遗传算法的改进及其验证  

Improvement of Genetic Algorithm and its Verification

作　　者：李爱军[1] 崔雁松[1] 王晓星[1] 

机构地区：[1]天津职业大学,天津300410

出　　处：《天津职业大学学报》2012年第1期85-87,共3页Journal of Tianjin Vocational Institute

摘　　要：由于传统的轮盘赌选择算子随机操作会产生较大的随机误差,使适应度较高的个体往往失去了选择的机会。而最佳个体保存方法的选择算子可以确保某一代的最优解可不被交叉和变异操作所破坏,但是会破坏种群的多样性,容易造成"早熟"现象的产生。为了弥补上述两种方法各自的不足,提出同时包含两种选择算子的思想,通过实验证明该方法能有效地提高遗传算法的收敛速度。As traditional roulette selection operator randomly operation will have a greater random error,or even a ＂degradation＂ phenomenon,that is,individuals with higher fitness tend to lose the opportunity to choose.The best way to save the individual choice of operator can ensure the optimal solution of a certain generation may not be destroyed by crossover and mutation operations,but it will destroy the diversity of the species,likely to cause ＂premature＂ phenomenon of generation.Therefore,in order to make up for their deficiency of the above two methods,the paper puts forward the thinking of contain both of the selection operator.Proved by experiments,this method can effectively improve the convergence speed of genetic algorithm.

关 键 词：轮盘赌规则 最佳个体 选择算子 遗传算法 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]