TY - GEN
T1 - The effect of neighborhood structures on examination timetabling with artificial bee colony
AU - Bolaji, Asaju La aro
AU - Khader, Ahamad Tajudin
AU - Al-Betar, Mohammed Azmi
AU - Awadallah, Mohammed A.
AU - Thomas, J. Joshua
N1 - Publisher Copyright:
© PATAT 2012.
PY - 2012
Y1 - 2012
N2 - Artificial Bee Colony (ABC) algorithm is among the most effective nature-inspired algorithms for solving the combinatorial optimization problems. In this paper, ABC is adopted for university examination timetabling problems (UETP) using a defacto dataset established by Carter et al. (1996). ABC has three main operators that drive the search toward the global minima: employed bee, onlooker bee, and scout. For UETP, the employed bee and onlooker bee operators are manipulated to be workable where three neighborhood structures are employed: move, swap and Kempe chain. The effect of these neighborhood structures on the behaviour of ABC for UETP is studied and analyzed in this paper. The experimental design is intentionally made with various convergence cases of different neighborhood structure. The result suggests that the ABC combined with the three neighborhood structures is an effective method for UETP. Comparative evaluation with previous methods is also provided. The results produced by the proposed method are competitive in comparison with state of the art methods. Theoretically, this study contributes to the examination timetabling community through an ABC template which is both efficient and flexible for UETP.
AB - Artificial Bee Colony (ABC) algorithm is among the most effective nature-inspired algorithms for solving the combinatorial optimization problems. In this paper, ABC is adopted for university examination timetabling problems (UETP) using a defacto dataset established by Carter et al. (1996). ABC has three main operators that drive the search toward the global minima: employed bee, onlooker bee, and scout. For UETP, the employed bee and onlooker bee operators are manipulated to be workable where three neighborhood structures are employed: move, swap and Kempe chain. The effect of these neighborhood structures on the behaviour of ABC for UETP is studied and analyzed in this paper. The experimental design is intentionally made with various convergence cases of different neighborhood structure. The result suggests that the ABC combined with the three neighborhood structures is an effective method for UETP. Comparative evaluation with previous methods is also provided. The results produced by the proposed method are competitive in comparison with state of the art methods. Theoretically, this study contributes to the examination timetabling community through an ABC template which is both efficient and flexible for UETP.
KW - Artificial bee colony
KW - Examination timetabling problem
KW - Nature-inspired algorithm
UR - https://www.scopus.com/pages/publications/85026736779
M3 - Conference contribution
AN - SCOPUS:85026736779
T3 - PATAT 2012 - Proceedings of the 9th International Conference on the Practice and Theory of Automated Timetabling
SP - 131
EP - 144
BT - PATAT 2012 - Proceedings of the 9th International Conference on the Practice and Theory of Automated Timetabling
A2 - Kjenstad, Dag
A2 - Riise, Atle
A2 - Nordlander, Tomas Eric
A2 - McCollum, Barry
A2 - Burke, Edmund
PB - SINTEF
T2 - 9th International Conference on the Practice and Theory of Automated Timetabling, PATAT 2012
Y2 - 29 August 2012 through 31 August 2012
ER -