活動

We consider an online retailer selling multiple products to multiple zones over an infinite horizon. The retailer replenishes the products from a single supplier and stores them at multiple warehouses. At the start of each period in the selling horizon, the retailer determines the order quantities of the products and their storage quantities to each warehouse subject to its capacity constraint. At the end of the period, after knowing the demands, the retailer determines the retrieval quantities from each warehouse to fulfill the demands. The retailer’s objective is to maximize her expected total discounted profit over the selling horizon. We prove that solving the infinite-horizon problem reduces to solving a single-period problem: For the case with a single zone, we solve the problem optimally and characterize the optimal ordering, storage, and retrieval policies. For the case with multiple zones, the problem is intractable analytically and we propose three efficient heuristics to solve it. The first heuristic controls the over-stocking risk through virtual demand pooling. The second heuristic is based on virtual capacity allocation. The third heuristic is a hybrid of the first two heuristics. Numerical experiments suggest that our best heuristic outperforms the sample average approximation (SAA) method, and achieves at least a 93% efficiency relative to an upper bound. A case study based on data from a major fashion online retailer in Asia suggests that all the heuristics outperform the SAA method, generating at least a 15% profit improvement over the retailer’s current practice.