Asignación dinámica de recursos en redes VANET mediante aprendizaje por refuerzo

Abstract

In this paper we propose the deployment of the RSU infrastructure through a reinforcement learning algorithm to optimally distribute the resources in the vehicular network. The main objective of our study is to use the Q-Learning algorithm to allocate channels from a controller to the RSUs in the planning scenario. With this initial deployment and its mobility, an analysis will be performed through an optimization model to obtain a minimum number of devices in the simulated VANET infrastructure. The learning of the algorithm on the scenario is dynamically established in relation to the vehicular demand and its coverage restrictions for a V2I communication.