适应度排序改进惯性权重的粒子群算法  

Use adaptation sequence modified inertia weight particle swarm optimization

作　　者：陶俊波[1,2] 吴彰敦[1] 蔡德所[1,3] 

机构地区：[1]广西大学土木建筑工程学院,南宁530004 [2]福建水利电力职业技术学院,福建永安366000 [3]三峡大学土木水电学院,湖北宜昌443002

出　　处：《计算机工程与应用》2009年第14期53-55,57,共4页Computer Engineering and Applications

摘　　要：改进PSO算法的惯性权重。惯性权重不仅随代数纵向线性变化,也根据当前和迄今粒子的适应度重排序横向线性变化。横向线性变化上限不变,下限逐渐减小,使得横向线性变化数值范围随代数逐渐增大。惯性权重数值随着代数逐渐取负,并且适应度差的粒子取负的几率更大。得到基于粒子适应度排序改进惯性权重的粒子群算法(ASMIWPSO算法)。通过仿真学解释ASMIWPSO算法。Rastrigrin函数测试对比ASMIWPSO算法、PSO算法,说明ASMIWPSO算法具有更好的优化结果。It has modified PSO inertia weight.The inertia weight not only through its longitudinal linear change by generation;but also think about current and up to now best results of particles adaptation sequence＇s lateral linear change which rearrange by adaptation good or bad.The lateral linear upper bound un-change and lower bound become small,so the lateral linear range expend gradually.The inertia weight will generate more negative value by generation increasing.That obtain use adaptation sequence modified inertia weight particle swarm optimization（ASMIWPSO）.It has been through simulation to explain ASMIWPSO.And make contrastive analysis on the results of PSO and ASMIWPSO by Rastrigrin function.It shows that the ASMIWPSO has better optimization results.

关 键 词：粒子群算法 惯性权重 适应度排序 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]