Scalable Multistage Stochastic Optimization for Freight Procurement in Transportation-Inventory Systems

The procurement of freight services is an important element for the supply chain management of a shipper (i.e., a manufacturer or retailer) that sources transportation services from the third-party logistics market. Motivated by a practical freight procurement problem faced by shippers, we provide a holistic approach to designing freight procurement strategies for transportation-inventory systems that captures the interconnections among freight procurement, transportation, and inventory management. In view of supply and demand uncertainties, we consider the problem on the basis of a multistage decision process with uncertain data. To handle instances of realistic size, we propose an enhanced stochastic dual dynamic programming solution approach. We conduct extensive numerical experiments to test the performance of the approach. The results demonstrate that our approach scales to very large instances with up to ${50}^{18}$ scenarios and that the proposed enhancement strategies significantly improve its performance. Compared with methods commonly adopted to solve similar problems, our approach achieves significant cost improvements based on generated instances calibrated from established benchmark instances and stochastic simulations.

History: Accepted by Andrea Lodi, Area Editor for Design & Analysis of Algorithms–Discrete.

Funding: This work was supported by the National Natural Science Foundation of China [Grants 72301230, 72288101, and 72322022]; the Research Grants Council of the Hong Kong Special Administrative Region, China [Grant 25223223]; and the Fundamental and Interdisciplinary Disciplines Breakthrough Plan of the Ministry of Education of China [Grant JYB2025XDXM211].

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.2025.1141) as well as from the IJOC GitHub software repository (https://github.com/INFORMSJoC/2025.1141). The complete IJOC Software and Data Repository is available at https://informsjoc.github.io/.