Constraint Programming Approach for a University Timetabling Decision Support System with Hard and Soft Constraints

This paper develops a university timetabling decision support system that considers both hard and soft constraints of the problem. A timetable solution must satisfy all hard constraints, and may be only capable of partially meeting soft constraints. We modeled the problem as a Constraint Satisfactory Problem and adapted lexicographic optimization approach to implement the solution procedure; where each soft constraint is treated as an objective with a priority. The solution approach utilizes the solution space reduction nature of constraint propagation to optimize objectives sequentially. Different value assignment strategies for constraint propagation are investigated to explore their robustness and effectiveness in performances. The system allows department management to indicate different combinations of preferences and parameters, and view the resulting timetable and related statistics in real time mode. The resulting timetabling contributes to a better teaching environment for both faculty and students.