QTM 7515: Reinforcement Learning and Sequential Decision Making in Practice
3 credits
This course is a practical hands-on introduction to analytical models for sequential decision making, which involves making a series of decisions over time with the goal of improving metrics of interest. Reinforcement Learning (RL) - a field of Artificial Intelligence (Al) that brought us Google DeepMind's AlphaGo and self-driving cars - is a focal area of the course. You will learn about classical RL concepts and techniques, such as bandit problems, Markov Decision Processes, dynamic programming, Monte Carlo methods, and temporal difference learning. However, the main overall objective of the course is to illustrate the interaction and use of numerous analytical techniques necessary to solve actual business problems from diverse areas like marketing, inventory management, supply chain management, healthcare operations, manufacturing, financial services, and humanitarian logistics. The course is set up as a guided journey through the lifecycle of several projects that the instructor has completed with actual organizations, increasing in complexity as the semester progresses. The open-source programming language Python and possibly other software tools will be utilized as needed during the course.
Note: This course is self contained. However, you will get more out of the course if you take it after taking at least one other modeling course at Babson (e.g., QTM7800, QTM7571/QTM6300, QTM9510, ECN7520/ECN6300/ECN63100). You do not need to have programming experience before taking the course.
Prerequisite: (QTM 7800 for MBA students)