Performance of adaptive RAT selection algorithms in 5G heterogeneous wireless networks

Abstract

Radio access technology (RAT) selection has recently received much attention from the research community due to the increasing deployment of heterogeneous wireless networks. Most prior works mainly focus on proposing an efficient algorithm that yields good performances, and evaluate their solutions on a certain network model, particularly in cases when every user can connect to all the available base stations (BSs). In this paper, we evaluate the impact of different aspects of network models, including (i) network topology and (ii) bandwidth allocation, on the performance of RAT selection algorithms to fully understand their limitations. Using simulations on realistic network models, this paper evaluates how different network parameters such as link density, user density or bandwidth distribution can lead to significant differences in algorithm performance. The paper demonstrates that among all the parameters, the performance of adaptive RAT selection algorithms are most significantly effected by the number of base stations that a user can connect to.