Logistics network planning constitutes a strategic decision in the quest to maintain business competitiveness. The present study develops a Mixed Integer Nonlinear Programming model that encompasses the main cost components that should be taken into account in network planning in Brazil, involving transportation, inventories and taxation. The model was tested in the petrochemical company Nova Braskem. Sensitivity analyses were performed in order to gauge the main trade-offs regarding the study's assumptions. It was found that applying the model in the case studied produced savings in taxation and transportation costs and thus in total costs. More precisely, the results indicate that, despite the importance of transportation costs compared to other components, the tax component showed the largest potential financial return to be pursued and in this way it was possible to demonstrate the value associated with the issue within the atmosphere of Brazilian business. For the optimal design of a network, it is confirmed the argument that it is not plausible the dissociation between logistics and tax planning. Moreover, given the nature of the case study, it is feasible to consider the planning of logistic networks as an important tool capable of supporting the search for synergies in mergers and acquisitions. There are, for example, several other recent cases in Brazilian business scenario, as the merger of Sadia and Perdigão and among other agribusinesses. To continue the studies of the trade-offs between logistics and Brazilian tax aspects, we suggest the inclusion of new taxes (PIS, Cofins etc.) as well as the granting of tax benefits in future modeling logistics networks. Specifically with regard to the case study, it is observed that the form of aggregating consumer markets should be enhanced, since the array of customers in the petrochemical industry is highly dispersed. The correlation between the demands could also be analyzed in a future case study. Finally, in order to examine the conclusions reached in this study, it is suggested to apply the model here developed in other case studies.
Logistics network planning; Mixed integer nonlinear programming; Petrochemical industry; Nova Brasken; Case study
