Multi-colony ant optimization with dynamic collaborative mechanism and cooperative game

Ant Colony Optimization easily falls into premature stagnation when solving large-scale Travelling Salesmen Problems. To address this problem, a multi-colony ant optimization with dynamic collaborative mechanism and cooperative game is proposed. Firstly, Ant Colony System and Max-Min Ant System form heterogeneous colonies. Secondly, to diversify the solutions of the algorithm, the Shapley value in the cooperative game is applied to share the information by distributing the pheromone payoff of the sub-colonies. In addition, the dynamic collaborative mechanism that contains two methods is designed to enhance the co-evolution of the heterogeneous populations. One, called public path recommendation strategy, is proposed to improve the astringency of Max-Min Ant System. The other is the pheromone fusion mechanism to regulate the pheromone distribution of Ant Colony System when the algorithm falls into stagnation, which can help the algorithm jump out of the local extremum effectively. Finally, the results demonstrate that the proposed methodology can improve the accuracy of solution effectively in solving large-scale TSP instances and has strong competitiveness with other swarm intelligent algorithms.