Learning and Control
ECE 412/512

This course is an introduction to learning and control for dynamical systems. The course covers both model-based and data-driven approaches with an emphasis on dynamic programming techniques. Topics include robot modeling, Markov decision processes, reinforcement learning, and multi-agent systems.

    1. Tabular Methods
      1. Introduction
      2. Multi-armed bandits
      3. Discrete planning
      4. Dijkstra's algorithm and A*
      5. Deterministic dynamic programming
      6. Markov decision processes
      7. Dynamic programming
      8. Value iteration
      9. Policy iteration
      10. On-policy Monte Carlo
      11. Off-policy Monte Carlo
      12. Temporal difference learning
      13. SARSA and Q-learning
      14. Bootstrapping
      15. Planning and learning
    2. Approximation Methods
      1. Linear function approximation
      2. Features
      3. On-policy control with approximation
      4. Eligibility traces
      5. Policy gradient methods