In this article we propose a hybrid simulation-optimization model to analyze alternative operational policies for the rail transport of grains in a closed loop system aiming to increase its overall capacity. The core of the proposed model is a discrete event simulation system developed in "Visual Basic for Applications" (VBA) that allows more flexible rules to be analyzed for the main decisions regarding queues. In our model ten different decisions related to rail operations (mainly rail asset allocation and queue management) are replaced by prioritization rules that aim to more closely reflect how decisions are made in practice. We also propose an optimization approach based on Hooke & Jeeves algorithm in order to determine the best subset of prioritization rules. Our model was applied to a real-world problem related to a grain transportation system composed of Centro-Atlântica and Vitória a Minas Railways and the Port of Tubarão. The results show that these prioritization rules can effectively yield to increased overall system capacity when compared to simpler decisions allowed by a discrete event simulation model.
simulation-optimization; rail transport; agriculture grains; heuristics