Priority-driven Constraints Used for Scheduling at Universi- ties

Introduction At the University of Siegen like at most other universities scheduling is a complex problem. Only a few individual wishes of the lecturers as well as of the students can be considered in the planning phase, so dissatisfaction cannot be avoided. Due to the large number of courses (> 3,000) and the increasing number of students registered (> 12,500), lecture halls and seminar rooms are scarce resources (< 90). A computerised solution for this NP-hard problem has failed because of insufficient search algorithms. As a matter of fact, there is no commercial software available satisfying the comprehensive requirements of university scheduling. As a project group made up of 6 advanced students of computer science, we are attempting to solve the problem with the class library of ILOG Solver [1 ] (abbreviated to "Solver" in the following). During the whole software development process we are using the object-oriented method of Grady Booch [2].