Contextual Learning for Content Caching With Unknown Time-Varying Popularity Profiles via Incremental Clustering

With the rapid development of social networks and high-quality video sharing services, demand for delivering large quantity high quality contents under stringent end-to-end delay requirement is increasing. To meet this demand, we study content caching problem modelled as a Markov decision process in network edge server when popularity profiles are unknown time-varying. In order to adapt changing trends popularity, context-aware learning algorithm proposed. We prove that error scheme sublinear number requests. light learned popularities, reinforcement learning-based designed on top state-action-reward-state-action with function approximation. A reactive also proposed reduce complexity. The time complexities both schemes studied demonstrate their feasibility real systems theoretical analysis performed cache hit rate asymptotically converges optimal rate. Finally simulations presented superiority algorithms.