PERFORMANCE IMPROVEMENT OF EDGE USERS BY UTILIZING JOINT SCHEDULING AND DEEP Q NETWORK

Abstract

Improving the user performance, especially for users at cell edge areas or edge users, is one of the most important requirements of 5G and 6G cellular networks which have a high densification of base stations. This paper presents the analysis and optimization of edge user performance under the dual-slope path loss model and Rayleigh fading by utilizing advanced techniques, called Joint Scheduling (JC), and Deep Q Network (DQN) utilizing Q-learning. The dual-slope path loss model is widely used to analyze the network performance of 5G and beyond since it is able to capture the attenuation properties of the real wireless transmission environment. In a JC system, a user can proactively select the base station with the highest downlink signal to communicate with. Meanwhile, the DQN dynamically adjusts the base station transmission power to adapt to change in the wireless environment. The analytical results indicate that the deployment of JC with DQN can increase the edge user performance up to 35% or 200% more than that of the system which applies DQN only or Maximum power (MP) respectively.