Analyzing the Effects of Instance Features and Algorithm Parameters for Max–Min Ant System and the Traveling Salesperson Problem

Ant colony optimization (ACO) performs very well on many hard optimization problems, even though no good worst case guarantee can be given. Understanding the effects of different ACO parameters and the structural features of the considered problem on algorithm performance has become an interesting problem. In this paper, we study structural features of easy and hard instances of the Traveling Salesperson problem for a well-known ACO variant called Max Min Ant System MMAS) for several parameter settings. The four considered parameters are the importance of pheromone values, the heuristic information, the pheromone update strength and the number of ants. We further use this knowledge to predict the best parameter setting for a wide range of instances taken from TSPLIB.