| 基于适应度分组的进化策略 |
| |
| 引用本文： | 张成,李影,邢伟. 基于适应度分组的进化策略[J]. 系统仿真学报, 2007, 19(21): 5081-5083 |
| |
| 作者姓名： | 张成 李影 邢伟 |
| |
| 作者单位： | 1. 沈阳化工学院,数理系,辽宁,沈阳,110142
2. 东北大学,系统科学研究所,辽宁,沈阳,110004 |
| |
| 摘 要： | 一般的进化策略都采用选择子代中最好的一部分个体替代原来的父代个体进入下一代进化,这样使那些相对劣的个体没有机会进入以后的进化过程。但每一个个体都有优良的一面,它们都有可能进化成最优解。鉴于此,根据每代群体的平均适应度将种群分为两部分:优势种群和劣势种群。父代个体以一定的比例在两组中选取,这样使劣质个体也有机会参加到进化中。仿真结果表明该方法对求解优化问题是有效的。
|
| 关 键 词： | 进化策略 平均适应度 优秀个体 劣质个体 |
| 文章编号： | 1004-731X（2007）21-5081-03 |
| 收稿时间： | 2006-08-28 |
| 修稿时间： | 2007-02-20 |
Evolutionary Strategies Using Fitness-based Grouping |
| |
| ZHANG Cheng,LI Ying,XING Wei. Evolutionary Strategies Using Fitness-based Grouping[J]. Journal of System Simulation, 2007, 19(21): 5081-5083 |
| |
| Authors: | ZHANG Cheng LI Ying XING Wei |
| |
| Affiliation: | 1.Department of Mathematics and Physics, Shenyang Institute of Chemical Technology, Shenyang 110142, China; 2.Institute of System Sciences, Northeastern University, Shenyang 110004, China. |
| |
| Abstract: | Best individuals are selected to replace parent individuals to generate new individual in general evolutionary strategies.This makes worst individuals have no chance to reproduce themselves.But every individual can provide good message,all of them can become optimal through evolution.So the population can be divided into two groups based on average fitness:best population and worst population.Parent individuals are selected between the two groups.That makes each individual have chance to reproduce.The result shows that the modified algorithm performs well in optimization problems. |
| |
| Keywords: | evolutionary strategies average fitness best individual worst individual |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
|
