In RL, the main goal is to learn from interaction. We want agents to learn a behavior, a way of selecting actions in given situations, to achieve some goal. The main difference between classical programming or planning is that we do not want to code the planning software explicitly on…