Graph Neural Networks for Evaluating the Reliability and Resilience of Infrastructure Systems: A Systematic Review of Models, Applications and Future Directions

Interdependent infrastructure systems serve as the foundation of modern society, ensuring the delivery of essential services, economic stability, and public well-being. The increasing integration of these systems, driven by advancements in cyber-physical systems (CPS) and the Internet of Things (IoT), has improved operational efficiency and adaptability. However, this interconnectivity and interdependency have also introduced vulnerabilities, making infrastructure networks more susceptible to cascading failures, cyber attacks, and disruptions from natural disasters and human-induced events. Graph neural networks (GNNs) have emerged as a powerful tool for addressing these challenges, offering data-driven solutions for reliability and resilience analysis across interconnected infrastructure networks. This paper provides a comprehensive review of GNN applications in interdependent infrastructure systems, including transportation networks, power distribution networks, water distribution networks, and communication networks. By leveraging spatial-temporal modeling, multi-modal data integration, and physics-informed learning, GNN-based approaches enhance predictive accuracy, system resilience, and decision-making efficiency. This review underscores the potential of GNNs in infrastructure management optimization and risk mitigation, and enabling the development of more sustainable, adaptive, and resilient infrastructure systems in the face of evolving challenges.