Estrutura para apoiar a decisão de alocação de recursos integrando critérios de vulnerabilidade em comunidades carentes

Abstract

This dissertation addresses the challenge of allocating resources to underserved communities, recognizing the structural complexity that characterizes these territories, marked by interdependent social, economic, and environmental inequalities. Given this reality, the central objective of the study was to propose a framework to support resource allocation decisions that integrates vulnerability criteria in underserved communities. The methodological approach involved identifying and selecting indicators through a systematic literature review and the use of secondary databases, culminating in the construction of a Composite Vulnerability Index (CVI). These indicators were weighted using the Swing Weighting (SW) method, enabling their aggregation into a single metric. The data were then spatialized in a Geographic Information Systems (GIS) environment, used as a comparative visualization tool across the analyzed regions. Subsequently, the IVC was incorporated as a criterion in the multi-criteria decision model, developed using the Multi-Attribute Value Theory (MAVT) approach and operationalized in the VISA software, which also integrated intervention costs and preferences for public resource categories. These preferences were defined based on an interview with a decision-maker responsible for Social Assistance Reference Centers (CRAS). Based on this, a case study was conducted in the municipality of Sorocaba, São Paulo. The results demonstrate that the proposed framework is capable of synthesizing multiple dimensions of vulnerability and supporting decision-making based on transparent and territorially informed criteria. Although the IVC identified regions such as Ipiranga, Vila Helena, and Laranjeiras as the most vulnerable territories, the decision-making stage with MCDA prioritized Laranjeiras, Ipiranga, and João Romão, demonstrating partial convergence between the technical indicators and the final prioritization of the areas. Limitations include the difficulty of standardizing indicators in per capita terms and restrictions associated with terminology to characterize underserved communities, in addition to the limited perspective of a single decision-maker. These limitations point to avenues for future research, including the use of more granular primary data and increased community participation in defining criteria and weights.