摘要
微粒群算法是一种新颖的群智能仿生进化优化算法,其原理简单,控制参数少,容易实现,在连续空间中有很强的优化能力。研究了将微粒群算法应用于基于排序的组合优化问题,进行了算法设计,给出了算法的流程,提出了计算两个排列的差及由置换求微粒群算法的速度的具体操作方法。为加快算法的收敛速度,增强全局搜索能力,运用矩阵的逐行最小元法来初始化微粒群,引入了突变算子。对一些测试旅行商问题利用新算法进行了模拟仿真,结果表明算法是可行的。
Particle swarm optimization （PSO） is a new swarm intelligence simulation evolution optimum algorithm. It has a simple principle, few controlling parameters, and strong optimizing ability in continuous space. It is easy to be implemented. The applying PSO to combination optimum problems is studied based on permutation, the flow of the algorithm is proposed, the computational method of difference between two permutations is designed, the method of deriving PSO＇s velocity from transforms, and the detailed operations is shown. In order to increase the convergent speed and to improve the overall searching ability of the algorithm, the rule of matrix minima line-by-line to is used initiate the particle swarm, and introduces a mutation operator. Finally, the paper makes simulating computation on some testing TSP problems. The results show that the algorithm is feasible.
出处
《计算机工程与设计》
CSCD
北大核心
2006年第21期4025-4027,共3页
Computer Engineering and Design
基金
广东省科技攻关基金项目(A10202001)
广州市科技攻关基金项目(2004Z2-D0091)
