University Class Schedule Assignment by a Tabu Search Algorithm

Historically, the University class scheduling problem has been difficult to solve efficiently, due to the long time it takes to find satisfactory solutions. This work seeks to automate the process that solves the following problem: assign to each mathematics class of an academic period of the Universidad Tecnológica de Panamá (Task) a professor from the Department of Exact Sciences (Resource) of said institution. In an academic period, the Department of Exact Sciences has approximately 205 subjects and 71 professors. Due to the huge space of possible solutions, a local search has been chosen, in which iteratively tries to improve the solution through slight changes. The effective implementation of this search requires solving two great challenges. First, define an objective function that allows considering the most important qualities of an assignment, and at the same time control its computational complexity, allowing multiple iterations to be carried out in a reasonable time. Second, toast to the search for a structure. The problem will start from a solution obtained by means of a greedy algorithm, where for each subject, a teacher is assigned so that the lowest value of the objective function is recorded. Subsequently, the search is provided with a Tabu Search metaheuristic that allows it to escape local optima and better control its path. The results obtained show evidence that the metaheuristic works, providing a fast feasible solution that includes the desired characteristics and that serve as a starting point for the career coordinator.