Reinforcement Learning and Learning-based Control
This is an advanced course intended for master students in mechanical engineering or computer science who are interested in learning-based control and reinforcement learning.
Course description
The class Reinforcement Learning and Learning-based Control covers state of the art methods for data driven learning of controls. The first part of the course introduces reinforcement learning, starting from basic concepts and building to current state-of-the-art algorithms. The second part of the course gives an overview over methods for learning-based model generation and controller tuning. The class emphasizes both solid theoretical understanding of the different mechanisms, as well as hands-on programming exercises to apply them to problems in the context of engineering.
Contact
Contact
Class Outline
- Reinforcement learning basics
- Reinforcement learning with function approximation
- Learning-based control: Model learning
- Learning-based control: Controller tuning
Lectures and Exercises
The course is given during the summer term and is held in English.
For information on lectures and exercises check with RWTHonline.
Consultation Hours
Consultation hours are offered on appointment.