A Data-driven Distributionally Robust Newsvendor Problem with Information Learning for Demand Censoring
Date:
From July 11 to 14, 2025, I had the incredible opportunity to attend and present at the 2025 POMS International Conference in Danzhou, Hainan Province. The conference was a massive academic feast, bringing together over 1,400 experts and scholars from around the world to explore the theme “Digital and Intelligent Operations Empowering Management Innovation and Transformation.”
My talk, titled “A Data-driven Distributionally Robust Newsvendor Problem with Information Learning for Demand Censoring,” aligned perfectly with the conference’s focus on data-driven management innovation. I addressed a critical challenge in supply chain management: how to make accurate inventory decisions when historical demand data is censored due to stock-outs. During the presentation, I introduced our novel data-driven repair framework based on Distributionally Robust Optimization (DRO). I demonstrated how leveraging local data (DRN model) and multisource external information synergy (DRNS model) can effectively mitigate the negative effects of demand censoring on Sample Average Approximation (SAA) decisions.
The scale and depth of the event were truly inspiring. With 24 parallel sessions and 192 sub-forums focusing on various research hotspots, the intellectual collisions and academic exchanges were invaluable. Discussing our real-world validated models with top scholars, and absorbing insights from the keynote speeches, gave me fresh perspectives on my own research in proactive fulfillment and supply chain resilience.
Beyond the rigorous academic exchanges, experiencing the beautiful tropical setting of Hainan made this conference an unforgettable milestone in my Ph.D. journey. I look forward to continuing to push the boundaries of operations research and applying these insights to my future work!
