计算机工程与应用 ›› 2007, Vol. 43 ›› Issue (1): 70-70.

• 学术探讨 • 上一篇    下一篇

一种基于选择的遗传算法

乔超,唐慧佳,王春红   

  1. 山西省运城学院
  • 收稿日期:2006-01-17 修回日期:1900-01-01 出版日期:2007-01-01 发布日期:2007-01-01
  • 通讯作者: 乔超 chaoqiao

a kind of selection-based genetic algorithm(GA_S)

,,   

  1. 山西省运城学院
  • Received:2006-01-17 Revised:1900-01-01 Online:2007-01-01 Published:2007-01-01

摘要: 鉴于标准遗传算法比较容易产生早熟现象和模式欺骗而收敛于局部最优解,本文对标准遗传算法的遗传操作进行了改进,提出了基于选择的遗传算法(GA_S)。在该算法中,首次提出了基因选择算子、广义精英算子、引进选择算子、基于精英集的成长期变异等概念,并对其进行了比较详细地描述。之后,使用7个经典测试函数对其进行了大量实验。实验表明本算法对早熟和模式欺骗具有较强的突破能力。

关键词: 遗传算法, 遗传操作, 模式欺骗

Abstract: Considering the easy premature convergence and schema deceptive problem in SGA (standard genetic algorithms), the selection-based genetic algorithm (GA_S) is proposed to improve the genetic operation. Some concepts such as gene selection operator, generalized elitists selection operator, introducing individual selection operator, and elitist mutation operator, are brought up for the first time in the algorithms. Then they are described in detail in this paper. Many experiments using traditional testing functions are conducted on this algorithm, and the results show that the algorithm can solve the premature convergence and the schema deceptive problem to a certain extent.

Key words: genetic algorithm, genetic operation, schema deceptive