Data-Driven Spectrum Partition for Multiplexing URLLC and eMBB

Multiplexing ultra-reliable low-latency communications (URLLC) and enhanced mobile broadband (eMBB) are critical in the next generation mobile network. URLLC requires ultra-high reliability and extremely low latency, whereas eMBB demands high data rates. Thus, the coexistence system of URLLC and eMBB faces the challenge of sharing the spectrum efficiently and effectively. In this study, we comprehensively investigate the state of the art spectrum partition methods in combined URLLC and eMBB services. We formulate a joint optimization problem for maximizing the eMBB throughput and guaranteeing the URLLC performance. For the eMBB and URLLC multiplexing system, a full separative spectrum partition scheme based on data-driven genetic algorithm-based spectrum partition (DDGSP) is proposed. Our simulation results demonstrate that the proposed DDGSP can make the URLLC and eMBB coexistence system outperform the state-of-the-art methods in terms of the error rate and computational efficiency.