Value Based Methods Reinforcement Learning