A variable neighborhood search approach for the multi-compartment vehicle routing problem with time windows considering carbon emission

Abstract

In recent years, fresh food e-commerce has developed rapidly and become a new strategic field for traditional e-commerce firms. Logistics is one of the bottlenecks in its development. Each type of fresh food has its own temperature requirements during transportation. In this paper, the multi-compartment vehicle routing problem with time window (MCVRPTW) arising in fresh food e-commerce is proposed to include the consideration of carbon emission. A variable neighborhood search (VNS) approach with the key steps of local search and shaking is proposed to solve the MCVRPTW formulation. For the VNS approach, three operators are used in the local search phase, and two kinds of operators based on path relinking are adopted in the shaking phase. Extensive efforts are made to verify the proposed approach. Firstly, the parameters are tuned based on Taguchi method. Secondly, the approach is compared with representative algorithms in the literature to solve VRPTW instances and MCVRPTW instances. For the VRPTW cases, VNS outperforms a well-known benchmark algorithm by 4–7%, and the overall average gap between the VNS solutions and the known best solutions is only 1.93%. For the MCVRPTW instances, VNS produces the same or improved “best solution” in the majority of the 24 instances with less than 30% of running time, compared with two existing MCVRPTW algorithms. Finally, VNS is employed to solve a real-world case, and the results show that the total cost of the solution is 36.38% lower than the actual operation. Overall, this study takes the minimum total of travel cost, fixed cost, refrigeration cost and carbon emission cost as the optimization objective, and the findings are conducive to reducing energy consumption and carbon emission in the fresh e-commerce delivery process.