蚁群算法及其实现方法研究  

Research on the Ant Colony Optimization and Its Implementation Strategy

作　　者：胡娟[1] 王常青[1] 韩伟[1] 全智[1] 

机构地区：[1]中国科学院软件研究所,北京100080

出　　处：《计算机仿真》2004年第7期110-114,共5页Computer Simulation

基　　金：中国科学院知识创新工程方向性研究课题资助项目(KGCX2-JG-09)

摘　　要：蚁群算法是一种相对较新的启发式方法,通过模拟蚂蚁的觅食行为解决问题,是目前昆虫算法中较成功的例子。蚁群算法的本质是一种并行的、自组织的算法,它可应用于更好地组织大数目实体的相互作用过程,如货郎担问题、车辆绕径问题、排程问题等。该文简述了蚁群算法的起源和发展,总结了蚁群算法的特点和不足及针对这些不足提出的各种改进方法,并介绍了和蚁群算法相关的几种具体应用。最后,文章探讨了蚁群算法研究中仍存在的问题和以后的发展方向。The Ant Colony Optimization(ACO) is a relatively new meta-heuristic algorithm and a successful paradigm of all the algorithms which take advantage of the insects' behavior. The ACO solves problems through mimicking ants' foraging behavior. Essentially, the ACO is a parallel and self-organizing algorithm, which can be applied to improve the management and control of large numbers of interacting entities such as TSP(Travelling Salesman Problem), Vehicle Routing Problem and Scheduling Problems. This article presents the origin and enrichment of the ACO, summarizes this algorithm's advantages,disadvantages,methods to overcome these disadvantages, and introduces some applications of the ACO. After discussing several problems existing in the research,this article puts forward the research foreground of ACO.

关 键 词：蚁群算法 行为启发 AS算法 ACO算法 信息素 局部搜索 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]