一种改进的自适应蚁群算法及其应用研究

蚁群算法作为一种新型的模拟进化算法，具有分布计算和信息正反馈等优点，但蚁群算法与其他进化算法一样存在收敛速度慢，易陷于局部最优等缺陷。针对这一问题，提出一种改进的蚁群算法，结合遗传算法和图论中的最邻近算法，并自适应地初始化信息素和限定信息素的大小范围。将该算法应用于旅行商问题（traveling salesman problem，TSP）求解，与基本蚁群算法比较，数值实验结果表明，这种改进算法能有效抑制算法陷入局部最优的缺陷，从而提高了解的全局搜索能力和解的质量。

An improved adaptive ant colony optimization algorithm with application

Ant Colony Optimization algorithm is a novel evolutionary simulating algorithm, which has the merit of distributed computation and positive feedback of information. But Ant Colony Optimization algorithm also has some shortcomings such as needing longer computing time and tending to go into local optimum as other evolutionary algorithms. To solve this problem, An improved Ant Colony Optimization is proposed. The algorithm combines genetic algorithms and the nearest neighbor algorithm of graph theory to adaptively initialize pheromone and limit the extent of pheromone. To apply this algorithm to TSP problem, numerical results show that this algorithm compared with fundamental algorithm can effectively suppress falling into local optimization, as a result it enhances the ability of global searching and leads to better results.