Predictive Multi-Tier Collaborative Edge Caching in Mobile Edge Networks: A Digital Twin Approach

The evolution of mobile edge networks demands efficient content delivery mechanisms, as traditional caching algorithms struggle with dynamic user demand. We propose a predictive caching framework combining digital twins and machine learning to optimize content delivery. By predicting content popularity and dynamically adjusting Time-To- Live parameters, our system efficiently manages cache size, retaining content based on demand predictions to optimize space utilization and access times. The framework also enhances performance through load balancing and content sharing among edge nodes. Integrating digital twins enables real-time predictive caching, improving cache hit rates, reducing latency by 75%, and achieving a cache hit ratio of 0.83, thereby enhancing user Quality of Service.