Distributionally Robust Operations of Networked Distribution Centers at Scale

48 Pages Posted: 11 Oct 2022 Last revised: 7 May 2024

See all articles by Chenxi Li

Chenxi Li

Beijing Institute of Technology - School of Management & Economics

Sheng Liu

Rotman School of Management

Wei Qi

Tsinghua University - Department of Industrial Engineering; McGill University - Desautels Faculty of Management

Lun Ran

Beijing Institute of Technology

Aiqi Zhang

Wilfrid Laurier University - School of Business & Economics

Date Written: August 31, 2022

Abstract

Problem definition: How should retailers and sellers distribute inventory across a large network of (potentially hundreds of) distribution centers? Motivated by the emerging challenges of inventory allocation facing volatile demand, we study distributionally robust network inventory management with cross-fulfillment using a scenario-based approach.
Methodology/results: We consider a scenario-wise ambiguity set that is adaptable to feature information and demand correlation. By exploiting the supermodularity property of the multilocation newsvendor problem, we obtain a scalable linear programming formulation for inventory placement optimization. We extend this framework to account for inventory dynamics in a multiperiod setting and derive a tractable policy. We conduct a case study using data from a major logistics service provider in China and demonstrate the superior performance of the proposed approach. Analytically, we characterize the optimal inventory decision for a symmetric two-location problem under single-scenario and two-scenario ambiguity sets. Our analysis shows how the optimal safety stock level varies according to the correlation and network structure.
Managerial implications: Our framework yields tractable and robust solutions for large distribution networks. Based on out-of-sample evaluation, our approach outperforms alternative data-driven methods, especially when the lead time is positive. It reduces the total inventory and fulfillment cost by up to 19% over the current policy used by the company. The scenario-wise ambiguity set provides a promising modeling option for delineating useful distributional information on the demands of multiple locations.

Keywords: Inventroy Network, Distributionally Robust Optimization, Supermodularity

Suggested Citation

Li, Chenxi and Liu, Sheng and Qi, Wei and Ran, Lun and Zhang, Aiqi, Distributionally Robust Operations of Networked Distribution Centers at Scale (August 31, 2022). Available at SSRN: https://ssrn.com/abstract=4207042 or http://dx.doi.org/10.2139/ssrn.4207042

Chenxi Li

Beijing Institute of Technology - School of Management & Economics ( email )

Beijing, 100081
China

Sheng Liu

Rotman School of Management ( email )

105 St. George st
Toronto, ON M5S 3E6
Canada

Wei Qi

Tsinghua University - Department of Industrial Engineering ( email )

Beijing
China

McGill University - Desautels Faculty of Management ( email )

1001 Sherbrooke Street West
Montreal, Quebec H3A 1G5
Canada

Lun Ran

Beijing Institute of Technology ( email )

Beijing, 100081
China

Aiqi Zhang (Contact Author)

Wilfrid Laurier University - School of Business & Economics ( email )

Waterloo, Ontario N2L 3C5
Canada

Do you have a job opening that you would like to promote on SSRN?

Paper statistics

Downloads
563
Abstract Views
1,573
Rank
92,825
PlumX Metrics