面向TSP求解的混合蚁群算法  

Hybrid Ant Colony Algorithm for TSP

作　　者：张泓[1] 李爱平[1] 刘雪梅[1] 

机构地区：[1]同济大学现代制造技术研究所,上海200092

出　　处：《计算机工程》2009年第8期34-37,共4页Computer Engineering

基　　金：上海市"十一五"重点科技攻关计划基金资助项目(06DZ22915;06DZ11302;071111005)

摘　　要：针对蚁群算法的早熟和停滞等现象,将免疫算法机制引入蚁群算法,提出用于TSP求解的混合算法。该算法具有蚁群算法的自适应反馈机理、收敛速度快和免疫算法操作算子简单和维持种群多样性、防止种群退化等特性。从算法解的质量与效率方面与基本蚁群算法和免疫算法进行比较,结果表明融合免疫机制的蚁群算法性能显著提高,也为解决其他组合优化问题提供一个新的思路。Aiming at the phenomena such as precocity and stagnation of ant colony algorithm, the mechanism of immunity algorithm is put into ant colony algorithm. This paper presents a new hybrid algorithm absorbed the adaptive feedback mechanism and fast convergence merits of ant colony algorithm and simple operation operators and keeping diversity of groups without degenerating merits of immunity algorithm. The new hybrid algorithm is compared with basic ant colony algorithm and immunity algorithm from the aspects of solution quality and algorithm efficiency. Results show that the new hybrid algorithm performance is improved markedly. It also provides a new idea for solving other combinational optimization problems.

关 键 词：蚁群算法 免疫算法 旅行商问题 混合算法 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]