改进的萤火虫群优化算法及其非线性盲源分离  

Nonlinear Blind Source Separation Algorithm Using Glowworm Swarm Optimization with Baffle Effect

作　　者：黄浦博 尉宇[1,2] 

机构地区：[1]武汉科技大学,武汉430070 [2]华中科技大学,武汉430047

出　　处：《电信科学》2015年第9期97-102,共6页Telecommunications Science

摘　　要：根据萤火虫算法的特点,针对萤火虫算法较早进入局部收敛的不足进行改进,引入了"挡板效应"的方法,扩大种群多样性,并提出了BGSO算法。在分析了非线性盲源分离模型的基础上,利用系统两次遍历萤火虫群,比较评价函数,获得最优解的方式。其有效性被仿真结果所证实。According to the characteristics of glowworm swarm optimization （GSO） algorithm, to overcome the disadvantages of premature convergence, a new method called ＂baffle effect＂ was proposed. The diversity of swarm was expanded. Based on the analysis of the model of nonlinear blind source separation （NBSS）, using traversing glowworm swarm twice, comparing evaluation function, the best result was get. Their validity is confirmed by effect of the signal separation test.

关 键 词：非线性盲源分离 萤火虫优化 挡板效应 

分 类 号：TN911.7[电子电信—通信与信息系统]