Deep Q-Learning and Double Deep Q-Learning for optimizing transitions within deterministic environments

Gherbia, OussamaKohil, YasserLachekhab (Kahoul), Fadhila (Promoteur)2024-12-022024-12-022024https://dspace.univ-boumerdes.dz/handle/123456789/14837120 p. : ill. ; 30 cmThe Development of intelligent decision-making algorithms has become important for technological growth in the era of artificial intelligence (AI), machine learning, and deep learning. These sectors have not only transformed whole businesses, but they have also completely changed the way we use technology every day. Reinforcement learning, a subfield of machine learning that focuses on teaching agents to make logical choices in dynamic environments. This document offers a deep review of reinforcement learning, with special attention paid to three essential algorithms: Q-learning , deep Q-learning, and double deep Q-learning. Each of these algorithms offers ever more advanced methods for handling challenging issues in a variety of fields, matting an important turning point in the pursuit of intelligent decision-making.enProcédés de fabrication : AutomatisationCommande automatiqueApprentissage profondApprentissage automatiqueAlgorithmesRéseau neuronal artificielDeep Q-Learning and Double Deep Q-Learning for optimizing transitions within deterministic environmentsThesis