analiza aiqbench
This paper introduces ORBITER, an agentic framework that leverages Large Language Models to enhance decision-making in last-mile delivery by explicitly reasoning over spatial and temporal constraints. The system employs a structured verification loop where an LLM gathers evidence on candidate orders while an independent critic validates decisions against this data, addressing the reliability issues often seen with direct LLM predictions. Evaluated across four cities, the approach demonstrates significant performance gains over traditional learning-based baselines by making the reasoning process transparent and robust.
ORBITER: Conflict-Aware Decision-Making for Agentic Last-Mile Delivery
arXiv:2608.18846v1 Announce Type: new Abstract: Last-mile delivery aims to handle dynamically arriving orders with couriers while modeling complex spatial and temporal correlations. Recent learning-based methods model spatiotemporal dependencies among orders to predict courier service sequences, but leave next-order decision making unexplained. Describing the current delivery state in language allows LLMs to reason explicitly about the spatial, temporal, and behavioral cues behind an individual decision. As direct predictors, however, LLMs remain sensitive to task presentation and often produce unreliable decisions. To address these challenges, we introduce ORBITER, an agentic Order Arbiter for next-order decision-making in last-mile delivery. ORBITER models courier service through decision points, each containing the courier's spatiotemporal state and visible orders and exposing local trade-offs for modeling and verification. Fixed proposers rank the candidates, and a structured report identifies where their rankings disagree. The LLM uses task-specific tools to gather evidence on the leading alternatives, while an independent critic checks the resulting decision against that evidence. We conduct extensive evaluations on data in four cities, where ORBITER outperforms existing state-of-the-art baselines by up to 9.2% on average showing its effectiveness.
Abstract:Last-mile delivery aims to handle dynamically arriving orders with couriers while modeling complex spatial and temporal correlations. Recent learning-based methods model spatiotemporal dependencies among orders to predict courier service sequences, but leave next-order decision making unexplained. Describing the current delivery state in language allows LLMs to reason explicitly about the spatial, temporal, and behavioral cues behind an individual decision. As direct predictors, however, LLMs remain sensitive to task presentation and often produce unreliable decisions. To address these challenges, we introduce ORBITER, an agentic Order Arbiter for next-order decision-making in last-mile delivery. ORBITER models courier service through decision points, each containing the courier's spatiotemporal state and visible orders and exposing local trade-offs for modeling and verification. Fixed proposers rank the candidates, and a structured report identifies where their rankings disagree. The LLM uses task-specific tools to gather evidence on the leading alternatives, while an independent critic checks the resulting decision against that evidence. We conduct extensive evaluations on data in four cities, where ORBITER outperforms existing state-of-the-art baselines by up to 9.2% on average showing its effectiveness.
Artificial Intelligence (cs.AI)
arXiv:2608.18846 [cs.AI]
(or arXiv:2608.18846v1 [cs.AI] for this version)
https://doi.org/10.48550/arXiv.2608.18846 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 270 cuvinte
Subiecte conexe
Problems of Proliferation
LessWrong · 29 sept. 2026
The current bottleneck is political will, not research
AI Alignment Forum · 11 iul. 2026
What We're Up Against: An AI Safety Crash Course
LessWrong · 24 sept. 2026
Inside the suddenly explosive world of AI safety
The Verge · 17 sept. 2026