双种群改进蚁群算法

基本蚁群优化（Basic Ant Colony Optimization，BACO）算法在进化中容易出现停滞，其根源是蚁群算法中信息的正反馈．在大量蚂蚁选择相同路径后，该路径上的信息素浓度远高于其他路径，算法很难再搜索到邻域空间中的其他优良解．对此，提出一种双种群改进蚁群（Dual Population Ant Colony Optimization，DPACO）算法．借鉴遗传算法中个体多样性特点，将蚁群算法中的蚂蚁分成两个群体分别独立进行进化，并定期进行信息交换．这一方法缓解了因信息素浓度失衡而造成的局部收敛，有效改进算法的搜索性能，实验结果表明该算法有效可行．

Dual population ant colony optimization algorithm

The Basic Ant Colony Optimization (BACO) algorithm often gets into premature stagnation during evolution due to the positive feedback of the pheromone. When a mass of ants select the same path, the pheromone on the selected path is denser than those on others, so the algorithm is difficult to explore other solutions in the neighbor space. And a Dual Population Ant Colony Optimization (DPACO) algorithm is presented to deal with it. Referred to the individual diversity feature in genetic algorithm, the algorithm separates the ants into two populations which evolves separately and exchanges information timely. The method can restrain the local convergence caused by the misbalanced of the pheromone and can improve the searching performance of the algorithm effectively. The experiment indicates that the algorithm is efficient and feasible.