Bergmann, Lisa KatharinaLisa KatharinaBergmannFischer, KathrinKathrinFischerZurheide, SebastianSebastianZurheide2020-07-132020-07-132014-01-01PATAT 2014 - Proceedings of the 10th International Conference on the Practice and Theory of Automated Timetabling, 26 - 29 August 2014, York, United Kingdom / edited by Ender Özcan, Edmund K. Burke, Barry McCollum. - [S. l.] : PATAT, 2010. - Seite 82-101http://hdl.handle.net/11420/6736An examination timetable has to satisfy a vast variety of requirements to be not only feasible, but also to be convenient to all parties involved. Many different aspects, as e. g. spreading of exams for students’ convenience or fixing exams to certain days or rooms for teachers’ convenience, have been discussed in the literature. However, there are no model formulations which take all aspects relevant for this work into account. Therefore, in this work a new linear mixed-integer programming model for the exam timetabling problem is presented. The model uses a penalty-based goal programming approach to assure the construction of timetables which fulfill important requirements made by teachers, students and administrators. Based on this model, feasible solutions are derived by a standard solver and subsequently are further improved by a tabu-search procedure. The trade-off between different criteria is shown and some very promising results of the approach for a real-world data set are presented.enExamination timetablingLinear programmingMixed-integer programmingUniversity timetablingInformatikWirtschaftHandel, Kommunikation, VerkehrConference Paperhttp://www.patatconference.org/patat2014/proceedings.htmlConference Paper