基于拓展性和魔方变换的自适应蚁群算法

Adaptive Ant Colony Algorithm Based on Expandability and Magic Cube Transformation

针对传统蚁群算法在求解过程中搜索时间过长、易于出现早熟停滞的缺陷,提出一种具有拓展性的自适应蚁群算法.蚁群综合启发式信息、信息素轨迹和拓展性信息自适应地调整状态转移规则,并采用全局信息素非均匀更新策略,有效增强了蚁群的全局搜索能力.同时,受魔方变换的启发,提出了一种新颖的魔方变异策略,以加快对迭代最优解进行局部优化的速度.旅行商问题仿真验证了文中改进蚁群算法的有效性,其收敛速度、稳定性远高于传统蚁群算法.

There are the shortcomings such as longer computing time and precocity and stagnation in classical ant colony algorithm. Based on the expandability, an adaptive ant colony algorithm is presented. The algorithm dynamically adjusts state transition rule by integrating expandability with heuristics and pheromone. Meanwhile, an uneven strategy based on the global pheromone updating is adopted to enhance the ant＇s excellent ability in searching the whole best solution. In addition, a novel magic cube mutation strategy, inspired by the magic cube transformation, is employed to accelerate evolution speed after each iteration. The experimental results on TSP demonstrate that the proposed algorithm has much higher convergence speed and Stability than that of classical ant colony algorithm.