一种基于免疫抗体聚类算法的复杂函数寻优  

A New Immune Clonal Algorithm Based on Antibody Cluster for Complicated Function Optimization

作　　者：徐雪松[1] 章兢[1] 周泉[1] 贺庆[2] 

机构地区：[1]湖南大学电气与信息工程学院,湖南长沙410082 [2]中南大学信息科学与工程学院,湖南长沙410083

出　　处：《湖南大学学报（自然科学版）》2007年第9期39-43,共5页Journal of Hunan University:Natural Sciences

基　　金：国家自然科学基金重点资助项目(60634020);教育部博士点基金资助项目(20060532026)

摘　　要：通过在克隆选择过程中引入抗体聚类机制,提出了一种用于复杂多模函数优化的新算法.通过聚类将抗体群分成多个子种群来实现其克隆选择策略,加速克隆扩增,从而提高抗体成熟力及亲和性.采用了混合超变异算子,使其能快速获取全局及局部最优.实验仿真结果表明:该算法对复杂函数寻优的过程是相当有效的,具备良好的全局和局部收敛可靠性.A new algorithm for multimodal function optimization was presented by introducing a cluster mechanism in the clonal selection process. The cluster operation divided the population into subpopulations for the stage of selection and reproduction, thus improving the variety of antibodies and affinity maturation. In order to quickly obtain the global optimum and local optimum, a hybrid hyper-mutation operator was adopted. Simulation results have illustrated that the efficiency of the proposed algorithm for complicated function optimization has proved its remarkable quality of global and local convergence reliability.

关 键 词：变异 聚类 多模态函数 克隆选择 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]