粒子带质量的粒子群算法  

Improved particle swarm algorithm which considers quality of particle

作　　者：王柏竹[1] 邢光龙[1] 

机构地区：[1]燕山大学信息科学与工程学院,河北秦皇岛066004

出　　处：《计算机工程与应用》2011年第17期31-33,共3页Computer Engineering and Applications

基　　金：河北省自然科学基金(No.D2008000767);燕山大学博士基金(No.B121)

摘　　要：针对粒子群算法在陷入局部最优时难于跳出的缺陷提出了一种带有质量的粒子群算法。该算法受运动学原理启发,粒子位置的更新不仅受自身最优和种群最优的影响,还受到由粒子质量引起的梯度场的影响。当粒子群出现早熟现象时,用电磁学原理与动量守恒定理更新种群的最优位置,使群体能及时摆脱局部最优区域。仿真结果表明,该算法优化4种具有代表性的基准函数,无论是在优化精度方面还是在优化效率方面,均较以往提出的改进粒子群算法在性能上有所改进。To solve the problem that once Particle Swarm Optimization（PSO） algorithms find a local optimization it is hard for them to jump out and continue a global optimization, an improved particle swarm optimization which considers the quality of the particle is proposed.This algorithm is inspired by the kinematic theory.The particle is updated not only by the best previous position and the best position among all the particles in the swarm, but also by the role of the gradient field which is caused by the particle mass.When the particle swarm appears premature convergence,law of conservation of momentum is used to update the best previous position so that groups jump out from the local optimum area in time.Simulation results show that, compared with other improved PSO algorithms proposed before, it improves both optimization precision and efficiency when the improved PSO algorithm is used to optimize 4 typical benchmarks.

关 键 词：粒子群算法 粒子质量 运动学原理 梯度场 电磁学 动量守恒定理 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]