Multi-Agent Deep Reinforcement Learning Joint Beamforming for Slicing Resource Allocation

In 5G Radio Access Networks (RAN), network slicing is a crucial technology for offering a variety of services. Inter-slice resource allocation is important for dynamic service requirements. In order to implement inter-slice bandwidth resource allocation at a large time scale, we used the Multi-Agent deep reinforcement learning (DRL) Asynchronous Advantage Actor Critic (A3C) algorithm with a focus on maximizing the utility function of slices. In addition, we used the K-means algorithm to categorize users for beam learning. We used the proportional fair (PF) scheduling technique to allocate physical resource blocks (PRBs) within slices at a small time scale. The results show that the A3C algorithm has a very fast convergence speed for utility function and packet drop rate. It is superior to alternative approaches, and simulation results support the proposed approach.