一种基于社团划分的小生境遗传算法  

Niche genetic algorithm based on cluster division

作　　者：祝希路[1] 王柏[2] 

机构地区：[1]北京邮电大学北京市智能通信软件与多媒体重点实验室,北京100876 [2]北京邮电大学计算机学院,北京100876

出　　处：《控制与决策》2010年第7期1113-1116,共4页Control and Decision

基　　金：国家自然科学基金项目(60402011);国家十一五科学技术支撑计划项目(2006BAH03B05)

摘　　要：提出一种相似性算子,并根据该算子建立超级个体之间的关系网,运用社团划分算法将该关系网划分成小生境;提出收敛算子,运用该算子判断种群的进化程度,以决定是否进行生境划分.根据生境中维持的共有模式,采用多位变异策略,有效维护生境中个体的多样性,避免早熟收敛.从理论上证明了该算法的收敛性.实验结果表明,该方法具有良好的收敛速度和全局搜索能力,其综合性能优于确定性排挤遗传算法,有利于并行实现.A similarity operator is proposed to construct the network among the super individuals. The cluster division algorithm is used to divide the network to find the niche. In order to judge the evolutionary degree,the convergence operator is introduced. The multi-position mutation policy is used to avoid the premature convergence and maintain the diversity of the population,according to the schema existing in the niche. The convergence of the algorithm is proved theoretically. Simulation results show that the improved algorithm can effectively avoid premature and greatly improve the global efficiency,and the overall performance exceeds the deterministic crowding genetic algorithm,which is convenient in parallel implementation respectively.

关 键 词：遗传算法 社团划分 小生境 多峰函数 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]