Computational models for timetabling problem

Abstract

The timetabling problem is to find a schedule of activities in space/time that satisfies a prescribed set of operational and resource constraints and which maximizes an objective function that reflects the value of the schedule. Constructing an effective timetable is always a challenging task for any scheduler. Most literature research focuses on specic applications and the resulting models are not easily applied to problems other than those for which they were designed for. In this paper, we construct a general model for university course timetabling. Our model incorporates a total of 17 different types of requirements identified in the literature as well as three new constraint types that we think should be part of the restrictions in a general university based timetabling model. An integer programming (IP) model is presented which incorporates restrictions that need to be satisfied and requests that are included in the objective function. We implement and test our models using the AIMMS mathematical software package. Computational results on a number of case studies are favorable and demonstrate the value of our approach.