The field of wireless communication systems is moving towards the integration of Unmanned Aerial Vehicles (UAVs) to enhance system flexibility and performance. Recent developments have focused on optimizing beamforming, phase shifts, and UAV positioning to maximize system capacity. Deep reinforcement learning algorithms have been proposed to achieve this goal, balancing exploration and evaluation to improve system performance. Another area of research has been on using UAVs to address the performance gap between cell-edge and cell-center users in cellular networks, with distance-based criteria and power control policies showing promise. Additionally, multi-agent deep reinforcement learning frameworks have been developed to jointly optimize UAV positioning, resource allocation, and quality of service in 5G network slicing. Notable papers include:

• One that proposes a convolution-augmented deep deterministic policy gradient algorithm to maximize the sum rate of a STAR-RIS-UAV-assisted wireless communication system.
• Another that employs a Deep Q-Network learning framework to optimize transmission power allocation in a UAV-assisted cellular network.
• A comparative study of multi-agent deep reinforcement learning algorithms for UAV-assisted 5G network slicing.