Value-based Methods in Deep Reinforcement Learning