改进粒子群算法的研究与仿真  

An improvement algorithm of particle swarm optimization

作　　者：李凌[1] 倪远平[1] 孙婧雅[1] 

机构地区：[1]昆明理工大学信息工程与自动化学院,云南昆明650051

出　　处：《云南大学学报（自然科学版）》2009年第S2期178-181,186,共5页Journal of Yunnan University(Natural Sciences Edition)

摘　　要：针对标准粒子群算法(PSO)存在收敛慢、易陷入局部极值的缺点,分析了粒子群算法中惯性权重和加速因子的作用,对它们作了修改,进行了相关的实验仿真.仿真结果表明:修改后的粒子群算法迭代次数减少了,收敛性有了明显的提高,全局搜索和局部搜索能力达到平衡,有利于更快地寻找全局最优解.Because the standard particle swarm optimization(PSO)algorithm has some shortcomings of slow convergence and easy to fall into the local extreme value,we analyzed the effects of inertia weight and accelerating factor in particle swarm optimization,modified the inertia weight and accelerating factor,and carried out the relational test simulation.The simulation results showed that the number of iterations has reduced,convergence has improved,the global search and local search capabilities can achieve balance,and the global optimal solution can be found quickly.

关 键 词：粒子群算法 惯性权重 加速因子 最优解 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]