一种改进的遗传算法及其收敛性分析  

An improved genetic algorithm and its convergence

作　　者：卢玉宇[1] 陈建能[2] 张翔[1] 

机构地区：[1]福建农林大学机电工程学院,福州350002 [2]浙江理工大学机械与自动控制学院,杭州310018

出　　处：《安徽农业大学学报》2007年第4期608-612,共5页Journal of Anhui Agricultural University

基　　金：福建省自然科学基金(Z0511031);福建省教育厅科技计划项目(JA03056)共同资助

摘　　要：为了克服基本遗传算法只考虑种内竞争的缺陷,提出了基于种间竞争的基本遗传算法的改进算法,并用Markov链分析了遗传算法的收敛性问题,证明了改进的遗传算法收敛到全局最优解的概率为1。通过考核实例计算,发现改进后的算法具有更强的鲁棒性。In order to overcome a defect of basic genetic algorithm that only competition among one population is considered, an improved algorithm based on basic genetic algorithm was proposed. The convergence of the im- proved algorithm was analyzed by Markov chain, and a proof was presented that the convergent probability of im- proved genetic algorithm was 1. Some evaluative examples showed that the results obtained by the improved algo- rithm were robuster than the original one.

关 键 词：遗传算法 收敛性 MARKOV链 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]