最优子种群实数编码的遗传算法  

Decimal genetic algorithm of optimal filial-population

作　　者：谢俊[1] 黄隽[2] 刘军[3] 郭晨海[3] 

机构地区：[1]江苏大学机械工程学院,镇江212013 [2]镇江高等专科学校数理系,镇江212003 [3]江苏大学汽车与交通工程学院,镇江212013

出　　处：《机械设计与制造》2006年第10期4-6,共3页Machinery Design & Manufacture

基　　金：江苏省高校自然科学基金项目(03KJB610019);江苏大学高级人才启动基金(05JDG015)

摘　　要：提出了最优子种群实数编码遗传算法理论,通过从种群中选出适应值最高的若干数量的个体,组成该代最优子种群,将最优子种群中的个体与种群中其它个体进行交叉变异、最优子种群中的个体间也进行交叉变异,从而产生新的种群。该遗传算法使得遗传过程中落入局部最优解的几乎不可能,对于多极值问题非常有效,收敛速度也非常快。A kind of new decimal genetic algorithm of optimal filial - population is put forward. A certain number of individual with high fitness are chosen from the population and made up an optimal filial - population, the individual of optimal filial- population are made crossover and mutation with the individual of population and the individual of population are made crossover and mutation each other, then the next new population is produced. This genetic algorithm makes it almost impossible to fall into local optimum, and it is very effective on multi - extremum programming, its convergence velocity is very fast.

关 键 词：遗传算法 最优子种群 实数编码 

分 类 号：O224[理学—运筹学与控制论]