Resilient Healthcare and Critical Urban Infrastructure Design Using AI-Driven Engineering and Project Management Systems

Abstract

Ensuring resilient healthcare accessibility under infrastructure disruptions remains a critical challenge for rapidly urbanizing environments characterized by complex and interdependent systems. This study proposes and evaluates an AI-driven engineering and project management framework designed to enhance the resilience of healthcare delivery and critical urban infrastructure under diverse disruption scenarios. The framework integrates machine learning–based infrastructure criticality prediction, network-level resilience analysis, healthcare accessibility modeling, spatial impact assessment, and AI-enabled project management optimization within a unified decision-support system. Results show that ensemble-based machine learning models achieve consistently high predictive performance, with accuracy exceeding 98% and area-under-the-curve values approaching 0.99, enabling reliable identification of infrastructure components whose failure disproportionately affects healthcare accessibility. Network resilience analysis demonstrates that AI-guided strategies significantly outperform random and clustered failure scenarios by preserving connectivity and limiting efficiency loss. Connectivity degradation under clustered failures is four times higher than under AI-guided strategies, while AI-informed prioritization reduces connectivity loss by approximately 75%. Healthcare accessibility outcomes further confirm the effectiveness of the proposed framework. AI-guided strategies consistently limit travel-time increases and preserve population coverage within acceptable emergency thresholds, maintaining coverage above 99.7% even under severe disruptions. In contrast, clustered failures produce abrupt performance collapses and localized isolation. Spatial analyses reveal that accessibility degradation and failure impacts are highly heterogeneous, occurring in localized pockets rather than uniformly across the urban network. Importantly, AI-guided strategies distribute impacts more evenly, preventing cascading spatial failures and mitigating inequitable access outcomes. From a project management perspective, AI-driven risk prediction and reinforcement learning–based optimization significantly reduce schedule delays and cost overruns by aligning execution strategies with infrastructure and service criticality. Overall, the results demonstrate that integrating AI across infrastructure analysis, healthcare accessibility modeling, and project management enables robust, adaptive, and equitable resilience planning for urban healthcare systems.