A hybrid Grasp and Scatter Search for the exam timetabling problem

In this paper we describe a new hybrid evolutionary algorithm and its implementation for the NP hard exam timetabling problem. This approach uses the good features of the Greedy Randomized Adaptive Search Procedure (GRASP) and the evolutionary Scatter Search (SS) approach which uses strategies for search diversification and intensification to maintain and evolve a population of solutions. New solutions are improved before replacing others according to their quality and diversity. The implementation of the algorithm has been experimented on the popular Carter's benchmarks and compared with the best recent results.