摘要
针对粒子群优化(PSO)算法存在早熟收敛问题,提出了一种改进算法——带有柯西扰动的重分布粒子群优化(RPSO)算法,并应用于IIR数字滤波器的优化设计。RPSO在检测到粒子群早熟收敛时,自动触发粒子重分布机制,帮助粒子逃离局部收敛区域,同时在迭代过程中对种群的全局最优位置施加柯西扰动以保持种群的多样性。仿真实验结果表明,在对IIR数字滤波器设计时,RPSO算法的性能优于粒子群、量子粒子群以及基于混沌变异的粒子群优化等算法。
Due to the shortcoming of particle swarm optimization (PSO) algorithm that it is often premature convergence, an improved PSO algorithm called redistributing PSO with Cauchy disturbance (RPSO) is proposed for designing infinite impulse response (IIR) digital filters. When premature convergence is detected, RPSO automatically triggers particles redistributing mechanism to help particles escape from local convergence regions. Moreover, Cauchy disturbance on the global best position of the swarm is employed in RPSO to maintain the swama diversity. The computer simulations show that IIR digital filters based on RPSO are superior to the ones based on PSO, quantum-behaved PSO (QPSO) and chaotic mutation PSO (CPSO).
出处
《计算机工程与设计》
CSCD
北大核心
2011年第8期2853-2856,共4页
Computer Engineering and Design
关键词
粒子群优化算法
粒子重分布机制
柯西扰动
早熟收敛
IIR数字滤波器
滤波器优化设计
particle swarm optimization
redistributing mechanism
Cauchy disturbance
premature convergence
IIR digital filter
filter optimization design