A Deep Learning-Based Safety Management Framework for Enhancing Fire Resilience and Addressing Fire Hazards in Karachi: An Analysis of Infrastructure Interdependencies Using IN-Core and MXL

Fire hazards in densely populated urban areas like Karachi pose significant threats to life, property, and critical infrastructure. This study explores the development of a deep learning-based safety management framework aimed at enhancing fire resilience and addressing fire hazards. The framework integrates predictive modeling techniques, including Random Forest and Stacking Regressors, with IN-Core (Interdependency Core) and MXL (Multi-Zone Layers) models to analyze the interdependencies between critical infrastructure systems and simulate fire spread across different zones. While the predictive models demonstrated limited accuracy in forecasting fire size, the IN-Core analysis revealed significant vulnerabilities in Karachi's infrastructure systems, where fire hazards can lead to cascading failures across electricity, water supply, and transportation networks. The MXL simulations identified high-risk zones, such as industrial areas and warehouse districts, where fires are likely to spread rapidly due to environmental and structural factors. The study emphasizes the importance of integrating advanced machine learning techniques with infrastructure resilience planning to improve fire safety management in urban environments. The findings provide valuable insights for policymakers and urban planners to optimize resource allocation and implement proactive fire prevention strategies.