排挤小生态遗传算法的改进方法  

Improvement of niching genetic algorithms using crowding

作　　者：谭竹梅[1] 余晓峰[1] 郭观七[1] 

机构地区：[1]湖南理工学院机械与电气工程系,湖南岳阳414000

出　　处：《控制理论与应用》2004年第4期651-654,共4页Control Theory & Applications

基　　金：湖南省教育厅科学研究基金项目(2001C380;2002A052).

摘　　要：提出了基于搜索空间聚类分析的聚类排挤小生态遗传算法.通过分析适应值曲面的拓扑结构和扩大相似个体的搜索范围,聚类排挤可确定搜索空间的局部性,减少排挤的替换错误并抑制种群的遗传漂移;通过结合确定性替换和概率替换策略,聚类排挤提高了并行局部爬山能力和并行子种群维持能力.对不同多峰问题的仿真优化结果表明,聚类排挤小生态遗传算法的有效峰数量、平均峰值比和全局最优解比等综合性能一致地优于适应值共享、简单确定性排挤和概率排挤等小生态遗传算法.A class of niching genetic algorithms using clustering crowding is proposed.By analyzing topology of fitness landscape and extending the space for searching similar individual,clustering crowding can determine the locality of search space more accurately,thus decreasing the replacement errors of crowding and suppressing genetic drift of the population.The integration of deterministic and probabilistic crowding increases the capacity of both parallel local hill_climbing and maintaining multiple subpopulations.The experimental results optimizing various multimodal functions show that,the performances such as the number of effective peaks,average peak ratio and global optimum ratio of genetic algorithms using clustering crowding are uniformly superior to that of the genetic algorithms using fitness sharing,simple deterministic crowding and probabilistic crowding.

关 键 词：遗传算法 小生态 排挤 聚类分析 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]