Imported from https://github.com/jihoonerd/Deep-Reinforcement-Learning-with-Double-Q-learning