Complex underground public spaces (CUPS) enhance urban functions but face high disaster risks. Integrating resilience theory into their planning and use can help mitigate these risks and improve safety in dense urban environments.
This study focused on developing a resilience evaluation indicator system (REIS) to enhance CUPS safety resilience. A comprehensive-objective REIS was established based on the pressure-state-response (PSR) model. The entropy weighting method assigned weights to the indicators based on a fuzzy integrated evaluation model. A Back Propagation (BP) neural network validated the accuracy and reliability of the REIS. Its effectiveness was verified by application to a real-world project.
The REIS rated the real-world case project as “good,” scoring 84.77. This result reflects a high level of safety resilience. The results were consistent with the BP neural network output, with a maximum error of just 0.97% and a minimum error of 0.13%.
This study fills a key gap in CUPS resilience research by shifting from fragmented, static assessments to a more systematic and validated approach. Fragmented approaches often fail to detect latent risks, delay emergency responses and lead to poor coordination in underground environments. These limitations can threaten user safety and disrupt system operations. The proposed REIS addresses these issues by integrating dynamic, multidimensional indicators into a unified framework. The findings provide a reference for future resilience efforts in other complex infrastructure systems.
