Optimizing Regression on Genetic Algorithms for Effective scheduling based on the Predicted Number of Students

Scheduling courses is a routine job that every educational institution always carries out at the beginning of the semester. Scheduling is a complicated task because the scheduling problem is a combinatorial problem with limitations that must be met. These limits are divided into absolute limits that must be met, such as the availability of classrooms and their capacity, the availability of lecturers and students, and soft limitations, such as lecturers’ preference in teaching time. The best solution to the scheduling problem is to-predict the number of course takers in the next semester within a certain period. This research solves this problem by predicting the number of applicants for the next semester’s courses. This study uses a genetic algorithm to optimize the data and parameters obtained from the data of enthusiasts of previous courses. After optimizing the data, regression is carried out to predict the number of enthusiasts for the following semester’s courses. The results of this test show that MSE was 0.445138347 in 2020, and the error rate based on all accurate and predictive data is 14.34%.