In this course:
- Students will learn how to formulate finite and infinite horizon Markov decision processes (MDPs) under various objectives.
- Students will understand the concept of a structured policy and deriving sufficient conditions for its existence.
- Students will implement various MDP solution techniques (backward induction, value iteration, policy iteration, modified policy iteration, linear programming).
- Students will be exposed to the concepts of continuous time MDPs and partially observed MDPs (POMDPs).
- Students will be able to read and appreciate recent MDP publications in top journals.
For more detailed information on the course and the course program please read the course description.
|Date:||27-28 September, 9-10 November 2017|
|Time:||Precise schedule to be determined (per block, start at 10:00 on the first day, end at 16.00 on second day)|
|Location:||Eindhoven University of Technology|
|Course leader:||Prof. Lisa Maillart|
|ECTS:||1 (attendance) / 4 (attendance & assignment)|
|Course fee:||Free for TRAIL/Beta/OML members; others please contact the TRAIL office|
|Registration:||The registration form is closed. Please contact the Beta office via firstname.lastname@example.org|
This course is part of the GP-OML program (www.GP-OML.nl)