Abordagem heurística para o problema de programação de horários de cursos

Abstract

This work focus on the Curriculum based Course Timetabling. This problem consists in defining the schedule and classroom allocation, respecting a set of constraints, such as classroom capacity, unavailability of teachers, and so on. In order to solve the problem, heuristic methods based on Iterated Local Search (ILS) and Variable Neighborhood Descent (VND) have been developed. A construction procedure was also developed, which is capable of generating good initial solutions with relatively small computational times. Due to the particularities of the problem, since each institution normally considers very different aims and constraints, the validation of this work was done adopting instances from the International Timetabling Competition 2007 (ITC). The results obtained indicate a satisfactory performance for the method that uses VND as local search procedure for ILS. Furthermore, a software was developed to build and manage a timetable.