Radio Resource Allocation for 5G Networks Using Deep Reinforcement Learning

The emerging 5G mobile network technology is intended to deliver an effective platform for the communication of devices within users. In this study, the proposed method enables the improvement of the connectivity for the IoT with network slicing concepts. The allocation of resources is based on individual network slices specified as audio, texting, video, and browsing. Then, to maximize the average resource allocation performance, a deep reinforcement learning (DRL) optimization method is proposed. For user resource request queue round-robin scheduling algorithm is applied to control the traffic for sharing resources. Finally, this work addresses two leading issues as allocating network slices to the users, balancing resource blocks, and quality of service for fair resource allocation.