Efficient Resource Allocation for NOMA-MEC System in Ultra-Dense Network: A Mean Field Game Approach

Abstract

Mobile edge computing (MEC) has become a promising technology to reduce the computational pressure and task delay of the users. Meanwhile, non-orthogonal multiple access (NOMA) can effectively improve the utilization of spectrum resources. Considering the advantages of MEC and NOMA, this paper investigates the resource allocation problem of the uplink NOMA-MEC system in an ultra-dense network (UDN), where each user will offload tasks to the MEC server according to the offloading policy. The optimization goal is to minimize energy consumption and task delay of users, which can improve the quality of service (QoS) for users. Firstly, a user cluster matching algorithm (UCMA) is proposed to improve the data transmission rate of users. Then, the UDN as a mean field game (MFG) framework, and a novel mean field-deep deterministic policy gradient (MF-DDPG) algorithm is proposed to obtain the solution of MFG because the DDPG method can reduce the complexity of the solution. The simulation results show that the MF-DDPG algorithm not only converges faster, but also effectively optimizes the energy consumption and task delay of the users.