A multi-commodity two-echelon vehicle routing problem with split-delivery and time windows for decisions of delivery-terminal systems

Purpose This study aims to explore the collaborative information system supporting decisions between delivery transport (commodities from DCs to satellites) and terminal transport (commodities from satellites to customers) of e-commerce enterprises, a novel variant of 2E-VRP named MC-2E-VRP-SDTW is investigated by jointly considering three important practical factors: multi-commodity, split delivery and time windows. Design/methodology/approach A mixed-integer linear programming (MILP) model is formulated, based on which three families of valid inequalities are proposed to enhance its solution performance. A two-stage heuristic is constructed to quickly obtain an initial solution, which is improved by incorporating three exchange operators and four/two destruction/repair operators into an enhanced adaptive large neighborhood search (ALNS) framework. Findings Experimental results show that the proposed valid inequalities can help obtain optimal solutions for more instances and improve the lower bounds by 28.7% on average. Our ALNS algorithm can obtain optimal solutions or near-optimal solutions for small-sized instances with far less computational time than CPLEX, and it is superior to the comparing algorithm for large-sized instances with an average improvement of solution at 4.8%. Practical implications This research provides integrated transportation decision support for practical operations of relevant enterprises, thereby helping to reduce operational costs and improve logistics efficiency. Originality/value This study proposes a new model for formulating MC-2E-VRP-SDTW and designs a tailored high-efficiency algorithm to solve the investigated problem.