Revenue Optimization in Video Caching Networks with Privacy-Preserving Demand Predictions
Yijing Zhang, Ferdous Pervej, Andreas F. Molisch
TL;DR
This work tackles revenue optimization for edge video caching under privacy constraints by combining privacy-preserving federated learning with Transformer-based multi-slot demand prediction. A two-stage approach first trains a Transformer via FL to forecast future user requests, then uses accuracy-weighted predictions to approximate actual demands and solve a greedy knapsack formulation for cache placement that balances delivery, backhaul, and update costs. The method yields near-centralized performance in prediction accuracy and outperforms baselines in cache-hit ratio and long-term revenue, demonstrating the practicality of privacy-preserving, revenue-oriented caching at the wireless edge. The framework enables coordinated decisions among users, ISPs, and CSPs without exposing private histories, with potential impact on real-world edge caching deployments.
Abstract
Performance of video streaming, which accounts for most of the traffic in wireless communication, can be significantly improved by caching popular videos at the wireless edge. Determining the cache content that optimizes performance (defined via a revenue function) is thus an important task, and prediction of the future demands based on past history can make this process much more efficient. However, since practical video caching networks involve various parties (e.g., users, isp, and csp) that do not wish to reveal information such as past history to each other, privacy-preserving solutions are required. Motivated by this, we propose a proactive caching method based on users' privacy-preserving multi-slot future demand predictions -- obtained from a trained Transformer -- to optimize revenue. Specifically, we first use a privacy-preserving fl algorithm to train a Transformer to predict multi-slot future demands of the users. However, prediction accuracy is not perfect and decreases the farther into the future the prediction is done. We model the impact of prediction errors invoking the file popularities, based on which we formulate a long-term system revenue optimization to make the cache placement decisions. As the formulated problem is NP-hard, we use a greedy algorithm to efficiently obtain an approximate solution. Simulation results validate that (i) the fl solution achieves results close to the centralized (non-privacy-preserving) solution and (ii) optimization of revenue may provide different solutions than the classical chr criterion.