Higher-Order Network Approach for Modeling Cascading Failure in Urban Critical Infrastructure

Cascading failure within interdependent urban infrastructure systems represents a critical threat to urban resilience. However, existing models often oversimplify intersystem dependencies of failure propagation. This study introduces a novel higher-order network framework that integrates both node-based and area-based failure modes via simplicial complexes. The proposed model captures multinode dependencies often overlooked in traditional pairwise interaction models. Simulation results demonstrate that area-based failures with >1 km impact radius propagate 2.3× faster and affect 3.1× more nodes than localized failures; higher-order interactions accelerate cascade progression by 40%–60% compared to first-order models; and self-healing mechanisms reduce peak failure count by 35±2% and recovery duration by 22±3%. The model is validated using historical disasters, yielding a high Jaccard similarity of 82%, thereby confirming its robustness.