Computerised Multi-Objective Faculty Exam Timetabling Problem

• Main Author: Phang, Min Hui
• Format: Thesis
• Language: English
• Published: Universiti Malaysia Sarawak, (UNIMAS) 2019
• Subjects: QA75 Electronic computers. Computer science
• Online Access: http://ir.unimas.my/id/eprint/27457/3/Phang%20Min%20Hui%20ft.pdf; http://ir.unimas.my/id/eprint/27457/
• Institution: Universiti Malaysia Sarawak

This research focuses on the final examination timetabling at the Faculty of Computer Science and Information Technology (FCSIT), Universiti Malaysia Sarawak (UNIMAS). In UNIMAS, each faculty needs to schedule a final examination timetable every semester. The large number of students and limited resources in each faculty may increase the problem complexity when arranging the final examination timetable. In this study, there is one existing system (FESS 1.0) to generate a clash-free timetable. However, student sectioning, room utilisation, continuous examination gaps and priority courses were not considered. The main objective of this research is to design and build a computationally bounded multi-objective two-stage heuristic algorithm to optimise examination room utilisation. The first stage, course grouping, is mainly to minimise the problem size and clash-free constraints. All the courses are divided into a smaller number of course groups. The course groups are then eased into the second stage of timeslot-room allocation. During the allocation, each course is allocated at ‘best fit room’ in order to maximise the room utilisation and minimise the student sectioning. A few real datasets were collected and experimented with the proposed solution. Overall, the proposed solution is proven to outperform the existing solution in terms of room utilisation, student sectioning, continuous examination and priority course. Besides that, it is also able to accommodate the priority courses constraints well in order to obtain earlier examination dates. Subsequently, a sensitivity analysis was conducted. The increment and decrement in the course size, course-student enrolment size and room size were tested respectively in the solution. All the sensitivity results proved that the proposed solution is effective and robust to solve different types of datasets.