Fuzzy Expert Ants to speed up big TSP Problems using ACS

Abstract

Ant colony algorithms are a group of heuristic optimization algorithms that have been inspired by behavior of real ants foraging for food. In these algorithms some simple agents (i.e. ants), search the solution space for finding the suitable solution. Ant colony algorithms have many applications to computer science problems especially in optimization, such as machine drill optimization, and routing. This group of algorithms have some sensitive parameters controlling the behavior of agents, like relative pheromone importance on trail and pheromone decay coefficient. Convergence and efficiency of algorithms is highly related to these parameters. Optimal value of these parameters for a specific problem is determined through trial and error and does not obey any rule. Some approaches proposed to adapt parameter of these algorithms for better answer. The most important feature of the current adaptation algorithms are complication and time overhead. In this paper we have presented a simple and efficient approach based on fuzzy logic for optimizing ACS algorithm and by using different experiments efficiency of this proposed approach has been evaluated and we have shown that the presented concept is one of the most important reasons in success for parameter adapting algorithms.