OPTIMIZING NETWORK RADIO RESOURCE MANAGEMENT USING MACHINE LEARNING TECHNIQUE FOR TRAFFIC IMPROVEMENT

In the world of communication networks, which provide services to a variety of highly demanding applications, effective resource allocation is essential to guaranteeing maximum efficiency and user experience. This study presents the use of Long Short-Term Memory (LSTM)-based Radio Resource Management (RRM) approach for network optimization. Grid Search Optimization (GSO) is used to optimise the LSTM model's hyper parameter tuning, guaranteeing peak performance in dynamic network settings. The system uses guard interval insertion and frequency interleaving to reduce Inter-Symbol Interference (ISI) and burst errors. According to simulation results, the LSTM-RRM technique outperforms the Dynamic Radio Resource Management (DRRM) approach in terms of dual connectivity, throughput, and fairness. The effectiveness of the suggested approach in allocating resources was demonstrated by the up to 50% increase in User Equipment (UE) throughput and the 12% increase in dual connectivity for 30 UEs. The LSTM-RRM system, which was implemented with MATLAB, demonstrated scalability, robustness, and efficacy in mitigating congestion and enhancing Quality of Service (QoS) for communications between machines and humans. This study opens the door for more advancement in network performance optimisation by demonstrating the potential of LSTM and machine learning approaches for resource management optimisation in next-generation networks.