【摘要】 针对基本蚁群算法存在收敛慢且常常停滞于局部最优的缺点,提出一种改进的蚁群算法,即将蚁群分为两个种群:一个种群的蚂蚁按照基本蚁群算法进行搜索;另一个种群的蚂蚁在选择路径时则考虑路径上蚂蚁密集度的因素,以减少算法初期信息素的正反馈。在每轮迭代搜索完成后对路径上的信息素进行更新时,对比当前全局最优路径与迭代最优路径,从而减少非优路径上的信息素增量,尽量淘汰劣质解的路径,加快收敛速度。仿真实验表明,改进后的蚁群算法比基本蚁群算法更快收敛于全局最优解。更多还原

【Abstract】 The traditional ant colony algorithm often tends to converge slowly and is detained at local points of optimization. To solve this problem two ant colonies are introduced: one colony searches its route according to basic ant colony algorithm,and the other selects its route based on ant crowding degree along the path. When the search is finished after each interation,by comparing global optimal path with current iteration optimal path,the inferior solution path is eliminated and the speed of convergence accelerated.Simulations indicate that this new method could avoid bad routes and speed up the convergence,and finally lead to a global optimal solution after certain rounds of iteration,and this algorithm is much faster than the traditional algorithm.更多还原