Bensaid Radjaà, Goual Samira2019-05-222019-05-222018-07-02salle des thesesms-003-250-01https://dspace.univ-tlemcen.dz/handle/112/14201The new concept "Green Networking" can take advantage of the broad paradigm of cognitive radio. The goal of our final master's project is to find a mechanism that minimizes energy consumption by integrating it into the cognitive radio network. For this, we used the algorithm of Q-Learning, a reinforcement learning technique that will help the cognitive user to find the optimal channel that has a low transmission power by guaranteeing the needs of his application, and therefore a reduction in battery consumption.frCognitive Radio, Green Networking, reinforcement learning, Q-Learning.Radio Cognitive, Green Networking, Apprentissage par renforcement, Q-Learning.Thesis