EUVIMEDEUVIMED entdecken
Crossref

Energy-Efficient Computing Using Adaptive Model Selection for Mobile Agents in Federated Learning

Benteng Zhang, Yingchi Mao, Tianfu Pang, Zhenxiang Pan, Jianxin Huang, Feng Mao, Xiaoming He, Jie Wu

IEEE Transactions on Cognitive Communications and Networking · 2026 · 12 · 8914-8928

DOI: 10.1109/tccn.2026.3704580 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IEEE Transactions on Cognitive Communications and Networking, 2026. Verfasst von Benteng Zhang und 7 weiteren Autor:innen. Die Arbeit nennt 46 Literaturangaben. Gefördert von: National Natural Science Foundation of China, Fundamental Research Funds for the Central Universities, 16th Batch of Science and Technology Development Programs (Innovation Consortium Projects) of Suzhou.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Multi-agent federated learning requires all agents to stay online until the global model converges, and its multi-round training incurs substantial on-agent computing. However, most edge mobile agents with limited computational capacity and battery life may drop out of FL training prematurely due to low battery, which can degrade the convergence speed and accuracy of the global model. Prior works assign heterogeneous models to different agents to reduce on-agent computational energy, but neglect the changes in data distributions and computational capacities on agents. This makes it difficult to save on-agent computational energy and to support global aggregation. To this end, we propose a hierarchical Federated learning framework based on Adaptive Local Training (FedALT). Specifically, FedALT can adaptively select the most suitable local model for each agent based on its computational capacity and data distribution to reduce on-agent computational energy. To achieve finer-grained savings in computational energy, FedALT dynamically adjusts the number of local epochs for each agent during local training. Furthermore, FedALT conducts edge aggregation through model layer sharing and knowledge distillation to improve the convergence speed and accuracy of the global model. We compare FedALT with three baselines by training six models on two Non-IID datasets. Experiments on both the simulation platform and the hardware platform demonstrate that FedALT can achieve the lowest computational energy, the highest average global model accuracy, and the fastest global model convergence.

Förderung

  • National Natural Science Foundation of China
  • Fundamental Research Funds for the Central Universities
  • 16th Batch of Science and Technology Development Programs (Innovation Consortium Projects) of Suzhou
  • Technology Talent and Platform Program of Yunnan Province
  • Technology Project of Huaneng Group
  • High Performance Computing Platform

Zitieren

APA

Zhang, B., Mao, Y., Pang, T., Pan, Z., Huang, J., Mao, F., He, X., & Wu, J. (2026). Energy-Efficient Computing Using Adaptive Model Selection for Mobile Agents in Federated Learning. IEEE Transactions on Cognitive Communications and Networking, 12, 8914-8928. https://doi.org/10.1109/tccn.2026.3704580

Vancouver

Zhang B, Mao Y, Pang T, Pan Z, Huang J, Mao F, et al. Energy-Efficient Computing Using Adaptive Model Selection for Mobile Agents in Federated Learning. IEEE Transactions on Cognitive Communications and Networking. 2026;12:8914-8928. doi:10.1109/tccn.2026.3704580

BibTeX
@article{Zhang2026,
  author = {Zhang, Benteng and Mao, Yingchi and Pang, Tianfu and Pan, Zhenxiang and Huang, Jianxin and Mao, Feng and He, Xiaoming and Wu, Jie},
  title = {Energy-Efficient Computing Using Adaptive Model Selection for Mobile Agents in Federated Learning},
  journal = {IEEE Transactions on Cognitive Communications and Networking},
  year = {2026},
  volume = {12},
  pages = {8914--8928},
  doi = {10.1109/tccn.2026.3704580}
}
Kostenlos starten →