An on-line radio access technology selection algorithm in an LTE-WiFi network

Abstract

In a Heterogeneous Network (HetNet) comprising of multiple Radio Access Technologies (RATs), a user can be associated with a particular RAT and can be steered to other RATs in a seamless manner. To handle the rapid growth of data traffic, offloading of mobile data to Wireless Fidelity (WiFi) has been proposed in a Long Term Evolution (LTE) based HetNet. In this paper, we consider an optimal RAT selection problem in an offload-capable LTE-WiFi system with an objective of maximizing the total system throughput subject to a constraint on the voice user blocking probability. An on-line algorithm for optimal RAT selection is proposed based on a Relative Value Iteration Algorithm (RVIA) centric Q-learning approach. The proposed algorithm can be implemented without any explicit knowledge of arrival processes of voice and data users. Simulation results are presented to exhibit the convergence behavior of the proposed scheme to the optimal policy.