Lecture 8: Approximation in value space

 (Big slides Download Big slides) (Small slides Download Small slides) (Recording)

In this lecture, you will learn about

• More Q-learning examples
• Approximation in value space
  • Model-based/model-free and off-line/on-line
  • Multi-step lookahead
  • Enforced decompositions
  • Simplifying probabilistic structure

Recommended text:

Section 2.1-2.3 in Bertsekas' book. Links to an external site.

Further reading: Lincoln/Rantzer Relaxed Dynamic Programming Download Lincoln/Rantzer Relaxed Dynamic Programming