Digital Twin Assisted Proactive Management in Zero Touch Networks
Digital Twin Assisted Proactive Management in Zero Touch Networks
The rapid expansion of cellular networks and rising demand for high-quality services require efficient and autonomous network management solutions. Zero Touch Network (ZTN) management has emerged as a key approach to automating network operations, minimizing manual intervention, and improving service reliability. Digital Twin (DT) creates a virtual representation of the physical network in realtime, allowing continuous monitoring, predictive analytics, and intelligent decision-making by simulating what-if scenarios. This paper integrates DT with ZTN proactive bandwidth management in end-to-end (E2E) next-generation networks. The integrated architecture applies Few-Shot Learning (FSL) to a memoryaugmented Bidirectional Long Short Term Memory (BiLSTM) model to predict a new network state to augment the known and trained states. Using Q-learning, it determines the optimal action (e.g. traffic shaping) under varying network conditions such that user Quality of Service (QoS) requirements are met. Three scenarios have been considered: 1) normal ZTN operation with closed-loop control, 2) a what-if scenario of DT, and 3) network state unknown to DT. The simulation results show that the network can adapt to underlying changing conditions. In addition, DT-assisted ZTN achieves better performance than the other techniques.
Tamizhelakkiya K、Dibakar Das、Komal Sharma、Jyotsna Bapat、Debabrata Das
自动化技术、自动化技术设备计算技术、计算机技术
Tamizhelakkiya K,Dibakar Das,Komal Sharma,Jyotsna Bapat,Debabrata Das.Digital Twin Assisted Proactive Management in Zero Touch Networks[EB/OL].(2025-08-25)[2025-09-05].https://arxiv.org/abs/2508.17941.点此复制
评论