An ML-Based Resource Allocation Scheme for Energy Optimization in 5G NR

This paper proposes a machine learning (ML)-based energy optimization framework for 5G New Radio (5G NR) utilizing a Classification and Regression Tree (CART) algorithm. The methodology implements dynamic cell resource reconfiguration through predictive load forecasting, achieving a 42.3% reduction in energy consumption, while maintaining QoS parameters within 3GPP-specified thresholds. A case study with a network layout made up of an inter-band NR-NR Dual Connectivity (DC) was simulated to quantitatively validate our model.