Modelos de programación entera para el problema de aignación de horarios en cursos universitarios

• Published in: Revista de métodos cuantitativos para la economía y la empresa Vol. 36; pp. 1 - 20
• Main Authors: Arango Londoño, Carlos Andrés, Jiménez, Heriberto Felizzola, Hualpa Zuñiga, Andres Mauricio, Gómez Martínez, Paula Zuranny, Mora Garzón, Carlos Felipe
• Format: Journal Article
• Language: Spanish
• Published: Sevilla Universidad Pablo de Olavide 2023
• Subjects: classroom assignment; integer programming; Timetabling
• ISSN: 1886-516X; 1886-516X
• DOI: 10.46661/rev.metodoscuant.econ.empresa.7354

Timetabling scheduling is classified as a combinatorial problem for which there are multiple solution alternatives, including the integer programming. However, the modeling of operational rules and the size of real problems makes its use not common compared to other techniques. This article proposes an integer programming model (IP) model, which addresses the scheduling problem made up of variables associated with time slots, subject and assigned teacher. Parameters such as the number of minimum and maximum time slots to be taught by the teacher, time slot time, teacher availability, available classrooms and estimated cost of dissatisfaction generated by the assigned schedule are also included. Seven hard and seventeen soft constraints are integrated into the model, which provide higher quality to the final schedule solution. The IP model is validated with a global objective function, in which experiments, and results obtained in real instances of the Universidad de la Salle (ULS) are reported. The new solution approach offers improvements in the final schedules, as well as the interaction with the users during its construction. Finally, in the conclusions of the work, the design and development of a system that provides support for decisions is discussed, referencing suggestions for future developments.