User Mobility-Aware Proactive Video Caching Migration in Mobile Edge Computing

Abstract

In recent years, the rapid development of short-video platforms has led to a surge in network traffic generated by mobile devices. The frequent changes in user location pose significant challenges to the continuity and smoothness of video content services. To address the decline in service quality caused by user mobility, this paper proposes a user mobility-aware cache migration strategy within a cloud–edge collaborative framework. First, a cache migration model is constructed, taking into account both the benefits and costs of cache migration. Then, a Transformer–LSTM neural network model is developed to predict users’ next locations based on historical location sequences. Finally, combining the prediction results with migration costs, we design a cache migration algorithm named MGCM, which is based on migration gain to proactively migrate user-interested content, ensuring continuous and low-latency video services. Simulation experiments under the evaluated settings show that the Transformer–LSTM model outperforms the compared prediction models, achieving up to 88% prediction accuracy on the selected trajectories. Compared with the baseline migration strategies, the MGCM algorithm improves the cache hit rate by up to 13.5 percentage points in the evaluated scenarios. The proposed strategy enhances user experience and provides a feasible cache migration solution for mobile edge computing scenarios.