Modelo de otimização para a rede humanitária de suprimentos no Brasil

Abstract

This dissertation addresses the problem of designing a humanitarian supply network. With this purpose, a mixed integer linear programming model is proposed, which aims at maximizing the covering of demand in the affected municipalities. The proposed model considers facilities having two purposes (which operate as deposits and aid centers), and is used to decide on capacity planning upon the basis of the Social Vulnerability Index (SoVI) that prioritizes the care of more vulnerable areas. For the sake of testing the model, a real case based on the Brazilian disaster response planning was considered. Thus, data was gathered from the Integrated System Disaster Information (S2iD); in particular, the 11 types of natural disasters with the highest prevalence during the period 2007-2016 were used for the study. The goal was to estimate the demand for relief aid required to cover population needs in the affected municipalities. Computational experiments were implemented and executed to evaluate the mathematical model. The results suggest that the model proposed may be used as a decision support tool that allows for evaluating: (i) the configuration of the humanitarian supply network, regarding to the deposits and aid centers, as well as their characteristics in terms of capacity and long-term operation, (ii) the dynamics of transportation of relief aid throughout the country to meet the requirements (iii) sensitivity analysis of the budget to the care of victims and the role of SoVI in the level of emergency service; and (iv) the advantages and disadvantages of combining pre-positioning and acquisition of relief aid to take the care of victims.