文献类型：期刊文献

英文题名：Energy State Awareness-Oriented Federated Learning Collaborative Optimization and Scheduling Algorithm

作者：Xiao, Kaile[1]; Lao, Fengxue[1]; Ma, Yan[1]; Liu, Shiyuan[1]; Tong, Xiaolu[1]; Chen, Jingxia[1]

第一作者：Xiao, Kaile

机构：[1] School of Applied Science and Technology, Beijing Union University, Beijing, China

第一机构：北京联合大学应用科技学院

年份：2025

起止页码：36-41

外文期刊名：5th International Conference on Electron Devices and Applications, ICEDA 2025

收录：EI(收录号：20261520499928)

语种：英文

外文关键词：Artificial intelligence - Behavioral research - Collaborative learning - Computer aided instruction - Decision making - Distributed computer systems - Energy supply - Learning algorithms - Learning systems - Optimization - Power management - Scheduling algorithms

摘要：With the widespread application of energy-harvesting IoT devices, distributed machine learning in this scenario faces challenges such as device offline and reduced training reliability caused by the randomness of energy supply. Traditional federated learning cannot adapt to the dynamic characteristics of energy-harvesting scenarios due to its assumption of stable device energy supply. This paper proposes an Energy State-Aware Federated learning collaborative optimization and scheduling algorithm (ESAFed). By defining the collaborative design requirements of the battery management system and energy-harvesting hardware, a lightweight intelligent scheduling mechanism is constructed to realize adaptive participation decision-making of devices. The algorithm takes the real-time energy status and prediction information of devices as the core decision-making basis, collaboratively optimizes the global learning efficiency and system lifetime, and provides a solution for the deployment of distributed intelligence in self-powered IoT environments. Experiments show that the algorithm proposed in this paper performs better than other algorithms, proving the effectiveness of the proposed algorithm. ? 2025 IEEE.