基于自适应量子粒子群算法的FIR滤波器设计  

FIR filter design based on adaptive quantun-behaved particle swarm optimization algorithm

作　　者：方伟[1] 孙俊[1] 须文波[1] 

机构地区：[1]江南大学信息工程学院智能与高性能计算研究所,江苏无锡214122

出　　处：《系统工程与电子技术》2008年第7期1378-1381,共4页Systems Engineering and Electronics

基　　金：国家自然科学基金项目(NSF60474030)资助课题

摘　　要：针对量子粒子群优化(quantum-behaved particle swarm optimization,QPSO)算法的参数控制方式,提出了一种自适应调节方法,该方法根据粒子之间的位置关系来设定参数值,给出了具体的设计思想与实现步骤。然后针对有限脉冲响应(finite impulse response,FIR)数字滤波器的优化设计实质,即多参数优化问题,通过适当的编码方式将改进的QPSO算法(adaptive QPSO,AQPSO)应用在其优化设计中,设计了低通和带通FIR数字滤波器。实验结果表明,AQPSO在收敛速度、鲁棒性及优化效果等方面都优于遗传算法(genetic algorithm,GA)、PSO算法及QPSO算法,说明了AQPSO算法的有效性和可行性。To study the parameter control method of quantum-behaved particle swarm optimization （QPSO） algorithm, an adaptive control method is proposed. In the proposed method, the parameter value is evaluated according to the relations between the particles. The design idea and implement steps are introduced detailed. With the optimization design of FIR digital filters being multi-parameter optimization problems, the proposed QPSO （AQPSO） algorithm is employed to design FIR digital filters. Experimental results on the low-pass and band-pass FIR digital filters show that AQPSO is superior to GA, PSO and QPSO in convergent speed, robustness and the optimization results. The effectiveness and feasibility of AQPSO are also demonstrated.

关 键 词：滤波器设计 FIR数字滤波器 粒子群优化算法 优化设计 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程] TN911.7[自动化与计算机技术—控制科学与工程]