一种非线性递减惯性权重策略的粒子群优化算法  

A Particle Swarm Optimization Algorithm with the Strategy of Nonlinear Decreasing Inertia Weight

作　　者：李会荣[1] 高岳林[2] 李济民[2] 

机构地区：[1]商洛学院数学系,陕西商洛7260000 [2]北方民族大学信息与系统科学研究所,宁夏银川750021

出　　处：《商洛学院学报》2007年第4期16-20,共5页Journal of Shangluo University

基　　金：国家民委科研项目基金资助(05XBE05);宁夏高等学校科研项目资助(2007)

摘　　要：目的改进基本粒子群算法的一些缺点.基本粒子群算法是一种有效的寻找函数极值的演化计算方法,它简便易行,收敛速度快.但此算法也存在收敛精度不高,易陷入局部极值点的缺点.方法对原有算法中的固定惯性权重进行改进.结果提出一种非线性递减惯性权重策略的粒子群优化算法.结论对六种具有代表性的测试函数进行了仿真实验,并与基本粒子群算法中惯性权重分别取固定权重、线性递减权重进行了比较,说明了非线性递减惯性权重策略的粒子群优化算法具有更好的性能和全局搜索能力。The standard Particle Swarm algorithm is a validated evolutionary computation way of searching the extremes of function, which is simple in application and quick in convergence, but low in precision and easy in premature convergence. Because of this, a particle swarm optimization algorithm with the strategy of nonlinear decreasing inertia weight is proposed based on the standard particle swarm algorithm. It is show that a particle swarm optimization algorithm with the strategy of nonlinear decreasing inertia weight is better in global searching and performance by comparing with six representative benchmarks in fixed weight, the linear weight and non-linear weight respectively.

关 键 词：粒子群优化 线性递减 惯性权重 非线性递减 

分 类 号：O141.3[理学—数学]