Reproduce “SabreAgent: Language Models at Design Time for Lost-Sales Inventory Control” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Yang Liu, Yulin Huang, Xue Yu, Jiong Dong, Jianshen Zhang, Yongzhi Qi
SabreAgent uses a language model at design time to prepare a product-specific seasonal cue and select a policy family. Frozen artifacts feed statistical forecasting and operations-research control, which place orders without runtime language-model calls. Across 1,320 InventoryBench instances, it averages normalized reward 0.6311 versus 0.5380 for the strongest baseline and ranks first in all six cells; positive-lead-time scores are not fractions of attainable performance.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.