Femtocaching in wireless video networks: Distributed framework based on exact potential game

Femtocaching is a distributed caching system designed to assist the macro base station by handling the popular content downloading. Usually, femto base stations (FBSs) have weak backhaul links but large storage capacity. When the requested files are cached, FBSs serve mobile users (MUs) through high-rate wireless link and avoid the backhaul bottleneck. However, the distributed caching problem is NP-complete. Previous works solve this problem with greedy algorithms in a centralized way. In this paper, we formulate this problem as an exact potential game (EPG), where each FBS acts as a game player with the constraint of cache size. The best Nash equilibrium (NE) is the global optimum solution of the distributed caching problem. We design an iterative algorithm in a decentralized way, where only local information exchange is needed. Theoretical analysis shows the best NE can be achieved and simulation results illustrate that the proposed algorithm performs better than the simple popular cache system and the traditional methods.