Game Theory Based Base Station Association for Flexible Duplex Networks

Abstract

Flexible duplex has been seen as a promising technology for dealing with varying traffic asymmetries in the fifth generation (5G) system. However, the main issue that limits its performance is cross-link interference, i.e., the interference between downlink (DL) and uplink (UL). To solve this interference problem, base station (BS) association mechanism is considered to be an effective way, in which the cellular system determines which user equipment (UE) is associated with a particular BS before data transmission. In this work, we propose a distributed BS association algorithm for flexible duplex systems to maximize network sum rate. In particular, by taking the influence of interference caused by the UE's associated BS into consideration, we formulate the BS association problem as a BS association game, and prove that it is an exact potential game (EPG) which has at least one pure strategy Nash equilibrium (NE). Then a distributed BS association algorithm is proposed to achieve NE with local information. Simulation results show that our algorithm can not only achieve an outstanding performance in terms of network sum rate, especially when the DL and UL traffic are asymmetric, but also balance the sum rates between the DL and UL UEs by adjusting a weight parameter in the proposed algorithm.