Abstract:

Examination timetabling is an inevitable problem of educational institutions. Each institution has its own particular limitations; however, the main structure is the same: assigning exams to time slots and classrooms. Several institutions solve the problem manually, but it becomes more difficult every year with increasing numbers of students and limited resources. There are many studies in the literature addressing the examination timetabling problem (ETP) and providing high quality solutions within reasonable amounts of time. Nevertheless, almost none of them can be used in practice since they are not converted into a decision support system (DSS). Commercial DSSs, on the other hand, are generally transactionally based and do not have optimization capabilities, i.e. they prevent conflicts via functional user interfaces. In this study, we propose a mixed integer programming (MIP) model that addresses the ETP of the Industrial Engineering Department of Yıldız Technical University. The model, which is capable of solving a wide range of similar ETP instances, is embedded into a DSS in the form of a spreadsheet. Given the enrollment lists of the courses, it generates schedules with minimum conflicts and consecutive exams while addressing requests of the lecturers and students. It does not require any technical knowledge and can be used by an average spreadsheet user. Moreover, it is flexible in terms of use for scheduling problems of other educational institutions. Currently, the DSS is in use by the department and real-life instances can be solved within a few seconds, saving significant amount of man-hours.

Keywords: