Our paper is now published in Expert Systems with Applications (joint work with Wanru Chen, Dezhi Zhang, Guangming Xu, and Jing Guo). You can download it here.
This paper addresses the significant carbon emissions the transportation sector generates, focusing on the freight transportation segment, especially perishable freights in the cold chain. Cold chain logistics contribute nearly half of all transport-related carbon emissions. The paper emphasizes the need to adopt clean energy solutions, such as electric vehicles (EVs), and optimize routing and transportation schedules to mitigate these emissions. However, implementing EVs faces challenges due to range anxiety caused by limited battery capacity and charging infrastructure. The authors suggest a mixed fleet of EVs and traditional vehicles to balance operational costs and reduce emissions while improving service efficiency.
The paper highlights the role of government policies and regulations in promoting sustainable practices, such as implementing a carbon tax, which influences decision-making in cold chain logistics enterprises. The authors acknowledge the complexity of optimizing vehicle routes in this distribution network and propose a Cold Chain Green Multi-Depot Vehicle Routing Problem with Time Windows and Mixed Fleets (CC-GMD-VRPTW-MF). They establish a mathematical model and solve it using an improved Variable Neighborhood Search (VNS) algorithm tailored for this specific problem, with new mechanisms designed for perturbation and local search. The paper validates the model and algorithm through real-world testing and sensitivity analysis. It offers valuable insights for enterprises and governments to effectively integrate EVs into their delivery systems and reduce carbon emissions.
The paper’s contributions include presenting a sustainable cold chain distribution approach, introducing an improved VNS algorithm, and providing valuable management insights based on extensive real-world testing. The subsequent sections of the paper review related literature, outline problem features and the proposed model, introduce the VNS algorithm with its new mechanisms, describe the experimental design, validate the algorithm’s performance, present numerical results, and conclude the study.
Tom Van Woensel 