Using Scatter Search Algorithm in Implementing Examination Timetabling Problem

Abstract

University of Anbar like whatever other universities of higher education, faces comparable issue and is utilized as real case in this study. Setting up the examination timetable takes a significant part of the department head’s role and it includes all the distinctive tools the errand is overpowering. Timetabling is confounded because of various reasons table such as limited time to create schedule, number of halls usually is specific, number of staff work on create timetable is limited and etc. Attempt to find an optimal result for timetable scheduling of Anbar university by using Scatter Search Algorithm (SSA). To achieve an optimal solution for timetable scheduling of Anbar university with the accompanying targets: to create a tool for final examination time tabling for the colleges in university of Anbar and to investigate the efficiency and the performance of the SSA operators including adaptive mutation and other schemes on the SS. The approach has been presented to improve the efficiency and accuracy of scheduling examination timetables and improve the speed of preparing the schedule. The study investigates the most suitable parameters of Scatter Search algorithm for the population based algorithm. The findings outcome from this study have shown that. The results showed the ability to evaluate the performance of SS between the different operators, hence identifying the best results. The adaptive mutation demonstrated its ability as well to provide variety in the generations and as result the SS was able to exploremore variety of solutions and then better solution was found. The optimal settings and configurations found have demonstrated their accurateness and suitability to produce better quality timetable representing very good solution to the problem being addressed. The proposed technique is tested and validated with real case study data and the results are satisfactory. The SS performance is examined in light of the varieties of the chose parameters. The best outcomes are gotten in direct population of 50 tests, 150 generations