ORIE 5570
Reinforcement Learning with Operations Research Applications
The ongoing information revolution and the advent of the big data era make quantitative methods in the business context indispensable. This course introduces reinforcement learning, decision-making under uncertainty, and related algorithms through the lens of OR applications. Examples will be drawn from real-world problems in operations, revenue management, queuing, finance, transportation, healthcare, and other areas of interest. The course will cover modeling and applications, basic theory, and algorithms.
Topics
- Be able to formalize dynamic decision problems under uncertainty as Markov decision processes.
- Learn about finite-horizon and infinite-horizon MDPs.
- Know how to solve MDPs exactly via dynamic programming as well as know how to solve MDPs approximately via reinforcement learning.
- Learn to read the technical literature in operations research, machine learning, and control literature.
- Gain hands-on experience in implementing and applying various exact and approximate algorithms.
Official roster learning outcomes
Prerequisites & course information
ORIE 3500 or ORIE 5500; ORIE 3510 or ORIE 5510; or equivalents.
Sources
Catalogue variants
- Reinforcement Learning with Operations Research Applications · 3.0 · Graduate