蚁群算法的原理及其改进  

Theory and Improvment of Ant Colony System

作　　者：陈晓梅[1] 

机构地区：[1]广东财经职业学院,广东广州510420

出　　处：《广东技术师范学院学报》2006年第4期68-70,共3页Journal of Guangdong Polytechnic Normal University

摘　　要：蚁群算法来源于对蚂蚁群体搜索行为的追踪研究,其基于信息素的正反馈特性有助于快速找到最优解。但蚁群算法也有不足之处,主要表现在当问题规模较大时,容易陷入局部最优化从而导致算法过早停滞。本文以旅行商(TSP)问题为基准,介绍了蚁群算法的原理,然后讨论了三种改进策略,主要表现在对其关键因子———信息量增量进行调整,这些改进策略有效地改善了蚁群算法过早停滞的现象。Ant colony system（ACS） comes from study of searching activity of ant colony, it helps to find the optimized solution based on its positive feedback of pheromone. But ACS has its defect, if the scale is too big it will result in algorithm premature stagnation due to falling into partial optimization. This article introduces theory of ACS based on Traveling Salesman Problem（TSP）, and then brings up three improvement methods. The methods are to adjust their key factors,pheromone, they effect on improvement the phenomenon of algorithm premature stagnation.

关 键 词：蚁群算法 信息素 信息素浓度 

分 类 号：TP311.12[自动化与计算机技术—计算机软件与理论]