Federated Learning for RAN Slicing in Beyond 5G Networks

Radio access network (RAN) slicing allows the division of the network into several logical networks tailored to different and varying service requirements in a sustainable way. It is thereby considered a key enabler of 5G and next generation networks. However, determining optimal strategies for RAN slicing remains a challenging issue. Using machine learning algorithms to address such a difficult problem is promising. However, due to the large differences imposed by RAN deployments and the disparity of their required services it is difficult to utilize the same slicing model across all the covered areas. Moreover, the data collected by each mobile virtual network operator (MVNO) in different areas is mostly limited and rarely shared among operators. Federated learning presents new opportunities for MVNOs to benefit from distributed training. In this paper, we propose a federated deep reinforcement learning (FDRL) approach to train bandwidth allocation models among MVNOs based on their interactions with their users. We evaluate the proposed approach through extensive simulations to show the importance of such collaboration in building efficient network slicing models.