The DSS group participates in the project Data Science for University Management (DSUM) together with the Management Science group at DTU-Management Engineering and the private company MaCom A/S. The goal of the project is to develop a software to produce optimal course timetables for universities. The project will run from 2019 through 2022 and has a total budget of 13 milions Danish krone of which 6 milions provided by the Innovation Fund Denmark.
Scheduling university courses is the activity of producing timetables that indicate for each class of a set of courses the room and the time slot within a semester where the class will take place. Course timetables affect the life of the persons involved: students and teachers. Overlaps between classes that are attended by the same persons must be avoided and rooms can host one single lecture at a time. Moreover, undesirable slots like those late during the day must be avoided as much as possible. The resulting scheduling problem is difficult to solve. The number of different possibilities is huge and examining them all cannot be done either with the fastest super computer on Earth. This difficulty is caputred in computer science terms by the notion of NP-hardness. The DSUM project will develop techniques to overcome this burden and to produce anyway optimal or near optimal timetables. During the project different techniques from the fields of operations research and artificial intelligence will be applied, configured and compared. Examples of these techniques are: mixed integer programming, constraint programming, local search, metaheuristics, and algorithm portfolios.