Adaptive Genetic Algorithm for High School Time-Table

Abstract

Time-table scheduling in senior high school is complex combinatorial problem. The scheduling process needs much time and susceptible to human error. This human error can lead to the violation of hard constraints. Beside the hard constraints, the good schedule must minimize the violation of soft constraint to ensure the students' and teachers' preference can be satisfied. The aim of this study was to build the automated scheduling process using adaptive genetic algorithm (AGA). The AGA implements in this study to flexible the probability of mutation (pm) based on the fitness. The minimum pm was set to 0.01, while the maximum pm was 0.2. The experiment shows that the AGA give the better fitness compared to the original GA. The best fitness reached by AGA was 0.054 with the average of maximum generation was 402. The original GA with fixed pm 0.1 resulted best fitness 0.045 and the average generation was 387.