Otimização da localização de ecopontos em cidades brasileiras : uma abordagem baseada em análise geoespacial e modelagem preditiva

• Ano de defesa: 2024
• Autor(a) principal: Marques, Leonardo da Cunha
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: por
• Palavras-chave em Português: Gestão de resíduos sólidos; Regressão linear; Ecopontos; Sistema de informação geográfica - SIG; Modelagem preditiva; Solid waste management; Linear regression; Ecopoints; Geographic Information System - GIS; Predictive modeling; CNPQ::ENGENHARIAS::ENGENHARIA CIVIL
• Link de acesso: http://repositorio.ufc.br/handle/riufc/78612

In Brazil, the management of urban solid waste is a growing challenge due to population increase and urban expansion. One of the strategies used by the public authorities to reduce irregular waste disposal is the installation of ecopoints, which are public facilities designated to receive various types of voluntarily discarded waste. However, there is no methodology for choosing their locations. This research proposes the development of a model that identifies optimized areas for the installation of ecopoints in Brazilian cities using geospatial analysis and predictive modeling. Data on the weight collected per year at each ecopoint in the cities of Belo Horizonte, Fortaleza, and Guarulhos were analyzed, cross-referencing them with socioeconomic, environmental, and infrastructural variables. The results indicate that population density is the variable with the greatest influence on the efficiency of the ecopoints. Two models were developed and evaluated, with the first based on deterministic analysis and the second on linear regression. The deterministic model used the variables of population density, proximity to main roads, and Municipal Human Development Index (MHDI) and was found to be suitable for classifying areas with scores from 0 to 10 for the suitability of installing ecopoints, based on contour scenarios that can be defined by the manager, allowing local characteristics of the city in question to be incorporated into the decision-making process. The linear regression model obtained a determination coefficient of 58% and used the variables of population density, slope, road density, and health unit density in its final formula to estimate the weight to be collected in a given area. The developed models proved to be effective as decision-support tools for locating ecopoints, contributing to more efficient and sustainable management of urban solid waste.