Multi-Agent Resource Allocation for NOMA-Based Broadcasting Vehicular Networks

Abstract

Non-orthogonal multiple access (NOMA) has been considered a promising technology to further reduce latency and increase reliability in vehicular networks. Recently, most relevant studies have focused on maximizing the overall performance of NOMA-based vehicular networks, such as the total throughput or the average package receptive ratio. Different from existing works, our study investigates the optimal clustering and resource allocation in vehicle-to-infrastructure (V2I) communications that underlie a NOMA-based vehicle-to-vehicle (V2V) network. Our objective is to maximize the time-based proportional fairness for the V2V users while guaranteeing the quality of service for the embedded V2I users. Particularly, the position-based optimal clustering for V2V transmitters is derived first, and then subchannel assignment is obtained given the optimal clusters. Finally, a multi-agent distributed power control algorithm is performed to solve the problem efficiently. The simulation results validate that the proposed resource allocation algorithms outperform the reference algorithms.