Adaptive Digital Twin Server Deployment for Dynamic Edge Networks in IoT System

Due to the heterogeneity and dynamicity of mobile edge computing (MEC) networks, providing users with high-quality services presents a significant challenge. However, the existing solutions primarily focus on adjusting user association policies to enable real-time interaction with their corresponding digital twin (DT), ignoring the critical role of an efficient deployment mechanism for digital twin servers (DTSs). In this paper, a two-tier DT model for dynamic edge networks is first designed that captures the real-time edge network dynamics and service requirements. Furthermore, a multi-stage adaptive joint deployment optimization algorithm (Multi-AJDO) is proposed, which involves two optimization stages. Specifically, the initial deployment of the DTSs is realized by using the historical operational status information of the MEC systems. Then, the DTSs capture the dynamic changes of the MEC systems in real-time and dynamically update the deployment accordingly. Numerical experiments verify the effectiveness of the proposed Multi-AJDO algorithm in terms of response latency, workload, and energy consumption.