Optimization of Beamforming and Transmit Power Using DQN and Comparison With Traditional Techniques

As Artificial Intelligence (AI) and Machine Learning (ML) technologies continue to evolve, their integration into 5G and 6G networks has become critical for improving performance and efficiency. These technologies enhance wireless communication by leveraging deep neural networks and data-driven methods to optimize resource allocation, signal detection, and channel coding. They also address the growing need to reduce energy consumption in next-generation networks. Deep Q-Networks (DQN) play a key role in this transformation by enabling dynamic resource management, adaptive beamforming, and efficient network slicing. The use of DQN can improve spectrum utilization, power consumption, and beamforming in massive MIMO systems, supporting demands like eMBB and URLLC. In 6G, it is considered for optimizing resource allocation, dynamic beamforming, RIS control for Terahertz communication, and complex network slicing. This article focuses on maximizing channel capacity by deploying DQN to identify optimal beamforming angles and transmit power values for multiple transmitters and receivers. The model determines the best steering vectors and power levels to achieve maximum channel capacity while minimizing energy consumption and mutual interference. Its performance is compared to conventional optimization methods, demonstrating its effectiveness in enhancing the efficiency and reliability of 5G and 6G networks.