Integer programming for minimal perturbation problems in university course timetabling

In this paper we present a general integer programming-based approach for the minimal perturbation problem (MPP) in university course timetabling. This problem arises when an existing timetable contains hard constraint violations, or infeasibilities, which need to be resolved. The objective is to resolve these infeasibilities while minimising the disruption or perturbation to the remainder of the timetable. This situation commonly occurs in practical timetabling, for example when there are unexpected changes to course enrolments or available rooms. Our method attempts to resolve each infeasibility in the smallest neighbourhood possible, by utilising the exactness of integer programming. Operating within a neighbourhood of minimal size keeps the computations fast, and does not permit large movements of course events, which cause widespread disruption to timetable structure. We demonstrate the application of this method using examples based on real data from the University of Auckland. Powered by Editorial Manager® and ProduXion Manager® from Aries Systems Corporation Noname manuscript No. (will be inserted by the editor) Integer Programming for Minimal Perturbation Problems in University Course Timetabling Antony E. Phillips · Cameron G. Walker · Matthias Ehrgott · David M. Ryan Abstract In this paper we present a general integer programming-based approach for the minimal perturbation problem (MPP) in university course timetabling.In this paper we present a general integer programming-based approach for the minimal perturbation problem (MPP) in university course timetabling. This problem arises when an existing timetable contains hard constraint violations, or infeasibilities, which need to be resolved. The objective is to resolve these infeasibilities while minimising the disruption or perturbation to the remainder of the timetable. This situation commonly occurs in practical timetabling, for example when there are unexpected changes to course enrolments or available rooms. Our method attempts to resolve each infeasibility in the smallest neighbourhood possible, by utilising the exactness of integer programming. Operating within a neighbourhood of minimal size keeps the computations fast, and does not permit large movements of course events, which cause widespread disruption to timetable structure. We demonstrate the application of this method using examples based on real data from the University of Auckland.