摘要
针对粒子群算法对盲信号分离存在的问题,提出二次裂变技术。首先将粒子集中的权值按照逐渐递减顺时针的进行排序,减少搜索算法的时间开销;接着选取一定数目大权值的粒子进行裂变繁殖,繁殖后再生的粒子自粒子集权值的低端逆时针依次覆盖原小权值的粒子,数目正比于原来的粒子群;最后对第二次裂变的数目判断,进行相应扩展或者收缩运算。实验仿真结果表明:二次裂变粒子群算法具有分离效果好、收敛速度快的优点。
In view of the defects of the particle swarm optimization（PSO） algorithm to the blind source separation（BSS）,the secondary fission technique was proposed in this paper.First,the particle pooled weight value was sorted by the rule of clockwise declining gradually to lower the time cost of the searching algorithm.Then the large weight value particles with certain numbers were selected to carry out the particle reproduction by fission.The reproduced particles anti-clockwise covered the original particles with small weight value from the particle pool at low end,which the number of the particle was directly proportional to the original particle swarm.Finally,the second fission particle numbers were distinguished to carry out the corresponding contraction operation or extended operation.The experimental simulation results showed that the particle swarm optimization（PSO） algorithm of the secondary fission had the advantages of good separation effects and fast convergence.
出处
《压电与声光》
CSCD
北大核心
2012年第1期158-162,共5页
Piezoelectrics & Acoustooptics
关键词
大权值
二次裂变
收缩因子
large weight value
secondary fission
contraction factor
作者简介
邵明省（1980-），男，河南滑县人，讲师，硕士生，主要从事信号处理、通信技术的研究。
