Uma ferramenta flexível para a geração automática de horários de cursos de graduação baseada em algoritmos genéticos

Abstract

Elaborating class schedules for undergraduate courses in a complex task, because it involves many restrictions that affects logistics of teachers, students and classrooms. Some schedule solutions are not viable, for example when different classes of the same teacher are allocated at the same time, while other solutions are viable but have undesirable class times. The amount of possible class schedules is generally very large, which makes it a very difficult job to perform manually with not very clear criteria. Therefore, this work presents a tool to automatically generate class schedules for undergraduate courses based on genetic algorithms. Four criteria have already being developed to find the best solution, but it is also flexible to allow the inclusion of new criteria and customize the weights of each criteria according to the needs of each course. The evaluation was performed by generating class schedules for the Information Systems and Computer Science courses of the Federal University of Paraıba, campus IV, which rapidly converged to a viable and optimal solution.