JUROR: Joint UAV Flight and Opportunistic Routing for Delay-Tolerant Networks

A new research paper on arXiv (2608.04590) proposes JUROR, a reinforcement learning framework that jointly optimizes UAV flight and opportunistic routing in delay-tolerant networks (DTNs). The study addresses challenges in sparse connectivity, intermittent contacts, finite buffers, and limited message time-to-live (TTL), which often cause sparse delivery and congestion. JUROR employs centralized training and decentralized execution (CTDE) based on proximal policy optimization (PPO) to enable per-node replication under contact-limited observations. The approach aims to enlarge future contacts through discrete UAV headings, improving end-to-end performance. The paper is authored by researchers (names not provided) and was announced as a new submission. The work contributes to the field of networking and AI, specifically applying reinforcement learning to network optimization.