Energy-Efficient Resource Allocation With Flexible Frame Structure for Hybrid eMBB and URLLC Services

The key objective of the fifth generation (5G) wireless technology is to support services with highly diverse requirements, which necessitates the specification of flexible numerology and frame structure for radio resource allocation in 5G new radio (NR). In this article, flexible 2-dimensional resource allocation is investigated to exploit the frequency and time diversities of the resource grid, and eventually maximize the overall system energy efficiency (EE) in the scenario where enhanced mobile broadband (eMBB) and ultra-reliable and low latency communication (URLLC) services coexist. A new definition of EE is proposed to capture and fulfill the heterogeneous quality-of-service (QoS) requirements of eMBB and URLLC services. By first sliding a window on the resource grid to determine the candidate resource pool, and then allocating resource blocks (RBs) therein by leveraging linear programming (LP) relaxation, we propose a sliding window (SW) based algorithm with a low complexity for the NP-hard EE optimization problem. In addition, we also analyze the achievable EE performance of the proposed algorithm to demonstrate its scalability with respect to the grid sizes and the maximum transmit power to provide an in-depth understanding. Numerical results show that the proposed algorithm can bring about as much as 16.7% power consumption reduction, and improve EE by about 29.3% compared to other baseline systems.