Model-Based Methods in Reinforcement Learning - Part 2: Model-based Control