A Case-Based Reasoning Approach to Formulating University Timetables Using Genetic Algorithms

This paper presents a technique to construct generic university timetables using case-based reasoning (CBR) with genetic algorithms (GAs). The case-based reasoning methodology allows a past memory of timetables to be stored and accessed via retrieval mechanisms, finding a past solution most fitting to the new timetable input problem. In the instance that a past solution is not well suited to the new timetable requirements, a genetic algorithm is employed to adapt the past timetables in the case memory. The hybrid technique used implements a learning mechanism to aid in the revision and adaptation of new timetable solutions. The focus of this technique is a feedback mechanism which allows the system to diagnose the violation of hard-constraints fitted to timetable creation.