IEEE Transactions on Cognitive Communications and Networking · Journal article · 2026
Authors: Jiahao You, Ziye Jia, Can Cui, Chao Dong, Qihui Wu, Zhu Han
Published in: IEEE Transactions on Cognitive Communications and Networking, vol. 12, pp. 5161–5175.
Authorship: First author / 第一作者
DOI: 10.1109/TCCN.2025.3641588
Research overview
This paper studies coordinated trajectory planning and task offloading for low-altitude edge computing, combining auction-based assignment with diffusion-based multi-agent learning.
The problem
UAVs must serve changing task demands with limited energy and heterogeneous computing resources. Movement and offloading decisions affect one another over time, making them difficult to optimize independently.
The approach
An integrated air-ground network is optimized at two timescales. A Vickrey–Clarke–Groves auction assigns trajectories at the larger timescale. At the smaller timescale, diffusion-based heterogeneous-agent proximal policy optimization embeds conditional denoising in actor networks to generate offloading actions.
Evaluation
Simulations compare energy efficiency, task success rate, and learning convergence against baseline methods. The paper provides the network assumptions and experimental settings needed to interpret these results.
拍卖与扩散式多智能体强化学习驱动的低空智联网分层任务卸载和轨迹优化
这项工作研究低空边缘计算中的轨迹规划与任务卸载，将拍卖分配机制与扩散式多智能体学习结合。
研究问题
无人机需要在能源受限、计算资源异构的条件下服务动态任务。飞行轨迹与卸载决策相互影响，并在时间上耦合，难以分别优化。
研究方法
论文构建空地协同网络，并采用双时间尺度优化：大时间尺度通过 Vickrey–Clarke–Groves 拍卖分配轨迹，小时间尺度通过扩散式异构智能体近端策略优化，在策略网络中利用条件去噪生成卸载动作。
研究评估
论文通过仿真比较能量效率、任务成功率与学习收敛表现。具体网络假设、对比方法和实验设置见发表原文。
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. Cui, C. Dong, Q. Wu, and Z. Han, “Hierarchical task offloading and trajectory optimization in low-altitude intelligent networks via auction and diffusion-based MARL,” IEEE Trans. Cogn. Commun. Netw., vol. 12, pp. 5161–5175, 2026, doi: 10.1109/TCCN.2025.3641588.
