Machine Learning-Assisted Resource Allocation in 5G and Beyond Wireless Networks

Abstract

The increasing complexity of 5G and beyond wireless networks requires intelligent techniques for managing limited radio resources under changing traffic and channel conditions. This study proposes a machine learning-assisted resource allocation framework that combines short-term traffic prediction with deep reinforcement learning for dynamic bandwidth and power allocation. Simulation-based evaluation indicates that the proposed approach can improve network throughput, reduce communication delay and maintain efficient spectrum utilization compared with conventional scheduling approaches.