Modeling and Solution of a Complex University Course Timetabling Problem

The modeling and solution approaches being used to automate construction of course timetables at a large university are discussed. A course structure model is presented that allows this complex real-world problem to be described using a classical formulation. The problem is then tackled utilizing a course timetabling solver model that transforms it into a constraint satisfaction and optimization problem. The tiered structure of this approach provides flexibility that is helpful in solving the multiple subproblems that arise from decomposition of the university-wide problem. A production system has been partially implemented and results of early use are presented. Practical issues raised during the implementation of the automated timetabling system are also discussed.