José Joaquim Moreira[1]
(recebido em 14 de Abril de 2008; aceite em 5 de Maio de 2008)
Resumo. A complexidade do problema dos calendários de exames justifica-se pela dimensão do escalonamento de exames e pelo elevado número de restrições e critérios de atribuição. Este artigo apresenta um método de solução para o problema da construção automática dos calendários de exames. Entre vários modelos matemáticos de representação, a opção final foi por um modelo matricial, que é justificado pelas vantagens que este modelo apresenta quando utilizado no algoritmo de solução. O método de solução consiste numa meta-heurística que inclui um algoritmo genético. O modelo é direcionado para a construção de calendários de exames em instituições de ensino superior. Os resultados alcançados em reais e complexos cenários são satisfatórios; os calendários de exames cumprem as regulamentações impostas. Concluímos que, quando o algoritmo não determina uma solução, sem penalização, é porque essa solução não existe.
Palavras-chave: Escalonamento, Problemas de horários, Calendários de exames, Algoritmos genéticos.
]]> Abstract. The complexity the of exam timetables problem is justified by the scheduling size of the examinations and the high number of constraints and criteria for allocation. This paper presents a solution method to the problem of automatic construction timetables for the exams. Among several mathematical models of representation, the final option was for a model matrix, which is justified by the benefits that this model presents when used in the algorithm solution. The method of solution is a meta-heuristics that includes a genetic algorithm. The model is directed to the construction of exam timetables in institutions of higher education. The results achieved in real and complex scenarios are satisfactory; the exam timetabling meets the imposed regulations. We conclude that when the algorithm does not determine a solution with no penalty, is because that solution does not exist.
Keywords: Scheduling, Timetabling problems, Exam timetabling, Genetic algorithms.
Texto completo disponível apenas em PDF.
Full text only available in PDF format.
References
Chan, H. W (1997). School Timetabling Using Genetic Search. 2th International Conference on the Practice and Theory of Automated Timetabling, PATAT’97.
]]> Eley, M. (2006). Ant Algorithms for the Exam Timetabling Problem. 6th International Conference on the Practice and Theory of Automated Timetabling, PATAT’06.Fang, H. L. (1994). Genetic Algorithms in Tametabling Problems. PhD Thesis, University of Edinburgh. [ Links ]
Fernandes, C. (2002): Infected Genes Evolutionary Algorithm for School Timetabling. WSES International Conference.
Gröbner, M., Wilke P (2002). A General View on Timetabling Problems. 4th International Conference on the Practice and Theory of Automated Timetabling - PATAT’2002.
Holland, John (1975). Scheduling, Adaptation in Natural and Artificial Systems. The University of Michigan Press.
ITC (2002). International timetabling competition: http://www.cs.qub.ac.uk/itc2007/index_files/overview.htm
Oliveira, E., Reis L.P. (2000). A Language for Specifying Complete Timetabling Problems. 3th International Conference on the Practice and Theory of Automated Timetabling PATAT’2000.
PATAT (1995). The International Series of Conferences on the Practice and Theory of Automated Timetabling (PATAT) - http://www.asap.cs.nott.ac.uk/patat/patat-index.shtml
Queirós, F. H. (1995). Construção automática de Horários de Aulas. Tese de Mestrado, Universidade Portucalense.
Thompson et al, Jonathan e Dowland, Kathryn (1995). General Cooling Schedules for a Simulated Annealing Based Timetabling System. Proceedings of the 1st International Conference on the Practice and Theory of Automated Timetabling.
]]> Wren, Anthony (1995). Scheduling, Timetabling and Rostering — a special relationship? Proceedings of the 1st International Conference on the Practice and Theory of Automated Timetabling, 474-495.
[1] ISLA - Rua de Cabo Borges, 55, 4430-032 V. N. de Gaia, Portugal.
]]>