Markov Decision Processes (MDP)

Pre-work

The pre-work for this class includes reading the pages from 47 until 56 and from 58 until 67 of Sutton, R. S., & Barto, A. G. (2018). Reinforcement learning: An introduction. MIT press.

In this class, we will introduce Markov Decision Processes (MDPs), which are a fundamental framework for modeling decision-making problems in reinforcement learning. We will cover the following topics:

• Definition of MDPs
• Examples of MDPs
• Value functions and Optimal policies
• Bellman equations