Q-Learning for Joint Access Decision in Heterogeneous Networks

In this paper, we focus on mobile-centered decision making in heterogeneous networks. We study the case where the JRRM decision is completely distributed so that mobile users have to decide to which of the available systems it is best to connect. To optimize their decision over the time, the mobiles implement a Q-learning algorithm that enables them to profit from their past experience. We study the performance of this decision-making framework in the case of a WiMAX/HSDPA heterogeneous network and show that the mobile decision is progressively enhanced until convergence.