A bi-level programming model for optimizing location, capacity, and pricing decisions of origin warehouses in an agricultural supply chain

As new infrastructures for the "first mile" in an agricultural supply chain, origin warehouses are highly valued and supported for reducing the loss of agricultural products. With this regard, we develop a bi-level programming model for optimizing the construction and operation of agricultural product origin warehouses. Specifically, the upper-level supply chain enterprise makes decisions regarding the location, capacity, and pricing of the origin warehouses, while the lowerlevel planters focus on determining the sales allocation plan of agricultural products to various origin warehouses to maximize their profits. To deal with the computational complexity of proposed model, we first divide the original problem into three subproblems related to the location, operation of origin warehouse and sales allocation plan of planters. Subsequently, we develop a hybrid adaptive large neighborhood search algorithm combined with the heuristic rulebased mechanism and the gurobi solver. Taking the Shilin area of Yunnan, China as a case study, we demonstrate the effectiveness of the proposed model and algorithm. The numerical results indicate that the construction of origin warehouses follows a "1 + N" layout strategy, featuring one larger core warehouse and N smaller warehouses, thereby balancing both logistics efficiency and cost. In addition, these origin warehouses attract higher quality agricultural products by adopting tiered pricing and premium-for-quality strategies, which benefit both the agricultural supply chain company and the planters.