence problem.

Key words: Evolutionary Programming(EP), probe mutation, dimensionality reduction, numerical optimization, Gaussian mutation

中图分类号: