Solving the classroom assignment problem with simulated annealing

Classroom assignment at ITESM, a Mexican university, was automated by developing a system capable of assigning the most adequate classroom to each class section. The problem is formulated as an optimization one and solved by using simulated annealing. ITESM offers more than 3,000 different sections with three or more meeting hours a week, almost 200 meeting times, and 190 classrooms, which make the problem a very large and complex one. To solve it, the problem was divided by solving the optimization problem for each meeting time. The optimization problem considers the capacity of the classroom as well the equipment it has in order to have a classroom assigned to a section of lower or equal number of students. It also considers the distance of the classroom to the professor´s office who will be the instructor of the section assigned to it (several floor buildings will be filled from bottom to top). Our simulated annealing based optimization system reaches a valid and optimal solution (in the sense of classroom capacity, distance, and equipment) in a lower time than considering the total number of sections and the previous way of doing it: manual