Abstract

The rapid digitization of civil infrastructure has precipitated an unprecedented convergence of physical engineering systems with computational intelligence, creating urgent requirements for sophisticated optimization frameworks capable of managing complexity, uncertainty, and real-time operational demands. This article presents a comprehensive examination of machine learning–driven optimization methodologies applied to smart infrastructure systems, addressing the fundamental engineering challenge of balancing performance, resilience, and sustainability across interconnected cyber-physical networks. The research synthesizes contemporary advances in supervised learning for failure prediction, reinforcement learning for adaptive control, and deep learning architectures for anomaly detection within transportation networks, energy grids, water systems, and urban infrastructure. Conceptual frameworks integrating physics-based modeling with data-driven approaches are examined, with particular emphasis on digital twin architectures that enable real-time synchronization between physical assets and computational representations. Predictive analytics methodologies for infrastructure resilience quantification are critically evaluated, encompassing uncertainty quantification, climate-adaptive simulations, and asset lifecycle optimization under stochastic operational conditions. Real-time engineering decision systems incorporating edge computing platforms and automated control feedback loops are analyzed for their translational potential in operational infrastructure management. The translational engineering implications extend to predictive maintenance scheduling, dynamic load balancing, and autonomous system reconfiguration under fault conditions. This review establishes that machine learning–driven optimization frameworks represent a transformative paradigm for smart infrastructure engineering, enabling unprecedented levels of operational efficiency while maintaining rigorous safety and reliability constraints essential for public infrastructure systems.