To assess the performance of a Q-learning technique for channel selection in LTE-U under different conditions of parameter settings and scenarios, and to optimize the performance of the technique by modifying their parameters and/or operation procedure.

The usage of LTE in unlicensed bands is a topic of great interest among the telecommunications companies and organizations, given the many advantages it can bring. For this reason, many feasibility studies have been conducted in order to define a standard for this technology. One of the approaches proposed is to apply the Q-Learning algorithm to the Channel Selection in LTE-U. Many promising results have already been achieved, so this technique deserves further insights. The objective of this Thesis is to assess the performance of this algorithm under different conditions of parameter settings and scenarios in order to optimize the performance of the technique by modifying the parameters and the operation procedure. The items analysed are the impact of the Initial Temperature, of the Learning Rate, of the positions of the SCs and of the number of active users.