Detalhes bibliográficos
Ano de defesa: |
2024 |
Autor(a) principal: |
Fabian, Matheus Felipe
 |
Orientador(a): |
Rebonatto, Marcelo Trindade
 |
Banca de defesa: |
Não Informado pela instituição |
Tipo de documento: |
Dissertação
|
Tipo de acesso: |
Acesso aberto |
Idioma: |
por |
Instituição de defesa: |
Universidade de Passo Fundo
|
Programa de Pós-Graduação: |
Programa de Pós-Graduação em Computação Aplicada
|
Departamento: |
Instituto de Tecnologia – ITEC
|
País: |
Brasil
|
Palavras-chave em Português: |
|
Área do conhecimento CNPq: |
|
Link de acesso: |
http://tede.upf.br:8080/jspui/handle/tede/2766
|
Resumo: |
Curriculum-Based Course Timetabling is a problem of scheduling subjects, teachers and courses into periods. As there is a wide variety of solutions, implementing a timetable without generating conflicts becomes a laborious task. Optimization algorithms are commonly used to solve this problem, generating viable solutions and making it possible to model the problem according to constraints. This can be used by the institution to adapt the solution to specific needs. Therefore, it is possible to build an algorithm that adapts to the National Assessment System carried out in higher education institutions in Brazil. These evaluations have as one of their factors for calculating the grade of a course the composition of the teaching staff, more specifically, the number of teachers with master's and doctorate degrees and the number of teachers with partial or full-time work. This work presents the modification of the genetic algorithm that generates the UPF timetable, designed to optimize the teaching staff's grade according to the assessment carried out by the National Higher Education Assessment System. By changing the calculation of the evaluation function and adding new constraints with weights defined through preliminary tests, it was possible to direct the solutions to allocate a greater number of teachers who meet the evaluated inputs. The optimization was applied to the 18 courses that participated in Enade in 2023. The results were positive, with 17 courses achieving a statistically significant increase in the allocation of teachers in at least one of the three optimized inputs, 16 courses having an increase in the allocation of doctors and teachers with the desired work regime with a statistically significant difference. significant in relation to the original algorithm, and three courses increased the allocation of masters in a statistically significant way, with 11 courses already having 100% of masters allocated in the original algorithm. It is also worth highlighting that two courses managed to fill 100% the number of allocated teachers who met all the requirements imposed by the assessment. |