An Adaptive Heterogeneous Multiple Ant Colonies System

This paper proposes an Adaptive Heterogeneous Multiple Ant Colonies System (AHMACS) Algorithm to solve the Traveling Salesman Problem (TSP). In this algorithm, heterogeneous multiple ant colonies with different pheromone updating mechanism have different searching traits. To leverage this feature, we design a set of adaptive rules to facilitate the collaboration between these colonies. An evolution coefficient is defined to adaptively determine ant sub-colonies´ information exchange partners and information exchange strategies. This approach can balance the diversity and convergence of solutions generated by different ant colonies. We use a set of TSP experiments to show that this algorithm can generate solutions with better quality and faster speed.