计算机工程与应用 ›› 2010, Vol. 46 ›› Issue (15): 28-30.DOI: 10.3778/j.issn.1002-8331.2010.15.009

• 研究、探讨 • 上一篇    下一篇

有导向交叉算子的研究

闭应洲1,2,丁立新2,李文敬1   

  1. 1.广西师范学院 计算机与信息工程学院,南宁 530001
    2.武汉大学 软件工程国家重点实验室,武汉 430072

  • 收稿日期:2010-01-06 修回日期:2010-02-25 出版日期:2010-05-21 发布日期:2010-05-21
  • 通讯作者: 闭应洲

Study on guided crossover operator

BI Ying-zhou1,2,DING Li-xin2,LI Wen-jing1   

  1. 1.School of Computer and Information Engineering,Guangxi Teacher Education University,Nanning 530001,China
    2.State Key Laboratory of Software Engineering,Wuhan University,Wuhan 430072,China

  • Received:2010-01-06 Revised:2010-02-25 Online:2010-05-21 Published:2010-05-21
  • Contact: BI Ying-zhou

摘要:

为了更有效地处理建筑块,提出有导向的交叉算子。首先反复运行快速演化算法找到多个局部最优解,然后识别这些局部最优解中的重要基因位,将其标识为潜在的建筑块,然后应用有导向的交叉算子,组合父代中的建筑块。基于4阶欺骗问题验证了新算法的可行性。

关键词: 演化算法, 有导向交叉算子, 建筑块识别与重组, 欺骗问题

Abstract: To process building-blocks effectively,a guided crossover operator is introduced.First,a fast evolution algorithm is performed repeatedly,and a local optimal solution is found in each independent run.Then identify the important genes in local optimal solution,and mark these genes as potential building blocks.Finally,the guided crossover operator is utilized to mix the building-blocks in different local optimal solutions to generate offspring.The new algorithm is verified with 4-order deceptive problem.

Key words: evolutionary algorithm, guided crossover operator, building-blocks identifying and mixing, deceptive problem

中图分类号: