The Implementation of Genetic Algorithms for Automation of Scheduling Systems at a Course Institution

Abstract

AHE Ngumpul Learning House located in Ngumpul Village, Jogoroto, Jombang, East Java had several problems including: 1) The scheduling system was still conventional, conducted manually by the owner and also a teacher; 2) The learning schedule was also still handwritten, so it could not be used together at the same time and must be copied; 3) Human resources (HR) at the AHE were limited; and 4) The conventional scheduling process was carried out by two people with a time of 1,622 seconds. Based on these problems, this current research proposes the application of a Genetic Algorithm (GA) on a web-based scheduling system with the aim of overcoming the problems in AHE Ngumpul. The data were obtained from the documents of learning scheduling system at AHE Ngumpul. The attributes used for this research were teachers, students, day, and time. In GA, this research used one cut point crossover and reciprocal exchange mutation. The trial showed good results compared to the conventional scheduling with a success rate of 86.5203%, though there were still conflicting schedules. The average time of making a schedule using: 1) GA was 203 seconds, with a fitness value of 0.0232; and 2) GA and an human (only correcting conflicting schedules from GA) was 715 seconds. The difference between conventional scheduling and the implementation of GA combined with humans saved 907 seconds.