Data-Driven Assessment of Complexity-Induced Risks in Infrastructure Projects

Abstract

Infrastructure projects are characterized by inherent complexities that often lead to their poor performance. Notwithstanding challenges posed by various risks and their interactions, the additional non-linear and dynamic interdependence-induced complexities make infrastructure projects susceptible to systemic risks—probable component disruption that can lead to cascade (system-level) disruptions. The study of teams/resource interdependence-induced systemic risks in an environment of interacting risks is scarce in the literature. In addition, several previous studies demonstrated that current risk interactions and systemic risk analysis models are impractical due to their complexity and limited theoretical application domains. In this respect, the current study fills this knowledge gap by developing a data-driven risk interactions and systemic risk management approach. This approach is formulated in three stages: (1) quantifying risk interactions and teams/resources interdependence; (2) building machine learning model (ML) models to predict project performance based on the quantified characteristics; and (3) devising relevant mitigation strategies. The study also includes a practical demonstration application of the approach to present a step-by-step demonstration for each stage—thus guiding practitioners to proactively safeguard against risk interactions and systemic risks. The current work contributes to the body of knowledge by laying out the foundations of investigating the compound phenomenon of risk interactions and systemic risks as well as by presenting an effective approach to achieve that endeavor. Overall, the current study introduces a reliable and practical approach to enhance the performance of infrastructure projects through interacting risks- and systemic risk-informed management strategies.