摘要
针对基本蚁群算法在处理旅行商问题（TSP）时会出现收敛速度慢且容易陷入局部最优解的缺陷,从城市选择策略和信息素挥发系数进行了改进,提出了一种基于赌盘算法的城市选择策略和挥发系数自适应的蚁群算法,并采用了任务提前终止策略来减少算法的运行时间。仿真结果表明,该算法与基本蚁群算法相比,收敛时间比基本蚁群算法运行时间缩短了60%~80%,改进后的算法最优解绝大部分优于基本蚁群算法,也有少部分不如基本蚁群算法,但都在可接受范围以内。
The basic Ant Colony Algorithm（ ACA） has some problems when processing Traveling Salesman Problems（ TSP）,such as slow convergence speed and easy to fall into local optimum. In order to solve these problems,this paper improved the city selection strategy based on roulette wheel algorithm and proposed an improved ant algorithm based on adaptive volatile coefficient. The run-time was reduced by using the strategy early to terminate the task. The simulations reveal that the proposed algorithm has some advantages compared to basic ACA. Its run-time is 60% to 80% less than basic ACA,and the most of optimization of improved algorithm is superior to that of basic ACA. Of course,it also have some problems inferior to basic ACA but within acceptable range.
出处
《计算机应用》
CSCD
北大核心
2015年第A02期114-117,共4页
journal of Computer Applications
关键词
蚁群算法
动态自适应
赌盘算法
TSP
Ant Colony Algorithm（ACA）
dynamic adaptive
roulette wheel algorithm
Travelling Salesman Problem（TSP）
