Proactive user association in wireless small cell networks via collaborative filtering

In this paper, we propose a proactive cell association framework for small cell networks, based on content recommendations. We focus on multimedia data services and characterize the user´s quality of experience (QoE) in terms of mean opinion scores (MOSs) that accurately reflect characteristics of the wireless transmissions and data applications. Based on such information, we propose a collaborative filtering approach that enables the small base stations (SBSs) to exploit the correlation across similar user and application requirements to predict the QoE that a UE-SBS association delivers. By leveraging such knowledge, the SBSs ultimately devise better-informed user-cell associations and more precise bandwidth allocations. We model this problem as a matching game with externalities, propose a decentralized algorithm based interference graph theory, and discuss the matching stability conditions in the proposed approach. Simulation results show that the proposed solution yields more precise bandwidth allocations reaching up to 19% spectrum savings, compared to conventional cell association approaches.