Please use this identifier to cite or link to this item:
https://ptsldigital.ukm.my/jspui/handle/123456789/513452| Title: | Enhanced Tabu Search approaches for university examination timetabling problems |
| Authors: | Ariff Md. Ab. Malik (P37888) |
| Supervisor: | Masri Ayob, Assoc. Prof. Dr. |
| Keywords: | University examination Examination timetable Tabu search Artificial intelligence Universiti Kebangsaan Malaysia -- Dissertations Dissertations, Academic -- Malaysia |
| Issue Date: | 24-Sep-2013 |
| Description: | Tabu Search (TS) is one of the local search techniques that has been studied and applied in solving and optimizing the timetabling problems. Based on the current reviews of literature, the utilization of TS is becoming less popular (perhaps) due to inability to produce better quality solutions and its search exhaustiveness (i.e. very time consuming). Therefore, the ultimate goal of this thesis is to enhance the Tabu Search approach by applying it to the university examination timetabling problem with some modifications and strategies. Several meta-heuristics algorithms and population's operators based methods were hybridized to the proposed method with the intention to investigate the effectiveness of this hybridization. The research work begins by applying basic TS method with a new separation concept of neighbourhoodtabu list- search strategies and tested on the Carter datasets. Due to some potential initial results of the overall of Integrated Two-stage Multi-neighbourhood Tabu Search technique, another two new strategies, adaptive switcher mechanism and stratified descent random selection, have been applied in enhancing the technique's exploration performance. On further testing, two outstanding local search methods, Exponential Monte Carlo with counter (EMCQ) and Great Deluge Algorithms (GDA), have been hybridized in order to improve the technique's diversification strategy. The hybridization of several population operators have also been conducted, such as population of solutions and reproduction procedures on the EITMTS, EITMTSEMCQ and EITMTS-GDA. The results of this enhanced ITMTS (EITMTS) show several best results against other state-of-the-art methods' results. Based on the results of the local search based technique, it was found only EITMTS-EMCQ has outperformed some other methods that were reported in the literature (on some problem instances). Meanwhile the hybrid evolutionary-EITMTS has shown better overall performance among these hybridized and other state-of-the-art evolutionary methods. As overall, the EITMTS-EMCQ technique has produced as a technique with an outstanding results' achievement among all ITMTS variation methods.,Ph.D. |
| Pages: | 236 |
| Call Number: | T57.6.A736 2013 3 tesis |
| Publisher: | UKM, Bangi |
| Appears in Collections: | Faculty of Information Science and Technology / Fakulti Teknologi dan Sains Maklumat |
Files in This Item:
| File | Description | Size | Format | |
|---|---|---|---|---|
| ukmvital_71614+Source01+Source010.PDF Restricted Access | 4.48 MB | Adobe PDF |  View/Open |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.