6G Deterministic Network Technology Based on Hierarchical Reinforcement Learning Framework

作者：Yanxia Xing, Luyuan Yang, Xinghong Hu, Chengli Mei, Heng Wang, Jinyan Li · 年份：2023 · DOI：10.1109/bmsb58369.2023.10211210 · 被引用次数：1 · 研究领域：Software-Defined Networks and 5G、Advanced Computing and Algorithms、Network Time Synchronization Technologies

With the rapid development of mobile networks, various IoT services are booming, and these services have different characteristics and requirements. Among them, services such as remote surgery and intelligent manufacturing are sensitive to delay and have high requirements for delay certainty. In order to meet the needs of delay-sensitive services, 3GPP defined the technical solution of 5G network as a TSN bridge in version R16 to support deterministic services of mobile networks. However, this solution has technical shortcomings, for example, it does not support aperiodic service scheduling, and does not define a coordination mechanism between applications and networks. Therefore, this paper sorts out the promotion process of 3GPP supporting mobile deterministic networks, clarifies the reasons for the existence of problems, and provides a GCN-based hierarchical reinforcement learning model to flexibly handle dynamic and diverse traffic for 6G networks. This model realizes joint scheduling of periodic services and aperiodic services and makes full use of the characteristics of the slow change of the whole traffic and the rapid change of individuals in the network to update sub-models with different frequencies and ensures the deterministic delay.