Construction of an automated examination timetabling system for École Polytechnique de Louvain

Examination timetabling is one of the major administrative task addressed each year in educational institutions. With an increasing number of proposed courses and upcoming possibilities to combine degrees, several universities are currently using or testing automated tools designed to perform this task. Examination timetabling is a well known combinatorial optimization problem without standardized approach to solve it, but several methods proposed, as the administrative requirements strongly vary from an institution to another. This paper introduces the examination timetabling problem at École Polytechnique de Louvain-La-Neuve. We present a solver tackling the problem in multiple stages using both linear programming and meta-heuristics. This solver produces schedules respecting the administrative constraints of the faculty while maximizing the students satisfaction. It is meant to be implemented in a tool usable by the administrative support in charge of scheduling. We also explore the mathematical aspects of the problem in order to introduce techniques speeding up the resolution.