User matching game in virtualized 5G cellular networks

Abstract

Recently, wireless virtualization has attracted more and more attentions from research communities. With virtualization, resource utilization is higher, system performance is improved and the investment capital is lower. However, there are remaining challenges to be addressed before wireless virtualization is widespread deployed. One challenge is user association which has great influence on the performance of the wireless virtualization network. Traditionally, user associates to the base station (BS) who provides the highest signal to interference plus noise ratio (max-SINR). However, this scheme may not guarantee the quality of service (QoS) of users and backhaul constraint of infrastructure providers (InPs). This motivates us to investigate the user association problem in virtualized cellular networks in this paper. We formulate the user-BS association as a matching game. We show that the preferences used to rank one another are independent and affected by the existing matching. Therefore, this game can be classified as the one-to-many matching game with externalities. In addition, we present a distributed algorithm to find the stable outcome of this game. Simulation results show that the presented association scheme is better than traditional scheme.