有限级信息素蚁群算法

Ant Colony Optimization Algorithm with Finite Grade Pheromone

提出一种新的蚁群算法,将信息素分成有限个级别,通过级别的更新实现对信息素的更新,并且信息素的更新量独立于目标函数值．文中采用有限马氏链的理论证明算法可以线性地收敛到全局最优解．针对TSP问题,通过与MMAS和ACS等蚁群算法的数值实验结果进行比较,表明所提出的算法是有效的、鲁棒的．

In the paper, a new class of ant colony optimization algorithm is proposed, in which pheromone is classified into finite grades, pheromone updating is realized by changing the grades, and the updated quantity of pheromone is independent of the objective function values. It is proved by means of finite Markov chains theory that the algorithm converges to the global optimal solutions linearly. Compared with MMAS, ACS and some other ant colony optimization algorithms for the Traveling Salesman Problem, the calculating results demonstrate that the proposed algorithm is effective and robust.