Authors: Alyssa Muller, Russ Ross
Mentors: Russ Ross
Insitution: Utah Tech University
University course timetabling assigns rooms and times to courses, considering instructor availability, curriculum conflicts, and quality of life considerations. Computing an optimal solution is computationally intractable. Researchers have refined approximation algorithms that yield far better results than the conventional pencil-and-paper approach used at many universities, including our own.
Transitioning to an automated system can be disruptive and poses both real and perceived risks to an institution. The implicit knowledge that humans bring to the problem is hard to fully capture within formal rules that a computer can understand. Our research addresses the impedance mismatch between abstract solutions and the messy real world.
In this project we build on prior research to fit the specific needs of our university. We will pilot our system with a set of departments in parallel with the traditional by hand process. We will analyze outcome quality through subjective assessment and quantitative comparison between human and machine generated timetables.