Dynamic Optimization and Reinforcement Learning, WS 2020/21, Jörg Lingens, Andreas Masuhr, Mark Trede

Machine Learning: Dynamic Optimization and Reinforcement Learning

Reinforcement learning is the third and most advanced pillar of machine learning (the other two being supervised and unsupervised learning). In this course, we will connect reinforcement learning to the classical deterministic and stochastic dynamic optimization approaches.

Students learn analytical and numerical methods to solve dynamic optimization problems. Both classical methods and modern machine learning methods (reinforcement learning) are introduced. Our focus is on the practical implementation.

Participants should have a solid knowledge in mathematics and statistics. Basic knowledge of R (or other modern programming languages like Matlab, Python or Julia) are helpful but not indispensable.

The course is designed for PhD students but also open to Master students.

The final examination will consist of a take-home exercise and a presentation of the results.