粒子群优化算法中粒子更新方法研究  

Research on updating algorithms in particle swarm optimization

作　　者：田雨波[1] 朱人杰[1] 李正强[1] 

机构地区：[1]江苏科技大学电子信息学院,江苏镇江212003

出　　处：《江苏科技大学学报（自然科学版）》2008年第5期67-72,共6页Journal of Jiangsu University of Science and Technology:Natural Science Edition

基　　金：江苏省高校自然科学基础研究资助项目(07KJB510032);江苏省高校"青蓝工程"优秀青年骨干教师培养资助项目

摘　　要：粒子群优化算法是根据鸟或鱼群居社会行为而提出的随机优化算法,但标准粒子群优化算法存在早熟收敛和搜索精度低等问题.因此模拟生物克隆选择中5%的B细胞自然消亡过程,在粒子群优化算法进化过程中分别基于代间差分、混沌理论、变异原理等方法设计了8种粒子更新算法,并按照模拟退火方法进行更新后粒子的选择.通过数值实验得出基于代间差分和混沌变异的粒子更新算法(即算法8)是一种很好的选择,并且当性能较差的20%左右粒子按照这种算法更新时效果较好.这种算法可以有效克服标准粒子群算法的早熟现象,并能够加快收敛速度.PSO（Particle Swarm Optimization） is a stochastic optimization algorithm inspired by social behavior of bird flocking or fish schooling. However, the standard PSO has some shortcomings, such as premature convergence, searching precision lowness and so forth. Based on the simulation of natural death process of 5 % B-cell in biology clone selection, this paper proposes 8 kinds of updating algorithms according to intergeneration differential, theory of chaos, principle of mutation respectively, and selects the updated particles in terms of simulated annealing method. Numerical experiments show that the updating algorithm by using intergeneration differential and chaotic mutation （ algorithm 8） is a good selection. Simultaneously, the updating effect is perfect when the updated particles are about 20%. The algorithm may overcome effectively the premature problem and speed up the convergence.

关 键 词：粒子群算法 克隆选择 混沌 变异 模拟退火 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]