Authors:
Nuno Leite
1
;
Rui Neves
2
;
Nuno Horta
2
;
Fernando Melício
1
and
Agostinho Rosa
2
Affiliations:
1
ISEL - Lisbon Polytechnic Institute, Portugal
;
2
TU-Lisbon, Portugal
Keyword(s):
Exam Timetabling Problem, Evolutionary Algorithms, Multi-objective Optimization, Combinatorial Problems.
Related
Ontology
Subjects/Areas/Topics:
Artificial Intelligence
;
Computational Intelligence
;
Evolutionary Computing
;
Evolutionary Multiobjective Optimization
;
Hybrid Systems
;
Memetic Algorithms
;
Soft Computing
Abstract:
This paper describes the construction of an university examination timetable using a hybrid multi-objective evolutionary algorithm. The problem instance that is considered is the timetable of the Electrical, Telecommunications and Computer Department at the Lisbon Polytechnic Institute, which comprises three bachelor degree programs and two master degree programs, having about 80 courses offered and 1200 students enrolled. The task of manually construct the exam timetable for this instance is a complex one due essentially to the high number of combined degree courses. This manual process takes, considering a two-person team, about one week long. A hybrid multi-objective evolutionary algorithm, based on the Non-dominated Sorting Genetic Algorithm-II (NSGA-II), is proposed for solving this problem instance, incorporating two distinct objectives: one concerning the minimization of the number of occurrences of students having to take exams in consecutive days, and a second one concerning
the minimization of the timetable length. The computational results show that the automatic algorithm achieves better results compared to the manual solution, and in negligible time.
(More)