粒子群与遗传算法的混合算法  

Mixed Research on Particle Swarm Optimization and Genetic Algorithm

作　　者：阳琼芳[1] 孙如祥[1,2] 

机构地区：[1]广西职业技术学院计算机与电子信息工程系,广西南宁530226 [2]广西大学计算机与电子信息学院,广西南宁530004

出　　处：《华侨大学学报（自然科学版）》2015年第6期645-649,共5页Journal of Huaqiao University(Natural Science)

基　　金：广西高校科学技术研究项目(2013YB295)

摘　　要：针对粒子群算法直接用于求解离散旅行商优化问题会存在诸多困难,通过分析粒子群算法、遗传算法各自优缺点,将粒子群算法、遗传算法有效结合组成混合算法用于求解离散旅行商问题.混合的目的在于保持两种算法各自的优点,并有效地避免各算法原有的不足.对3个不同规模的巡回旅行商问题进行实验,结果表明:混合算法提升了算法的局部搜索能力.There are many difficulties when particle swarm optimization is used directly to solve discrete travelling sales- man problem （TSP） optimization problems. Therefore, we analyze the advantages and disadvantages of particle swarm optimization algorithm and genetic algorithm, and then mix them to be an effective algorithm to solve discrete TSP. The purpose of combination is to keep the original advantages of the two kinds of algorithms and to avoid the existing deficien- cies. We conduct some experiments on the 3 TSP problems different scales. The result shows that the hybrid algorithm can highly improve the local search ability of algorithm.

关 键 词：离散旅行商问题 遗传算法 粒子群算法 自适应 启发策略 

分 类 号：TP302[自动化与计算机技术—计算机系统结构]