Mathematical Model and Algorithm Design for Collaborative Vehicle Routing in Three-echelon Transportation with Complete Information Sharing
LYU Xiaohui,WANG Nengmin,YANG Zhen
DOI: https://doi.org/10.12005/orms.2023.0083
2023-01-01
Abstract:Intercity goods distribution for e-commerce is usually realized in a three-echelon transportation network, where the delivery of goods from suppliers to customers passes two urban consolidation centres, one located in the origin city, and the other in the destination city, before arriving at the customers. In order to improve distribution efficiency, service quality, and reduce transportation costs, many cities in our country have carried out collaborative transportation pilot work, established resource-sharing urban consolidation centers, and constructed a collaborative three-echelon transportation network. The continuous development of digital platform technology provides information sharing foundation for collaborative transportation, reduces information asymmetry between logistics companies, integrates the supply and demand information of transportation requests, and reduces the empty load rate and improves transportation efficiency. The characteristics of transportation request delivery timeliness and mutual coupling in a three-echelon transportation network put forward new challenges for collaborative transportation. Based on the operation scenario of the digital platform providing information sharing for collaborative transportation, this paper studies a collaborative transportation problem in three-echelon transportation considering time windows constraints with the minimum of transportation cost. In this problem, all requests are served in a three-echelon transportation network, where goods are first delivered from suppliers located in a city（the city of origin） to their customers in another city（the city of destination） through two urban consolidation centers located at the outskirt of the two cities. Multiple shippers and customers are clustered in pick-up zones and delivery zones respectively. To handle the complicated linking constraints in a collaborative three-echelon transportation network, a multi-start iterated local search algorithm is proposed to obtain the approximate optimal solution of the problem. In the proposed algorithm, based on the coordination constraints among different echelons and the characteristic that each request in a three echelon transportation network must be served by the same carrier, the neighborhood structure is constructed by adding local branching constraints in local search operations and two perturbation strategies are adopted to select free variables in the perturbation operator. Extensive numerical experiment results have proved the effectiveness of the proposed algorithm, and demonstrated that in a three-echelon transportation network, collaborative transportation can significantly reduce the overall transportation cost. Moreover, the lower the proportion of reserved transportation requests, the more the number of transportation requests, and the larger the transportation network, the higher the possibility of collaboration among carriers, and the higher benefits can be achieved by collaborative transportation. This research enriches the study of collaborative transportation and provides quantitative decision support for carriers to solve the collaborative vehicle routing optimization problem in three-echelon transportation considering time windows with complete information sharing. However, this paper only considers the certain transportation requests, and ignores the randomness of transportation requests in the reality. Also, the proposed algorithm can only get approximate solutions for medium size and large size instances. In the future, the randomness of transportation requests should be considered when formulating a mathematical model and designing an algorithm to solve the problem with random transportation requests. In addition, accurate algorithms such as branch and cut algorithm and branch and cut and price algorithm can be adopt to obtain optimal solutions for this problem.