Aplicação de um algoritmo genético ao problema de rodízio de tripulações do sistema de transporte público urbano

This paper addresses the resolution of the Crew Rostering Problem (CRP). The problem consists of assigning duties to the crew members of a company over a given planning horizon, in order to minimize its total costs. The number of crews required to perform all journeys is considered as the fixed costs, while the accumulated overtime hours and idle hours for each crew are the variables costs. In the resolution of this problem, it must be considered the labor laws and the operational constraints of each company. In this paper, we solved the CRP in two stages. In the first of them, we defined the rest period, minimizing the total number of crews. In the second stage, we allocated the duties to be performed by crews, minimizing idle and overtime hours. Both stages were solved using a Genetic Algorithm, a novel CRP approach to Brazilian cases. The algorithm was designed to solve a real case from a company and its results were compared with the exact solutions obtained by an Integer Programming Model, indicating to be competitive