北京邮电大学学报

  • EI核心期刊

北京邮电大学学报 ›› 2002, Vol. 25 ›› Issue (1): 27-29.

• 学术论文 • 上一篇    下一篇

基于群体列突变遗传算法的遍历性

孟祥武,张玉洁   

  1. 北京邮电大学计算机科学与技术学院, 北京 100876
  • 收稿日期:2001-04-25 出版日期:2002-01-10
  • 作者简介: 孟祥武(1966—),男,博士,副教授
  • 基金资助:
     

The Traversing of Genetic Algorithms Based on #br# Mutation of Column in Population

MENG Xiang-wu, ZHANG Yu-jie   

  1. Computer Science and Technology School, Beijing University of Posts and Telecommunications, Beijing 100876, China
  • Received:2001-04-25 Online:2002-01-10
  • Supported by:
     

摘要: 通过分析遗传算法过早收敛的原因,提出了一种基于群体列突变的遗传算法,分析了这种遗传算法交换和突变操作的特性,证明该遗传算法能够进行遍历搜索。该算法克服了遗传算法的局限性,通过群体列突变能够避免过早收敛,并找到最优解。

关键词: 并行算法, 遗传算法, 突变操作, 过早收敛

Abstract: The reason for premature convergence of genetic algorithms is analyzed, and then an genetic algorithms based on mutation of column in population is presented, the properties of crossover and mutation in this genetic algorithms are analyzed, the traversing property of this genetic algorithmsis proved. The new algorithm copes with the limitations of genetic algorithms,can avoid premature convergence by mutation of column in population, and find optimal solutions.

Key words: parallel algorithms, genetic algorithms, mutation, premature conver-gence

中图分类号: