改进蚁群算法在二次分配问题中的应用  

Application of Improved Ant Colony Algorithm for Quadratic Assignment Problems

作　　者：袁东锋[1] 吕聪颖[1] 

机构地区：[1]南阳理工学院计算机与信息工程学院,河南南阳473000

出　　处：《计算机与现代化》2013年第3期9-11,16,共4页Computer and Modernization

基　　金：国家自然科学基金青年科学基金资助项目(81101490);教育部重点基金资助项目(02090)

摘　　要：为了解决基本蚁群算法在求解大规模二次分配问题时暴露出的缺陷,本文提出一种改进的蚁群算法。在基本蚂蚁算法中,采用全局信息素更新策略,使用距离及流量作为启发式信息并引入局部优化策略,对每代的最优解进行改进,进一步加快算法的收敛速度。通过对于二次分配问题的3种不同类型的问题进行实验,将改进的蚁群算法与基本蚂蚁算法及混合遗传算法进行比较,结果表明该改进算法具有更优的性能。In order to solve the problems that the basic ant colony algorithm for solving large scale quadratic assignment has revealed defects,this paper proposes an improved ant colony algorithm.This algorithm adopts the global pheromone update strategy,the use of distance and traffic as heuristic information and the introduction of local optimization strategy.The optimal solution for each generation is to improve and further accelerate the convergence speed.For the quadratic assignment problem through three different types of problems,and improved ant colony algorithm with the basic ant algorithm and the hybrid genetic algorithm are compared,the experiments show that the improved method has better performance.

关 键 词：改进蚁群算法 二次分配问题 全局 顶点重定位 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]