Thesis (BS Applied Mathematics) -- University of the Philippines Mindanao, 2006

Course timetabling problem is a mapping of the set of lecture courses to the set of periods and rooms subject to constraints. Genetic algorithm, a biologically inspired process based upon the analogy of natural selection and population genetics, is often used as a search and optimization algorithm in the field of timetabling. Although the genetic algorithm was found to be promising in timetabling courses of the College of Science and Mathematics-University of the Philippines Mindanao, it is still not known which kind of crossover and mutation operators are effective in producing better timetables. Thus, in this study, combinations of uniform, sector-based, and conflict-based crossover, with swap, swap/random, violation-directed mutation were inserted in the general genetic algorithm. Results showed that the combination of conflict-based crossover and violation-directed mutation gave the best performance in finding good solutions, followed by uniform crossover and violation-directed mutation combination and then by the sector-based crossover and violation-directed mutation combination. Results also showed that the six combination with swap and swap-random mutation, were performed badly, were not significantly different form each other. These results were confirmed further statistical analysis. Although the present results obtained indicate the effectiveness of the violation-directed mutation combined with conflict-based crossover, further studied are still needed to explore the potential of other crossover and mutation process.