Solving the Invigilation Scheduling Problem in Examination Timetabling System Using Tabu Search Technique

Liong, Jee Yuen (2019) Solving the Invigilation Scheduling Problem in Examination Timetabling System Using Tabu Search Technique. Final Year Project (Bachelor), Tunku Abdul Rahman University College.

Text Liong Jee Yuen_FULL TEXT.pdf Restricted to Registered users only Download (741kB)

Abstract

In this study, the invigilation duty scheduling problem in the examination timetabling system of Tunku Abdul Rahman University College (TAR UC) will be solved by using the Tabu Search technique. The purpose of conducting the project is to investigate the linear relationship between the size of Tabu List and the performance of Tabu Search and thus determining the most appropriate choice of the size of Tabu List during the implementation of the algorithm. Besides that, the project also aims to investigate the performance of Tabu Search in solving the invigilation scheduling problem. Furthermore, a comparison between the performance of Tabu Search and the current algorithm that was implemented in the examination timetabling system of TAR UC will be carried out in order to determine the suitability of both of the algorithms in solving this problem. In order to achieve these objectives, the implementation of Tabu Search with different size of Tabu List in the invigilation duty scheduling problem will be done and followed by a series of tests to be conducted. At the same time, tests will also be conducted on the invigilation module in the examination timetabling system of TAR UC and the results will be compared to the results from Tabu Search algorithm. Based on the results, the performance of Tabu Search is not linearly related to the size of Tabu List and therefore, the implementation of Tabu Search with Tabu List’s size of 20 has been chosen for further analysis. From the study, Tabu Search can be considered as an efficient algorithm in solving the invigilation duty scheduling problem but lack of consistency in maintaining the high efficiency all the time. However, as compared to the current algorithm, Tabu Search is able to outperform the current algorithm in a great extent in terms of efficiency as well as consistency in solving the problem. Therefore, this study has concluded that Tabu Search has a great potential in solving the invigilation duty scheduling problem as it has a high efficiency in solving the problem despite of the low consistency in maintaining the performance.