Dynamics of User Association in Non-Orthogonal Multiple Access Enhanced Fog Radio Access Networks

Abstract

In order to improve spectrum efficiency (SE) and efficiently support the massive connections in fog radio access networks (F-RANs), a non-orthogonal multiple access (NOMA) enhanced F-RANs architecture is proposed. Since the network congestion may cause system performance degradation, choosing appropriate communication mode is necessary to performance optimization in the complex scenarios. We study dynamic user association mode selection with two algorithms based on evolutionary game and reinforcement learning, respectively, where the payoff function jointly considers quality of service, delay cost, and power consumption. Simulation results show that adopting NOMA technology in F-RANs can effectively improve SE and both of two proposed algorithms can achieve the equilibrium of dynamic user association.