Energy-Efficient Neural Network-Based Handover Prediction in 6G Networks

This paper presents a novel approach for balancing energy efficiency and QoS in the context of neural network-based predictive handover management. Our solution encompasses a two-step approach. Initially, we collect data on the energy consumption of the neural network model. Subsequently, a grid search algorithm is deployed to ascertain optimal hyper-parameter thresholds. This process effectively optimizes the trade-off between energy efficiency and QoS. Our experimental results showcase substantial energy savings without significantly compromising QoS. To the best of our knowledge, this paper is the first study exploring balance between energy efficiency and the performance of neural-based predictive handover.