Incentivized Social-Aware Proactive Device Caching With User Preference Prediction

In order to offload network traffic, we design a device caching strategy by jointly considering a popularity model, social influence and incentive design in this paper. Firstly, we propose a prediction model by virtue of users' social network information to evaluate users' encounter probability. Moreover, users' content preference is predicted using users' context information. Based on these predicted values, a content placement algorithm is described provided that the users will fully cooperate to optimize system performance. Thereafter, a more practical scenario where users are selfish and unwilling to devote their resources is considered. A Stackelberg game is established between the mobile network operator (MNO) and users by providing an incentive to encourage cooperation. Device caching strategy and incentive price design are determined by analyzing the Stackelberg game and finding the Stackelberg equilibrium point. We verify the effectiveness of our prediction models utilizing real data sets. Simulation results show that the cache hit ratio can be considerably improved by exploiting social and context information. Incentive design and profit analysis are also thoroughly investigated.