Digital Twin-Driven MADRL Approaches for Communication-Computing-Control Co-Optimization
IEEE Journal on Selected Areas in Communications · 2025-10 · 43 (10) · 3596-3611
DOI: 10.1109/jsac.2025.3574616 ↗Kurz gefasst
Fachartikel in der Zeitschrift IEEE Journal on Selected Areas in Communications, 2025-10. Verfasst von Xiaoming Yuan und 6 weiteren Autor:innen. Die Arbeit wurde bisher 12-mal zitiert (Crossref) und nennt 38 Literaturangaben. Gefördert von: National Natural Science Foundation of China, Research Grants Council of Hong Kong Special Administrative Region, China, NSFC/RGC CRS.
Abstract (Originalsprache)
The unpredictability of network environments, limited edge resources, and the high complexity of collaborative policies are significantly hindering the development of the Industrial Internet of Things (IIoT). These challenges are particularly pronounced in healthcare, where high-priority, delay-sensitive medical tasks and large-scale personalized services face substantial obstacles. To address these challenges, this paper proposes the Self-Attention Enhanced QMIX with Multi-Pass Multi-Task Execution (SAE-MT-QMIX) algorithm, aimed at optimizing communication and computing resource allocation as well as task offloading strategies. By leveraging Digital Twin (DT) support, the algorithm achieves collaborative optimization of communication, computing, and control within the Internet of Medical Things (IoMT), significantly enhancing the quality of service for massive personalized applications. The algorithm adopts a distributed execution and centralized training framework: the distributed execution component uses the Multi-Pass Multi-Task Deep Q-Network (MPMT-DQN) algorithm to handle the complexity of parameterized action spaces in multi-task scenarios, while the centralized training component employs the Self-Attention Enhanced QMIX (SAE-QMIX) algorithm to dynamically optimize credit assignment across multiple users. Simulation results demonstrate that SAE-MT-QMIX significantly reduces delay and energy consumption compared to baseline methods. It ensures effective optimization of communication, computing, and control in dynamic IoMT, efficiently addressing diverse demands and tasks while enhancing service quality and system adaptability.
Förderung
- National Natural Science Foundation of China
- Research Grants Council of Hong Kong Special Administrative Region, China
- NSFC/RGC CRS
- Areas of Excellence Scheme
- Collaborative Research Fund
- Shenzhen–Hong Kong-Macau Technology Research Programme
Zitieren
APA
Yuan, X., Tian, H., Zhang, X., Du, H., Zhang, N., Huang, K., & Cai, L. (2025). Digital Twin-Driven MADRL Approaches for Communication-Computing-Control Co-Optimization. IEEE Journal on Selected Areas in Communications, 43(10), 3596-3611. https://doi.org/10.1109/jsac.2025.3574616
Vancouver
Yuan X, Tian H, Zhang X, Du H, Zhang N, Huang K, et al. Digital Twin-Driven MADRL Approaches for Communication-Computing-Control Co-Optimization. IEEE Journal on Selected Areas in Communications. 2025;43(10):3596-3611. doi:10.1109/jsac.2025.3574616
BibTeX
@article{Yuan2025,
author = {Yuan, Xiaoming and Tian, Hansen and Zhang, Xinling and Du, Hongyang and Zhang, Ning and Huang, Kaibin and Cai, Lin},
title = {Digital Twin-Driven MADRL Approaches for Communication-Computing-Control Co-Optimization},
journal = {IEEE Journal on Selected Areas in Communications},
year = {2025},
volume = {43},
number = {10},
pages = {3596--3611},
doi = {10.1109/jsac.2025.3574616}
}
EUVIMED