European Journal of Operational Research vol:195 pages:307-318
In this paper we present a decomposed meta-heuristic approach to solve real-world university course timetabling problems. Essential in this problem are the overlapping time slots and the irregular weekly timetables. A first stage in the approach reduces the number of ubjects through the introduction of new structures that we call ‘pillars’. The next stages involve solving the constraints one by one, instead of trying to find a solution for all the constraints at once. Test results for a real-world instance are presented.