An online scheduling framework for UAV emergency logistics in dynamic disaster environments

发表于 2026-9-23 11:13:35

DOI号:[size=1em]10.1016/j.comnet.2026.112607

文献链接:PADA: An online scheduling framework for UAV emergency logistics in dynamic disaster environments - ScienceDirect

文献作者:Ziwen Bao, Kaiwei Mo, Xianglong Li, Zongpeng Li, Fansheng Gao,

备注信息:Abstract: Unmanned Aerial Vehicles (UAVs) provide a flexible means of delivering emergency drugs in mountainous disaster regions, where online scheduling decisions must simultaneously handle dynamic no-fly zones, hard delivery deadlines, and payload-dependent endurance. Existing UAV scheduling methods often rely on offline training or static planning, and they rarely model life-value urgency as a primary scheduling objective. These limitations make them less suitable for no-prior-knowledge emergency response. This paper presents PADA, a pre-training-free online scheduling framework that combines Priority-driven Assignment (PA) with Dynamic Adaptation (DA). PA accepts and matches tasks according to life-value urgency and priority-discounted energy cost, whereas DA performs bounded real-time replanning, adaptive payload shedding, and fail-safe return-to-base fallback when sudden obstacles or residual-energy risks arise. To better reflect UAV flight physics, the commonly used linear energy approximation is replaced with a payload-aware nonlinear rotary-wing propulsion model. A primal–dual analysis shows that the online assignment rule achieves a logarithmic competitive factor β(γ) at the assignment layer, while the additional welfare degradation caused by execution-time adaptation is explicitly characterized. Extensive real-data-driven experiments over 30 independent runs show that PADA improves total social welfare by 14.26% over the best baseline, raises the Type-1 emergency delivery rate from 71.69% to 95.54%, and reduces the value-loss ratio by 31.40%. PADA also sustains millisecond-level assignment latency and a 95th-percentile dynamic-adaptation latency below 50 ms, demonstrating its practicality for online emergency UAV logistics.
Keywords: UAV scheduling; Online optimization; Emergency logistics; Dynamic adaptation; Path planning

发表于 2026-9-23 11:13:36

: 当前求助已完成，请重新发起求助。
已完结状态的附件将会在24小时内删除。

[Elsevier/Sciencedirect] An online scheduling framework for UAV emergency logistics in dynamic disaster environments

已采纳

同分类求助

最新回复

站点公告

今日热心研友

（热心研友QQ群）

[Elsevier/Sciencedirect] An online scheduling framework for UAV emergency logistics in dynamic disaster environments 一键复制