Meta-heurística aplicada ao planejamento da designação de maquinistas a trens de viagem de longa distância em ferrovias de carga

Abstract

Among all the modes of transporting goods and people, one of the most important in the whole world is the rail transport. Due to its high capacity to transport goods, in Brazil, its use has been increasing in recent years. For the transportation of these goods, the trains require qualified professionals to be operated, these professionals are the drivers. Due to the large lengths of the Brazilian railways, the travel times between origin and destination, in most cases, are longer than the time of the drivers' work schedule, which is defined by labor laws. In this way, it’s necessary, to meet all labor requirements, that the drivers are changed during trips. These exchanges occur in sections of the railway where there are detachments. Nowadays, the planning of the designation of train drivers on Vitória Minas Railway, which is the focused railway on this work, is still done manually without any aid of any type of software, which makes the process slow and ineffective. Thus, there is currently a need to create tools that can assist professionals in this task of assigning drivers to freight trains. This dissertation proposes the development of a meta-heuristic based on the Problem of Routing Vehicles with Multiple Deposits, Multiple Trips, for the designation of train drivers to trains according to the train circulation planning, aiming to minimize the costs with payment of overtime and minimize the number of drivers needed. This meta-heuristic stands out due to the fact that there is only one paper found that addresses this subject in the literature. However, the mathematical model proposed in this paper had very long execution times, making it practically impossible to be used. The meta-heuristic will be written in C language and based on the Simulated Annealing meta-heuristic.