A modified genetic method for automatic scheduling

Abstract

The problem of creating an optimal schedule is considered, which consists in finding the optimal distribution of educational classes for a certain period under given restrictions. Sequential and parallel scheduling methods based on genetic search have been developed. The proposed methods use adapted and modified initialization, crossover, and selection operators. Algorithms, using the objective function, minimize conflicts between classes and the time interval between classes, taking into account the recommended time and venue. The developed methods allow you to speed up the time for planning the educational process and avoid mistakes when creating a schedule. A comparative analysis was conducted between the classical and modified genetic algorithms, and it was found that the modified algorithm works faster and more efficiently than the classical one. The performance of the modified algorithm was also compared with different genetic algorithm operators and parameters to determine the best ones. The obtained results allow us to propose effective methods for improving the quality of scheduling and improving the learning process at the university.