ALGORITHMIC-SOFTWARE METHOD FOR CALCULATING OPTIMAL SHELTER PLACEMENT USING API AND MAPPING DATA

Abstract

In contemporary Ukrainian cities, shelter placement is primarily managed through static, GIS-based software that provides informational maps without real-time optimization or predictive analytics. These tools help residents locate shelters but do not dynamically address coverage gaps, population density shifts, or transportation accessibility. Lacking advanced mathematical models, existing solutions rely on predefined criteria and historical planning rather than data- driven optimization, resulting in potential disparities in shelter accessibility, particularly in high-density or mobility- limited areas. Unlike some international solutions that integrate GIS modeling and transportation analysis, Ukrainian software does not fully adapt to rapid urbanization and infrastructural changes, highlighting the need for a more advanced, real-time optimization approach to improve emergency preparedness. The article focuses on developing an optimized approach for shelter placement in urban environments using computational algorithms and geospatial data. The proposed algorithmic-software method integrates demographic statistics and spatial analysis to determine the most effective shelter locations. In large Ukrainian cities like Kyiv, where population density is high, the findings suggest that large-capacity shelters should be placed near major transportation hubs. In suburban areas surrounding cities like Lviv and Odesa, shelters should be positioned within a 5 km radius of densely populated regions to ensure accessibility via road and rail networks. For rural areas, a more evenly distributed network of smaller shelters is recommended to maximize coverage across larger territories.The research utilizes an official bomb shelter map of Kyiv to develop and implement the Huff model in Python, accounting for existing population distribution and shelter locations. The core steps of the methodology include computing a weighted central point for dense population clusters, generating potential shelter locations, interactive GIS based visualization, and refining placements based on real-world constraints. The proposed method integrates with GIS platforms such as QGIS and ArcGIS, enabling real-time decision-making for optimized shelter distribution.