求解TSP问题的蚁群算法研究

Research of Ant Colony Algorithm on TSP Problems

蚁群算法是一种新型仿生算法,但存在搜索时间长,收敛速度慢,易陷入局部最优等缺点.提出了一种改进蚁群算法,利用象限近邻表构造候选集和对偶象限近邻的方法初始化信息素,可以克服上述缺陷.TSP的仿真结果表明新算法大大缩小了其搜索范围,提高了搜索精确度并减少了搜索时间.

Ant Colony optimization （ACO） is a novel metaheuristic algorithm, which has the limitations of stagnation and poor convergence, and is easy to fall in local optima. An improved ant algorithm based on quadrant nearest neighbor list is presented to solve these shortcomings, constructing candidate list by quadrant neighbor list and initializing pheromone table by dual quadrant neighbor method. The simulation for TSP shows that the improved algorithm can efficiently reduce the scope of solution space and searching time, and enhance the precision.