粒子群算法的改进及其在优化函数中的应用  

Improvement of Particle Swarm Algorithm and Its Application in Optimization Function

作　　者：马发民 张林[1] 王锦彪[2] 

机构地区：[1]商洛学院数学与计算机应用学院,商洛726000 [2]中国民航大学计算机科学与技术学院,天津300000

出　　处：《计算机与数字工程》2017年第7期1252-1255,1293,共5页Computer & Digital Engineering

基　　金：国家自然科学基金项目(编号:60472121);商洛学院自然科学研究项目(编号:15SKY007)资助

摘　　要：针对粒子群优化算法自身的缺陷,即随着迭代次数的增大,种群多样性减小,引起早熟现象,从而可能出现局部最优结果。而生物免疫机制能够有效地克服这些缺点,因此将粒子群算法与免疫原理有机结合起来形成免疫粒子群优化算法(IMPSO);其次对PSO算法的惯性系数和学习因子做了一定的改进;最后通过经典优化函数的计算,验证了算法改进的效果。The defection of particle swarm optimization algorithm means that an increase in iterations decreases swarm diversi-ty and causes prematurity, thus probably producing local optimization results. However, immune mechanism of biology is capable of effectively overcoming these shortcomings. Firstly particle swarm algorithm is organically combined with immune principle to form immune particle swarm optimization algorithm （IMPSO） , then certain improvements will be made in inertia coefficient and learning factor of PSO algorithm and finally effect of algorithm improvement will be verified through calculation of typical optimization function.

关 键 词：粒子群优化算法 免疫原理 免疫粒子群优化算法 惯性系数 学习因子 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]