Every new academic year, scheduling new timetables due to disruptions is a major problem for universities. However, computing a new timetable from scratch may be unnecessarily expensive. Furthermore, this process may produce a significantly different timetable which in many cases is undesirable for all parties involved. For this reason, we aim to find a new feasible timetable while minimizing the number of perturbations relative to the original disrupted timetable. The contribution of this paper is a maximum satisfiability (MaxSAT) encoding to solve large and complex university timetabling problem instances which can be subject to disruptions. To validate the MaxSAT encoding, we evaluate university timetabling real-world instances from the International Timetabling Competition (ITC) 2019. We consider the originally found solutions as a starting point, to evaluate the capacity of the proposed MaxSAT encoding to find a new solution with minimal perturbation. Overall, our model is able to efficiently solve the disrupted instances.
CITATION STYLE
Lemos, A., Monteiro, P. T., & Lynce, I. (2020). Minimal Perturbation in University Timetabling with Maximum Satisfiability. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 12296 LNCS, pp. 317–333). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-3-030-58942-4_21
Mendeley helps you to discover research relevant for your work.