ADAPTIVE BEAMFORMING SYSTEM FOR 5G NETWORKS USING CONVEX OPTIMIZATION ALGORITHMS

Abstract

The research provides a novel flexible beamforming system to 5G Networks that involves Hybrid Convex Optimization and machine learning integration. The proposed system will address the issue of beamforming patterns optimization under 5G dynamic conditions as the latter is significantly influenced by the variables of interference and mobility. The system shows dynamically-adapting to changing network conditions to attain improved signal-to-interference-plus-noise ratio (SINR), throughput and energy efficiency, by executing time-varying convex optimization algorithms with PyTorch ecosystem machine learning models to allow dynamically-adapted machine learning systems in dynamically-changing conditions. The hybrid solution possesses 28 percent SINR combination, 18 percent superiority and 22 percent improvement in comparison to the normal techniques. Integrating machine learning permits modifying the operations to enhance their functioning in the dense settings. The simulation impediments demonstrate the system to be scalable and efficient in spite of the constraint of the pure, convex or machine learning-based approaches. The article provides a successful method of enhancing the beamforming of 5G networks in future.