Cellular Traffic Offloading via Link Prediction in Opportunistic Networks

Abstract

With the emergence of affordable smart mobile devices (such as smartphones and tablets) running innovative applications have severely overloaded the cellular network. To cope with this issue, there have been many efforts to offload the traffic from the cellular network to other complement networks, for instance, Wi-Fi and device-to-device (D2D) communications. In this paper, we formulate the traffic offloading issue as a link prediction problem in opportunistic D2D network, which is targeted to alleviate the overburdened cellular network traffic and reduce the delay time. Considering the complexity of realistic networks, we employ three indexes of link prediction: common neighbors, Katz, and LRW index. To measure the performance of our proposed algorithm, we analyze it is offloading traffic capacity along with delay minimization among users in different networks. It is demonstrated that our proposed link prediction solution can efficiently offload up to 80% of the cellular traffic.