The research from CASTLE Laboratory (and increasingly from its sister activity, PENSA) has contributed a substantial body of material that supports several courses that are taught by Warren Powell in the Department of Operations Research and Financial Engineering. Below are short summaries of courses currently being taught. For more information, click on the link associated with each course.
ORF 411 – Operations and Information Engineering (check out the set of lectures that are now provided).
This is the capstone course for ORFE (required of all majors), which has several major themes:
- Modeling dynamic systems – Students learn the five core components of a stochastic, dynamic system, which they use to apply to problems of increasing complexity. Special emphasis placed on organizing the flow of information and decisions.
- Dynamic resource management – The course is set in the context of managing physical, financial and informational resources.
- Optimization – In deterministic optimization, you are looking for the best decision; in the presence of uncertainty, we are looking for the best policy.
- Team work – The centerpiece of the course (if you ask the students) is the Orange Juice game, a major competition involving teams of five students who have to make decisions about pricing, purchasing futures contracts, planning manufacturing and distribution facilities. The game requires filling out a spreadsheet which guides the operation of the company for a year. We have four slow rounds (one spreadsheet per week), followed by the “Orange Bowl” where we run about 14-16 more iterations in a four hour period.
An article on the OJ game, which includes a short video of students playing the game, is available here.
This is an elective course that attracts about 30-40 undergraduates each year. It grew out of a body of research that also produced a book by the same name. The course focuses on the challenge of collecting information, when these decisions are time-consuming and/or expensive.
ORF 544 is a course that teaches a unified framework for stochastic optimization, spanning classical fields such as stochastic search, dynamic programming, stochastic programming, and optimal (stochastic) control, and related fields including reinforcement learning/approximate dynamic programming, model predictive control, online computation, and decision trees. We consider both offline formulations (e.g.. ranking and selection) and online (e.g. multiarmed bandit problems), in a simple, integrated framework. Students will see how these diverse fields can be viewed from this common framework, opening up both new problem classes, as well as new solution approaches.
The course will emphasize proper modeling, and the study of general computational frameworks rather than the highly specialized problems that yield analytical solutions. General modeling formulations will be illustrated with a wide range of applications drawn from different fields (depending on the makeup of the class), spanning business, economics, engineering, and the sciences. The course will be guided by extensive experience with computational work on a range of applications.