A parallel ant colony optimization for multi-depot vehicle routing problem

This paper presents a method for solving the multi-depot vehicle routing problem (MDVRP). Due to different vehicle capacities among companies, a mathematical formulation, which considers the variable costs and the vehicles' fixed costs simultaneously, is given. Since the inherent complexity of the MDVRP makes it difficult to be solved even for a relatively small scale, an improved ant colony optimization with a new strategy to update the increased pheromone (Ant-Weight) is developed for the MDVRP. Then, a parallelization strategy for ant colony optimization is used to increase computational efficiency. Finally, the proposed algorithm is examined with the data of Wuhu city in China. The results indicate that this method performs well in terms of the solution quality and run time consumed.