Joint Trajectory, RIS, and Computation Offloading Optimization via Decentralized Model-Based PPO in Urban Multi-UAV Mobile Edge Computing
IEEE Transactions on Mobile Computing · 2026-09 · 25 (9) · 14498-14512
DOI: 10.1109/tmc.2026.3679344 ↗Kurz gefasst
Fachartikel in der Zeitschrift IEEE Transactions on Mobile Computing, 2026-09. Verfasst von Liangshun Wu und 2 weiteren Autor:innen. Die Arbeit nennt 41 Literaturangaben. Gefördert von: Xi’an Key Laboratory of Advanced Control and Intelligent Process, Key R&D Plan of Shaanxi Province, Collaborative Innovation Project of Xi’an Science and Technology Bureau.
Abstract (Originalsprache)
Computation offloading in urban multi-UAV MEC networks is challenging because building blockage frequently disrupts line-of-sight links and user demand changes rapidly. RIS can create controllable reflected paths, but effective deployment requires jointly optimizing UAV trajectories, offloading schedules, and RIS phase shifts under limited inter-UAV information exchange. To address this problem, we propose a decentralized model-based MARL framework in which each UAV makes mobility and offloading decisions from local and k-hop observations and submits an RIS phase proposal to a lightweight RIS controller. To improve sample efficiency and training stability, each agent learns a local dynamics model and uses short-horizon branched rollouts in PPO updates. Simulations show that the proposed method approaches centralized performance while improving throughput and energy efficiency in large-scale urban scenarios.
Förderung
- Xi’an Key Laboratory of Advanced Control and Intelligent Process
- Key R&D Plan of Shaanxi Province
- Collaborative Innovation Project of Xi’an Science and Technology Bureau
- Open Research Fund of National Mobile Communications Research Laboratory
- Southeast University
- National Natural Science Foundation of China
- Open Fund of Guizhou Key Laboratory of Artificial Intelligence and Brain-inspired Computing
- Key Laboratory of Embedded System and Service Computing
Zitieren
APA
Wu, L., Du, J., & Qu, J. (2026). Joint Trajectory, RIS, and Computation Offloading Optimization via Decentralized Model-Based PPO in Urban Multi-UAV Mobile Edge Computing. IEEE Transactions on Mobile Computing, 25(9), 14498-14512. https://doi.org/10.1109/tmc.2026.3679344
Vancouver
Wu L, Du J, Qu J. Joint Trajectory, RIS, and Computation Offloading Optimization via Decentralized Model-Based PPO in Urban Multi-UAV Mobile Edge Computing. IEEE Transactions on Mobile Computing. 2026;25(9):14498-14512. doi:10.1109/tmc.2026.3679344
BibTeX
@article{Wu2026,
author = {Wu, Liangshun and Du, Jianbo and Qu, Junsuo},
title = {Joint Trajectory, RIS, and Computation Offloading Optimization via Decentralized Model-Based PPO in Urban Multi-UAV Mobile Edge Computing},
journal = {IEEE Transactions on Mobile Computing},
year = {2026},
volume = {25},
number = {9},
pages = {14498--14512},
doi = {10.1109/tmc.2026.3679344}
}
EUVIMED