A Novel Adaptive Ant Colony Algorithm with Application

Ant algorithm is a novel meta-heuristic optimization algorithm. Preliminary study has shown that the algorithm has great ability of searching better solution, but at the same time there are some shortcomings such as tending to go into stagnation behavior and needing longer computing time. In order to overcome the shortcoming of basic ant algorithm, a new ant algorithm, Adaptive Ant Colony Algorithm (AACA), is proposed in this paper. In AACA, the transition probability with which ant used to select next city is dynamically adjusted based on the number of average node branch to avoid going into stagnation behavior. The ability of searching better solution is great improved in this way. Simulated experiments show that AACA has a good ability of searching better solution in the last runs of the algorithm.