A multistage evolutionary algorithm for the timetable problem

It is well known that timetabling problems can be very difficult to solve, especially when dealing with particularly large instances. Finding near-optimal results can prove to he extremely difficult, even when using advanced search methods such as evolutionary algorithms (EA's). This paper presents a method of decomposing larger problems into smaller components, each of which is of a size that the EA can effectively handle. Various experimental results using this method show that not only can the execution time be considerably reduced but also that the presented method can actually improve the quality of the solutions.