A Case Study of Model-Driven Engineering for Automated Timetabling

Large educational institutions such as universities need to efficiently allocate their staff and students into teaching and examination timetables, while respecting hard constraints and taking into account preferences. The timetables produced automatically by commercial products are generally unsatisfactory, and considerable manual modification is required before they can be used. The authors have designed and implemented an algorithm which produces good results, but the next challenge is to integrate it into real universities. This involves synchronizing with existing information systems, as well as presenting a user interface that can be used by timetabling staff to view and manipulate the generated timetables and to configure the algorithm. Given the limited number of developers available, the high productivity offered by model-driven engineering was deemed necessary. In this paper, the authors show how they have combined several transformation tools, persistence frameworks and model-driven UI approaches to deliver a first version of an integrated solution for automated timetabling. The authors identify areas of future work and places where the current state of the art could be further developed.