Reinforcement Learning in R: An Introduction to Dynamic Programming


Reinforcement Learning is an approach to learning that attempts to maximize a cumulative reward based on a set of actions and states. The techniques are very popular within operations research and control theory. It does not fall under the traditional paradigms of supervised or unsupervised learning because correct inputs/outputs are never provided. The algorithm is instead presented with some notion of doing well or poorly at each step.

There are several different methods for optimizing a reinforcement learning problem; I will be focusing on Dynamic Programming. Dynamic programming (or DP) is a powerful optimization technique that consists of breaking a problem down into smaller sub-problems, where the sub-problems are not independent . This is useful both in mathematics (especially fields like economics and operations research) and computer science …