A Survey on Device Scheduling in Over-the-Air Computation-Enabled Federated Learning
Usman Iqbal, Haejoon Jung, Hyundong Shin · 2024
In this paper, we provide a survey on device scheduling strategies for over-the-air computation (AirComp)-enabled federated learning (FL). Due to the emergence of 6G networks, the combination of FL and AirComp offers significant benefits, especially in terms of spectral and energy efficiency. However, effective device scheduling is critical to maximize these benefits, ensuring precise and timely data aggregation. We mention various scheduling algorithms related to device heterogeneity, communication constraints, and network conditions. This survey highlights the impact of optimized scheduling on the performance and scalability of FL systems as well as aims to guide future research and development in enhancing federated learning with AirComp.