多信息结合离散粒子群算法及其应用  

More information with discrete particle swarm algorithm and its application

作　　者：任小波[1] 陈舒骅[2] 

机构地区：[1]宁波工程学院电子与信息工程学院,浙江宁波315016 [2]武汉大学计算机学院,武汉430072

出　　处：《计算机工程与应用》2011年第32期54-55,74,共3页Computer Engineering and Applications

基　　金：宁波市自然科学基金资助项目(No.2009A610090);浙江省教育厅项目(No.Y200906724)

摘　　要：为提高粒子群算法的搜索性能,提出了一种改进的离散粒子群算法:多信息结合离散粒子群算法。该算法在粒子群算法的基础上借鉴蚁群算法的信息素机制,重新定义了粒子的速度位置更新公式,并且引入双曲正切函数对粒子群进行初始化。通过求解背包问题对算法进行验证,实验结果表明所提算法性能较优。To improve the search capability of particle swarm algorithm,an improved discrete particle swarm optimization algorithm is proposed and named as More Information with discrete Particle Swarm algorithm（MIPSO）.The idea of phero- mone refresh mechanism of ant colony algorithm is used and the update equations of the speed and position of particles are redefined and a hyperbolic tangent function is used to initialize the particle swarm on the basis of PSO algorithm.This algorithm is verified by solving knapsack problem, the results of the experiment show that the proposed algorithm can result in better profits.

关 键 词：离散粒子群算法 信息素机制 双曲正切函数 背包问题 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]