Fairness-aware eMBB/URLLC spectrum resource multiplexing to ensure reliability in vehicular networks

Spectrum resource multiplexing of enhanced Mobile BroadBand (eMBB) and Ultra-Reliable and Low Latency Communications (URLLC) via puncturing in vehicular networks has become an important research direction. The main issue in this direction is how to minimize the adverse impact of preemptive puncturing on eMBB traffic. However, the available eMBB/URLLC coexisting schemes lack considerations on the quality of service of a single eMBB user and fairness between users. In view of this, we formulate the eMBB/URLLC multiplexing problem to maximize eMBB reliability and fairness under the premise of meeting the requirements of URLLC as a nonlinear integer programming (NIP) in this article. To balance the performance and computational complexity, we decompose this NIP into two subproblems that be solved efficiently. For the eMBB spectrum resource allocation subproblem of maximizing the reliability and fairness under resource constraints, we prove that this problem is NP-hard, and design a Deep Q-Network (DQN)-based spectrum resource allocation algorithm and a two-stage heuristic spectrum resource allocation algorithm, respectively. For the URLLC preemption subproblem, we further propose bandwidth-sensitive URLLC spectrum resource preemption algorithm and time-sensitive URLLC spectrum resource preemption algorithm, respectively. Simulation results show that the proposed scheme can guarantee better fairness and reliability for eMBB users compared with the comparison schemes. Besides, for the eMBB resource allocation problem, the proposed scheme can achieve near-optimal performance with low complexity. The article proposes Deep Q-Network-based and two-stage heuristic enhanced Mobile BroadBand (eMBB) resource allocation algorithms, and bandwidth-sensitive and time-sensitive Ultra-Reliable and Low Latency Communications (URLLC) preemption strategies to maximize the reliability and fairness of eMBB while satisfying differentiated URLLC requirements. Simulation results show that the personalized service requirements of eMBB and URLLC users are guaranteed. image