A Multi-Phase Hybrid Metaheuristics Approach for the Exam Timetabling

We propose a Multi-Phase Hybrid Metaheuristics approach for solving the Exam Timetabling Problem (ETP). This approach includes a pre-processing phase, a construction phase and an enhancement phase. The pre-processing phase involves two stages: the propagation of ordering constraints and implicit constraints discovery stages. The construction phase uses a variant of the Tabu search with conflicts dictionary. The enhancement phase includes Hill Climbing (HC), Simulated Annealing (SA) and our updated version of the extended ”Great Deluge” algorithm. In order to evaluate the performance of the different phases of our proposed approach, we conducted several experiments on instances taken from the ITC 2007 benchmarking datasets. The results are very promising and competitive with the well known ETP solvers.