针对蚁群算法在求解交通网络两点之间最短路径时存在收敛速度慢和容易出现停滞现象等缺点,为提高搜索效率,提出了一种改进的蚁群算法。通过在初始化信息素时加入方向引导因素,减少了劣质解,提高了解空间的质量;设计一个动态因子,使其自适应地更新全局信息素,很好地利用了较优的解,提高了全局搜索能力,避免算法求解出现早熟。仿真结果表明,不但在收敛速度有大幅度地提高,而且在避免易于陷入局部最优解方面取得了很好的效果。实例证明了改进算法是可行有效的。
Considering that the ant colony algorithm in solving the shortest path between two points of transport network has the shortcomings such as slow convergence and prone to stagnation phenomenon, the authors provided an improved ant algorithm which adds the heuristic direction information, reduces the inferior solution, improves the quality of solution, and designs a dynamic factor to adaptively adjust the renewal of pheromone on the optimal solu- tion. The algorithm is more conducive to optimal path, can improve the capability of global search and avoid the algo- rithm premature. The results of experiment show that the improved algorithm enhances the convergence speed effec- tively and avoids getting into local optimal easily.