具有遗传特性的粒子群优化算法及在非线性盲分离中的应用

粒子群优化算法是一种新的基于群智能的随机优化进化算法.文章将变异和交叉思想引入到粒子群优化算法中,其基本思想是利用粒子群优化算法每次迭代的最优粒子位置及速度为基础对部分粒子进行变异,然后对变异前后粒子的分量进行随机交叉操作,从而产生新一代粒子群.通过这种处理使得粒子群体的进化速度加快,从而提高了算法的收敛速度和精度.该算法应用于盲信号分离中而获得一种非线性盲信号分离算法.计算机仿真结果表明该算法的收敛性能优于粒子群优化算法,并且在非线性盲信号分离中是有效的.

A particle swarm optimizer with genetics and its application in nonlinear blind sources separation

Particle swarm optimization is a new swarm-intelligent-based stochastic optimization evolutionary algorithm.The present paper incorporated mutation operator and crossover operator into original particle swarm optimizers.In this algorithm,the new particle swarm are reproduced by operating on the current global best individual with mutation and on a stochastic selected components from the current individuals and mutation individuals with crossover.The particle swarm optimization speeded up the evolution process,and improved the convergence speed and accuracy.When applide to separation of blind signals,a nonlinear blind source separation algorithm was obtained.The computer simulation results showed that the proposed algorithm was superior to original particle swarm optimization algorithms and was effective in separating nonlinear blind sources.