一种改进的自适应惯性权重粒子群优化算法  

Improved Particle Swarm Optimization with Adaptive Inertia Weight

作　　者：董平平[1] 高东慧[1] 田雨波[1] 胡永建 

机构地区：[1]江苏科技大学电子信息学院,江苏镇江212003 [2]国网电力科学研究所,江苏南京210003

出　　处：《计算机仿真》2012年第12期283-286,共4页Computer Simulation

基　　金：船舶工业国防科技预研基金项目(10J3.5.2);江苏高校优势学科建设工程资助项目

摘　　要：研究粒子群算法优化问题,针对基本粒子群算法早熟收敛,易收敛于局部极值的缺点,提出了一种改进的粒子群算法,采用对全局最优微扰和调整惯性权重的方法,改善算法的优化速度和收敛精度。利用个体寻优能力来定义惯性权重,并且将其控制在0.9-0.4范围内,从而合理地调整全局探索能力和局部开发能力。在每次迭代时对当前全局最优粒子进行微扰,改变它的位置,避免它陷入局部最优。经过对一系列测试函数的计算和比较,证明改进方法无论收敛速度、搜索精度及稳定性均有显著改善。In order to overcome the drawbacks of standard particle swarm optimization （PSO） algorithm, such as prematurity and easily trapping into local optimum, a modified PSO algorithm was proposed, in which special techniques including global best perturbation and changing inertia weight were adopted. The convergence speed and accuracy of the algorithm were improved. In this study, individual search ability was used to define inertia weight and make it in the range of 0. 9 - 0. 4. Then the global exploration and local exploitation abilities were rationally adjusted. At each iteration, a perturbation was given to the current global best particle to prevent it from falling into local optimum. The experiments about some benchmark problems show that it achieves higher performance, including convergence precision, convergence rate and stability.

关 键 词：粒子群优化 惯性权重 微扰 

分 类 号：TP202.7[自动化与计算机技术—检测技术与自动化装置]