一种快速收敛的改进粒子群优化算法  

Modified Partical Swarm Optimization Algorithm with Fast Convergence

作　　者：林卫星[1] 陈炎海[1] 

机构地区：[1]宁波大学信息科学与工程学院,宁波315211

出　　处：《系统仿真学报》2011年第11期2406-2411,共6页Journal of System Simulation

基　　金：浙江省教育厅重点项目(Z201017236);浙江省钱江人才项目(2011R10074);浙江省自然科学基金(Y107010);宁波市自然科学基金项目(2008A610019)

摘　　要：采用离散线性系统的状态方程,根据系统稳定性理论,推出了保证粒子群优化算法收敛性的参数设置区域。在收敛性理论分析的基础上,提出了一种快速收敛的改进粒子群优化算法,它是基于二阶系统按最佳阻尼比的思想来设定粒子群速度更新公式中的惯性权重。通过标准测试函数的性能测试,验证了改进粒子群优化算法的收敛性和快速性,并和惯性权重线性递减的标准粒子群优化算法进行了比较。仿真结果表明,该算法具有可靠的收敛性能和更快的收敛速度。An area of parameters was proposed in order to ensure the convergence of partical swarm optimization （PSO）. According to the system stability theory, the area was obtained in a discrete and linear state equation. A novel algorithm of modified partical swarm optimization （MPSO） was proposed for the fast convergence. In MPSO, the inertia weight of velocity renewal formula could be obtained based on the second-order system that has the best damping ratio. A performance test of benchmark functions was taken to confirm convergence and rapidity of MPSO and standard partical swarm optimization （SPSO） algorithm. Simulation results show that MPSO algorithm has more stable and rapid convergence than SPSO algorithm.

关 键 词：粒子群算法 阻尼比 收敛性 函数优化 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]