Bayesian Optimization of Queuing-Based Multichannel URLLC Scheduling

This paper studies the allocation of shared resources between ultra-reliable low-latency communication (URLLC) and enhanced mobile broadband (eMBB) in emerging 5G beyond cellular networks. In this paper, we design a unique queuing mechanism for joint eMBB/URLLC system. The aim is to flexibly schedule URLLC traffic enhance total eMBB throughput reliability packets (i.e., probability not dropping each mini-slot) while maintaining satisfactory transmission latency as per 3GPP requirements. Precisely, by deriving steady-state probabilities queue backlog analytically, formulate stochastic optimization problem maximize normalized utility. Due nature objective function, it expensive evaluate any set inputs, thus Bayesian applied obtain optimal results such black-box function. Numerical demonstrate that proposed never violates requirement services but improves reliability. It also enhances compared method without queuing.