Joint Long-Term User Scheduling and Beamforming Design for Burst IIoT

With the spurt progress of the Industrial Internet of Things (IIoT), enhancing the IIoT’s ability to deal with burst traffic is essential. User perceived throughput (UPT) is an appropriate metric for evaluating the real user experience of burst traffic since it takes buffer state into account, which is overlooked by traditional throughput calculations. However, the inherent strong randomness coupled with the emphasis on long-term performance, poses significant challenges to UPT optimization in such a stochastic industrial environment with time-varying characteristics. In this article, we focus on the UPT to measure the Quality of Experience (QoE) for burst IIoT, formulating a long-term UPT-maximization problem in the system. To address this, joint user scheduling at the medium access control (MAC) layer and beamforming design at the physical (PHY) layer optimization is considered. By applying the Lyapunov optimization theory, the long-term optimization problem is decoupled into more manageable short-term problems. Subsequently, we propose both centralized and decentralized algorithms, based on fractional programming (FP) and successive convex approximation (SCA), respectively. Simulation results verify the effectiveness of the proposed algorithms compared to the baselines, and FP-based collaborative QoE-driven cross-layer co-design algorithm (FPCQA) demonstrates superior UPT gain.