ORBITER: Conflict-Aware Decision-Making for Agentic Last-Mile Delivery

ORBITER is a new framework for next‑order decision‑making in last‑mile delivery that uses large language models (LLMs) to reason about courier spatiotemporal states and visible orders. It structures decision points with local trade‑offs, ranks candidates via fixed proposers, and employs a critic to verify decisions against gathered evidence. Experiments on data from four cities show ORBITER outperforms state‑of‑the‑art baselines by up to 9.2% on average.

Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv AI.

• llms
• agents
• benchmarks