Enabling Ultrafast Online Order Fulfillment: Efficient Inventory Management for In-Store Microfulfillment Centers

The emergence of in-store microfulfillment centers (ISMFCs) is transforming omnichannel retailing by facilitating the rapid fulfillment of orders placed online. Effective management, particularly of the complex decisions related to the dynamic selection of products to place in the ISMFC and the determination of inventory levels of each product selected therein, can go a long way in maximizing the benefits of ISMFCs. In this paper, we first formulate the ISMFC inventory decision problem as a Markov Decision Process. We then leverage intuition from this representation and introduce a threshold policy based on the optimal multiperiod marginal profit-to-volume ratio to efficiently manage stochastic demand and make forward-looking decisions. We establish the quality of the proposed approach using two sets of computational experiments. Because key benchmark approaches do not scale well, we restrict the first set of experiments to simulated data involving three products. In the second set of experiments—based on a retail data set with 3,498 products—we benchmark the threshold policy against scalable methods, employing model parameters obtained partly from data estimation and partly from observed data values. The results from these experiments demonstrate that our approach outperforms state-of-the-art benchmarks, identifying near-optimal solutions in a few seconds. The scalability and effectiveness of the threshold policy underscores its practical viability and highlights the substantial economic gains achievable in managing ISMFC operations. History: Accepted by Ram Ramesh, Area Editor for Data Science and Machine Learning. Funding: Q. Jia was supported by the National Natural Science Foundation of China [Grants 72394373 and 72231004]. Supplemental Material: The software that supports the findings of this study is available within the paper and its Supplemental Information ( https://pubsonline.informs.org/doi/suppl/10.1287/ijoc.2024.0710 ) as well as from the IJOC GitHub software repository ( https://github.com/INFORMSJoC/2024.0710 ). The complete IJOC Software and Data Repository is available at https://informsjoc.github.io/ .