计算机工程与应用 ›› 2010, Vol. 46 ›› Issue (30): 169-172.DOI: 10.3778/j.issn.1002-8331.2010.30.050

• 图形、图像、模式识别 • 上一篇    下一篇

基于特征选择的智能化分组遗传算法

洪留荣,张建成   

  1. 淮北师范大学 计算机科学与技术学院,安徽 淮北 235000
  • 收稿日期:2009-03-12 修回日期:2009-05-20 出版日期:2010-10-21 发布日期:2010-10-21
  • 通讯作者: 洪留荣

Novel intelligentizing grouping genetic algorithm combining feature selection

HONG Liu-rong,ZHANG Jian-cheng   

  1. School of Computer Science and Technology,Huaibei Normal University,Huaibei,Anhui 235000,China
  • Received:2009-03-12 Revised:2009-05-20 Online:2010-10-21 Published:2010-10-21
  • Contact: HONG Liu-rong

摘要: 典型遗传算法在进化过程中易陷入局部收敛、过早收敛,效率低,针对这些问题,提出一种基于特征选择的智能化分组遗传算法,利用特征选择原理和分组优化思想对进化过程中的基因进行智能分组的遗传操作,在适应度函数中引入个体特征构建动态的环境适应度评价模型。算法通过分组的遗传操作,保证了父代的优秀模式遗传到下一代,加快了收敛速度,分组变异算子扩大了搜索范围,使结果容易走出局部最优解。应用实验验证表明,算法对局部最优解有较强的免疫能力,有效搜索到全局最优解的进化代数较典型遗传算法明显减少,收敛精度高,证明了算法的有效性。

关键词: 分组遗传算法, 智能化, 特征选择, 搜索空间

Abstract: Canonical genetic algorithm in the evolutionary process,fall into local convergence easy,premature convergence and low efficiency,to solve the problem,this paper presents a novel intelligentizing grouping genetic algorithm combining feature selection,intelligenting grouping genetic operation on genes in the evolution process using feature selection principle and grouping optimize idea,the fitness function at the introduction of individual feature and dynamic environmental fitness assessment model to build.This algorithm ensures the excellent genetic model of parent generation to the next generation and speeds up the algorithm convergence rate by the grouping genetic operation,grouping mutation operator expands the scope of the search,and make working out the local optimal solution.Application experimental results show algorithm has a strong immunity to the local optimal solution and the evolution algebra of searching the global optimal solution effectively less than canonical genetic algorithm,has high convergence precision and indicating the effectiveness of the proposed method.

Key words: grouping genetic algorithms, intelligentizing, feature selection, searching space

中图分类号: