QoE-Driven Wireless Communication Resource Allocation Based on Digital Twin Edge Network

Abstract

As a real-time representation of physical entities in the digital world, digital twin (DT) has been widely used in many industrial fields, which has brought remarkable efficiency improvement and cost reduction. With the evolution of 6G network, the requirements for ultra-low delay and intelligence are gradually improved, and digital twin edge network (DTEN) came into being. DTEN mainly collects real-time information of physical objects through edge nodes such as BSs and APs, and builds dynamic models on demand based on these information, which has the ability of description, diagnosis, prediction and decision-making. Aiming at the traffic management and resource allocation in DTEN, a QoE-driven wireless communication resource allocation method based on DTEN is proposed, and the overall architecture design of DTEN system is completed. By integrating DT and RL technologies, the DTEN system models can be self-improved and dynamically adjusted, and the data-driven resource allocation model can be updated efficiently. The simulation results show that the proposed algorithm guarantees a high level of QoE for users. Compared with Q-Learning algorithm alone, the proposed algorithm can support more users’ access under the same congestion performance and reduce the iteration times of the algorithm by about 70%.