2025 59th Annual Conference on Information Sciences and Systems (CISS) · Conference paper · 2025
Authors: Jiahao You, Ziye Jia, Chao Dong, Qihui Wu, Zhu Han
Published in: 2025 59th Annual Conference on Information Sciences and Systems (CISS), Baltimore, MD, USA, pp. 1–6.
Authorship: First author / 第一作者
DOI: 10.1109/CISS64860.2025.10944752
IEEE Xplore ↗ | Open preprint on arXiv ↗
Research overview
This paper combines UAV mobility and ground-station computing to support uncertain computational tasks generated by unmanned surface vehicles.
The problem
Task generation and surface-vehicle trajectories vary over time. Heterogeneous resources and limited UAV capacity complicate joint offloading and flight planning.
The approach
The problem jointly optimizes task offloading and UAV trajectories to minimize execution time. Generative AI-enhanced heterogeneous-agent proximal policy optimization improves actor representations and stabilizes critic learning in dynamic environments.
Evaluation
Simulations compare the proposed GAI-HAPPO algorithm with benchmark methods under the considered network settings. Model details and experimental configurations are available in the conference paper.
生成式人工智能增强的无人机与地面站协同边缘计算：面向无人水面艇
本文结合无人机的机动性与地面站的计算能力，为无人水面艇产生的不确定计算任务提供服务。
研究问题
任务生成和无人水面艇轨迹随时间变化。节点资源异构以及无人机容量受限，增加了联合卸载与航迹规划的难度。
研究方法
论文以最小化执行时间为目标联合优化任务卸载与无人机轨迹，采用生成式人工智能增强的异构智能体近端策略优化，改进策略表征并提高动态环境中的评论家学习稳定性。
研究评估
论文在所设网络条件下，通过仿真将 GAI-HAPPO 与基线方法进行比较。模型细节与实验配置见会议原文。
中文题名为便于阅读所作的译名，正式引用请使用英文原题。
Citation
[1] J. You, Z. Jia, C. Dong, Q. Wu, and Z. Han, “Generative AI-enhanced cooperative MEC of UAVs and ground stations for unmanned surface vehicles,” in Proc. 59th Annu. Conf. Inf. Sci. Syst. (CISS), Baltimore, MD, USA, 2025, pp. 1–6, doi: 10.1109/CISS64860.2025.10944752.
