Computational methodologies for critical infrastructure resilience modeling: A review

Abstract

Modeling the resilience of critical infrastructures (CIs) is broadly viewed as critical to maintaining the normal condition of CIs as a result of frequent threats that can disrupt safety, security, and business continuity. The goal of this research is to present a comprehensive review of computational methodologies for CI resilience modeling and development. Through both quantitative and qualitative analysis, a systematic review of the existing relevant literature is conducted to understand the current status and identify future demands. To begin with, this research conducts a scientometric review of 1130 journal articles to explore the research areas, keywords, and clusters. Then, an overview of CI resilience is provided, which briefly analyzes the resilience of CIs and the benefits of computational methodologies. Special attention has been given to five popular computational methodologies in CI resilience that have been widely adopted for managing CI resilience, including network-topology-based, simulation-based, optimization-based, data-driven-based, and physics-informed-machine-learning-based methodologies. To respond to the development of technologies and the advancement of devices, the challenges and future trends of embracing the coming of smart CIs, such as model transparency and explainability, multi-scale human–computer interaction, model portability, robustness, security, and safety, are identified to facilitate progress towards more resilient CIs.