A Stackelberg Based Leadership Algorithm for RAT Selection in Heterogeneous Networks

To address the increasing demand of data rate for data-hungry smart devices, next generation cellular networks are heterogeneous (HetNet) environments where multiple types of base transceiver stations provide service to users using different radio access technologies (RATs). Mobile network operators deploy WiFi or micro-cell base stations along with macro-cell base stations to enhance the coverage, provide load balancing and extensively improve the throughput of their networks. Various challenges arise when users have the option for their access connection, where the most important one is the selection of the appropriate RAT. Different objectives can be considered for RAT selection, including throughput, power consumption, fairness, and etc. In this paper, we study a RAT selection problem for a HetNet network with WiFi and a 5G gNB. We model this multi-objective problem using a Stackelberg game. To tackle the computational complexity of this model, we propose a novel technique by estimating the interacting functions between the users and the RATs. The numerical results show the tightness of the estimation. Moreover, the RAT selection heuristic solution is 90 % of the times in 30 % proximity of the optimal solution.