基于聚类的小生境克隆选择算法  

Cluster based niche clonal selection algorithm

作　　者：郑士芹[1] 邓凡星[1] 

机构地区：[1]北京信息职业技术学院计算机工程系,北京100018

出　　处：《计算机工程与应用》2009年第33期25-27,34,共4页Computer Engineering and Applications

摘　　要：基于聚类的小生境克隆选择算法是针对小生境克隆选择算法计算复杂、参数设置困难等缺点而提出的。新算法删除了计算复杂度较大的抑制算子,引入聚类算子,并对算法的部分流程进行了调整。新算法不仅计算复杂度降低,而且无需预知峰的个数等先验知识,仅根据样本数据即可找到全部峰值点。仿真实验验证了C-NCSA的完全收敛性;并且通过与小生境克隆选择算法的对比实验证明:在相同的实验条件下,C-NCSA的执行时间比NCSA明显降低。Cluster based Niche Clonal Selection Algorithm（C-NCSA） is an improved algorithm for the disadvantages of NCSA, which includes the high computation complex and the difficulties in parameter setting.The restrain-operator with high computation complexity is removed;a cluster operator is imported,and the algorithm flows are reconstructed.Those improvements not only decrease the computation complexity of C-NCSA sharply,but also make C-NCSA find all the peaks just from the sample data without any foreknown conditions such as the number of the peaks.The simulation tests validate the muhi-modal convergence of the C-NCSA,and the contrast tests with the NCSA show that：Compared with the NCSA,the executed time of C-NCSA is reduced obviously at the same test conditions.

关 键 词：人工免疫系统 免疫算法 基于聚类的小生镜克隆选择算法 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]