AI-Driven Digital Twins: Optimizing 5G/6G Network Slicing with NTNs

2025-05-13

AI Summary

To address inefficient network slicing resource allocation in 5G/6G non-terrestrial networks (NTNs) caused by high user mobility and dynamic service demands, this paper proposes an AI-driven digital twin architecture. The architecture uniquely integrates digital twin technology with deep deterministic policy gradient (DDPG)-based reinforcement learning to enable real-time slice-state sensing, accurate prediction, and closed-loop optimization. It supports adaptive bandwidth allocation for enhanced mobile broadband (eMBB) slices and is validated under representative NTN scenarios—including disaster relief and urban occlusion. Simulation results demonstrate a 25% reduction in end-to-end latency and significant improvement in resource utilization, thereby effectively supporting mission-critical, high-dynamics, low-latency services. This work establishes a scalable, intelligent paradigm for NTN slicing management.

Technology Category

• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems
• Search and Optimization: Sampling/Simulation-based Search
• Machine Learning: Deep Neural Architectures and Foundation Models

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Social Networks and Social Media: Generative AI / large language models and their impact on social systems

Abstract

Network slicing in 5G/6G Non-Terrestrial Network (NTN) is confronted with mobility and traffic variability. An artificial intelligence (AI)-based digital twin (DT) architecture with deep reinforcement learning (DRL) using Deep deterministic policy gradient (DDPG) is proposed for dynamic optimization of resource allocation. DT virtualizes network states to enable predictive analysis, while DRL changes bandwidth for eMBB slice. Simulations show a 25% latency reduction compared to static methods, with enhanced resource utilization. This scalable solution supports 5G/6G NTN applications like disaster recovery and urban blockage.

Problem

• Optimizing 5G/6G network slicing with AI-driven digital twins
• Addressing mobility and traffic variability in NTN resource allocation
• Reducing latency by 25% via dynamic DRL-based bandwidth adjustment

Innovation

• AI-driven digital twin for network slicing optimization
• Deep reinforcement learning with DDPG algorithm
• Dynamic resource allocation reducing latency by 25%