An Investigation of Multistage Approaches to Examination

Many successful approaches to examination timetabling consist of multiple stages, in which a constructive approach is used for finding a good initial solution, and then one or more improvement approaches are employed successively to further enhance the quality of the solution obtained during the previous stage. Moreover, there is a growing number of studies describing the success of approaches which make use of multiple neighbourhood structures. In this study, we investigate the methods of ordering neighbourhood structures within a Variable Neighbourhood Search approach using a great deluge move acceptance method. We also analyse how this approach performs as an improvement algorithm when combined with different initialisation strategies while performing multiple runs for examination timetabling. The empirical results over a well known examination timetabling benchmark show that the performance of Variable Neighbourhood Search great deluge performs reasonably well, ranking second among previously proposed approaches, with the right choice of initialisation and neighbourhood ordering methods. Syariza Abdul Rahman Universiti Utara Malaysia, School of Quantitative Sciences, College of Art and Science, 06010 Sintok, Malaysia E-mail: syariza@uum.edu.my Andrzej Bargiela and Ender Özcan University of Nottingham, School of Computer Science, Jubilee Campus,Nottingham NG8 1BB, UK E-mail: {abb,exo}@cs.nott.ac.uk Edmund Burke University of Stirling, Cottrell Building, Stirling FK9 4LA, UK E-mail: e.k.burke@stir.ac.uk Barry McCollum University of Queen Belfast, School of Electronics, Electrical Engineering and Computer Science, University Road, Belfast, BT7 1NN, Northern Ireland, UK E-mail: b.mccollum@qub.ac.uk