EExApp: GNN-Based Reinforcement Learning for Radio Unit Energy Optimization in 5G O-RAN
This AI breakthrough could slash billions from global telecom energy bills.
Deep Dive
Researchers have deployed a new AI system, EExApp, on a live 5G network that uses Graph Neural Networks and Reinforcement Learning to optimize energy use. It dynamically puts radio units to sleep and allocates resources, achieving significant power savings while maintaining service quality. Tested on commercial hardware with real traffic, it outperforms existing methods, directly tackling the massive 131 TWh annual energy consumption projected for global 5G base stations.
Why It Matters
This could dramatically reduce the operational costs and carbon footprint of the world's expanding 5G infrastructure.