Dynamic Resource Allocation in Vehicular Networks Using Reinforcement Learning for Coverage Optimization and Operational Efficiency in Urban Environments

This project addresses the inefficiency in communication resource distribution within vehicular networks (VANET) by implementing an intelligent system based on reinforcement learning (Q-learning) and Markovian analysis. Given the challenges posed by high vehicular mobility and route variability in urban environments, the system models and trains a network controller capable of adjusting the activation of Roadside Units (RSU) in real-time. The approach seeks to maximize operational efficiency and coverage, reducing unnecessary energy consumption and improving service continuity. By integrating machine learning techniques with mobility simulations, the project proposes a scalable and adaptive solution that responds to dynamic user demand. Expected results include the validation of this model through simulations in specialized software, contributing to the development of more sustainable and efficient transport infrastructures. This work aligns with the Sustainable Development Goals (SDG 9), promoting innovation in intelligent transport systems and connected cities.<br/><br/><b>Goal</b>: <br/>To propose a mathematical-algorithmic model based on reinforcement learning for dynamic resource allocation in vehicular networks (VANET). The goal is to optimize coverage and operational efficiency in urban environments using machine learning algorithms.<br/><br/><b>Research lines</b>: <br/>Telecommunications and information technologies