Optimal caching placement for cache-enabled mobile social networks

Abstract

Caching at user devices is one promising solution to alleviate the heavy base station burden in mobile social networks. In this paper, we consider both physical coupling and social relationships among users, and then devise the optimal content placement scheme to maximize the offloading probability for mobile social networks. In particular, because the problem is NP-complete, we first propose a weighted fast unfolding algorithm to divide the users into communities according to users' social ties, locations and communication radius in the area. We then prove the optimal problem based on the proposed algorithm to be monotone submodular maximization subject to matroid constraint and a community based greedy algorithm is then presented. Simulation results indicate the proposed caching placement scheme achieves a great performance gain over existing approaches, and show that user devices should cache most popular files while there are seldom social ties in the network, otherwise users in the same community should store the same contents and store different contents for different communities.