The objective of reinforcement learning is to discover a policy, and that is a mapping from states to actions, that maximizes the predicted cumulative reward as time passes. As an example, think about a spouse and children of up to three generations. A grandfather and mother have their little one