A Two-Stage Optimization Approach for Multi-Depot Vehicle Routing Problem with Time Windows

The surge in market demand poses challenges regarding delivery time and costs in logistics and distribution vehicle routing problems. However, conventional algorithms often fall short of addressing such complexities effectively. In this study, we tackle the optimization of multi-depot vehicle routing problem with time windows(MDVRPTW), aiming to minimize vehicle and transportation costs while optimizing distribution time. Key constraints, including vehicle loading capacity and soft time windows, are considered. Leveraging the K-means clustering algorithm, we simplify the problem-solution process. Additionally, we enhance the ant colony algorithm(ACO) with adaptive state transfer rules and dynamic pheromone updating strategies, complemented by eight local search operators, to bolster optimization capabilities. Employing the Solomon Corporation dataset, we benchmark our approach against established algorithms such as Simulated Annealing (SA), Genetic Algorithm (GA), and Particle Swarm Optimization (PSO). Results indicate that our proposed algorithm outperforms competitors in 18 out of 24 test cases, underscoring its efficacy and stability. Overall, these findings validate the efficiency and robustness of our proposed methodology.