Abstract: To settle the contradictory between convergence speed and precocity and stagnation in ant colony algorithm, an ant colony algorithm, which is based on adaptive selection of the paths and dynamic updating of pheromone, is presented. By dynamically adjusting the strategy of selection of the paths and the strategy of the trail information updating according to the distribution of the solutions, the algorithm can keep good balance between accelerating convergence and averting precocity and stagnation. Experimental results on traveling salesman problem show that the method presented in this paper has a better global searching ability, higher convergence speed and solution diversity than that of classical ant colony algorithm.

Key words: traveling salesman problem, ant colony algorithm, pheromone, dispersed degree

摘要： 针对蚁群算法加速收敛和早熟、停滞现象的矛盾，提出一种基于自适应路径选择和动态信息素更新的蚁群算法，以求在加速收敛和防止早熟、停滞现象之间取得很好的平衡。该算法根据优化过程中解的分布状况，自适应的调整路径选择策略和信息量更新策略。基于旅行商问题的实验验证了算法比一般蚁群算法具有更好的全局搜索能力、收敛速度和解的多样性。

关键词: 旅行商问题, 蚁群算法, 信息素, 分散度