Energy Efficiency Evaluation and Optimization in Wireless Networks Using Graph Neural Networks and Adaptive Base Station Management

Abstract

Energy efficiency has become an important issue in modern wireless networks due to the rapid growth of users and the dense deployment of 4G and 5G base stations. Since base stations consume a major share of network energy, improving their operation is essential for reducing power usage and operational costs. This paper presents an Energy Efficiency Evaluation Framework (E3F) that evaluates network energy performance using detailed base station power models and realistic traffic variations. The framework is applied to a 3GPP LTE network to identify practical energy-saving opportunities. In addition, adaptive base station management using ON–OFF switching is explored as an effective method to reduce energy consumption during low traffic periods while maintaining Quality of Service (QoS). The role of Graph Neural Networks (GNNs) in optimizing large-scale wireless networks is also discussed. Simulation results show that adaptive base station operation can significantly improve energy efficiency with minimal impact on network performance, making it a promising approach for future energy-aware wireless networks.