遗传选择算子的比较与研究  

Comparison and research over genetic selection operators

作　　者：杨平[1] 郑金华[1] 

机构地区：[1]湘潭大学信息工程学院,湖南湘潭411105

出　　处：《计算机工程与应用》2007年第15期59-62,65,共5页Computer Engineering and Applications

基　　金：国家教育部留学回国人员科研启动基金(The Project- sponsored by SRF for ROCS; SEM 教外司留[2005]546 号) ;湖南省自然科学基金 (the Natural Science Foundation of Hunan Province of China under Grant No.05JJ30125) 。

摘　　要：在改进的基本遗传算法的实验基础上,通过分析传统的基本选择算子的理论及其优缺点,提出了能够产生较好收敛速度的三种选择方法:基于上限的确定式采样、基于切断的轮盘赌选择以及无回放最大值选择法,通过实验证明其在收敛性和收敛速度上都有很大的改善,为一些需要更快速求到最优解的应用问题提供了更好的选择策略。This article analyses the theory of traditional basic selection operators and the advantage and disadvantage based on the experiment of simple genetic algorithm that is improved.Then three approaches as Deterministic Sampling based on Upper-limit,Roulette Wheel Selection based on Cutting and Biggest Value Selection with Replacement are proposed for quickening the convergence speed.The new methods that are proved to have a good advance on the convergence and speed can be used as an efficient way for those application problems that are wanted to reach their best value.

关 键 词：选择算子 基于上限的确定式采样 基于切断的轮盘赌选择 无回放最大值选择法 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]