Two years ago, the Rotman School of Management at the University of Toronto adopted an algorithm called Course Match to assign class schedules to students. Until that point, they had relied on an auction that was beginning to falter -- students found it frustrating and it occasionally failed to produce complete schedules.

Course Match was a radical departure: it creates a distribution of optimized class schedules for every student based on a ranking of their preferred classes. It was also a huge success: manual effort was mostly eliminated, students got schedules they liked, and the school was able to identify a number of unnecessary sections using the algorithm's demand data.

This was not entirely surprising, as Course Match had already seen years of successful deployment at the Wharton School at the University of Pennsylvania, where it was originally developed by applying Nobel Prize-winning economics to the problem of matching students to class schedules.

This session will walk participants through the origins and capabilities of the algorithm, what it meant to change the elective course enrolment system for a school, and how real-world results confirmed the algorithm's theoretical promise.

Particular emphasis will be placed on the generation and use of planning data, and how Rotman was able to make difficult decisions on the basis of an objective fact base that they did not have before.

Submission ID: 

5203

Day:

Tuesday, July 16, 2019

Time:

8:00 AM - 9:00 AM

Room:

Neopolitan II