Otimização da operação dos veículos de empresas do transporte público de Belo Horizonte.

This work tackles the mass transit vehicle scheduling problem through network flow approach. This problem is usually modeled as a quasi-assignment problem and solved with specific algorithms. However, in most real cases, the underlying network is of such size that the computational time required to solve the problem becomes prohibitive. In order to overcome this problem, it was applied the Arcgen methodology, which represents the problem as a flow circulation model and combines an arc generation technique with the out-of-kilter algorithm. So, is possible to decrease the number of arcs considered in the optimization process. This methodology was applied to different real cases of different sizes, operating in Belo Horizonte city. The results obtained to each bus company studied pointed out to a) possible reduction in the operational costs incurred by bus conduction and b) the practical applicability of the theoretic solutions, when compared with the current solutions adopted by the companies.