Optimal radio access technology selection policy for LTE-WiFi network

As a part of Fourth Generation (4G) wireless communication system, a user can associate with any radio access technology (RAT) when there are multiple RATs available and can move seamlessly among them. To handle the explosive growth of traffic in cellular network, the idea of mobile data traffic offloading to Wireless Fidelity (WiFi) has been proposed. In this paper, we focus on mobile data offload assisted optimal association problem in a heterogeneous network. The problem has been formulated under the framework of constrained Markov decision process (CMDP). Value iteration and gradient descent algorithm has been used to determine the optimal policy. We aim to maximize the expected average per-user throughput of the system subject to a constraint on blocking probability of voice users. In our model, we consider the possibility of mobile data user offload from one RAT to another during association or departure of a user. Optimal policy has been found to follow a threshold structure. Numerical results are presented to show how the average per-user throughput of the system varies under different load conditions.