Platform-enabled two-echelon collaborative logistics for cross-supply chain optimization in instant retail

The rapid expansion of instant retail poses significant challenges to logistics systems, as time-sensitive last-mile delivery and high-frequency supplier replenishment drive costs upward. Motivated by practical needs, a platform-enabled, two-echelon multi-owner collaborative logistics model is proposed from a cross-supply chain perspective to support alliance formation, routing optimization, and cost allocation. Firstly, participants form short-term, order-driven logistics alliances via the platform efficiently. Secondly, A two-echelon collaborative vehicle routing problem with time windows across supply chain (2E-MOCVRPTW-CSC) is formulated to minimize total logistics cost, while satisfying the stringent timeliness requirements of instant retail. The bi-level programming model is established to reflect the strategic interactions, with suppliers as leaders and retailers as followers to optimize routes in delivery and replenishment. Thirdly, the Shapley value is quickly calculated to allocate the cost. To solve this problem, a two-stage iterative adaptive large neighborhood search algorithm (TSIALNS) is designed. Extensive experiments, real-road network cases and sensitivity analysis confirm the model’s effectiveness. Compared with non-collaborative and decoupled collaborative models, the proposed approach achieves up to 31.65% cost reduction in real-world case, with retailers and suppliers saving 40% and 14% on average. This demonstrates substantial economic potential of cross-supply chain collaboration and practical value for resilient multi-level logistics systems.