The State-of-the-Art Survey on Optimization Methods for Cyber-physical Networks

TL;DR

This survey reviews optimization methods for cyber-physical systems, highlighting current practices, research trends, and future needs for scalable, data-driven solutions to enhance system resilience and decision-making.

Contribution

It provides a comprehensive synthesis of mathematical modeling and optimization techniques in CPS, categorizing methods and analyzing research trends through bibliometric approaches.

Findings

Identifies main mathematical modeling practices for CPS optimization.

Highlights the dominance of exact, heuristic, and meta-heuristic solution approaches.

Discusses the need for scalable, data-driven optimization algorithms in future research.

Abstract

Cyber-Physical Systems (CPS) are increasingly complex and frequently integrated into modern societies via critical infrastructure systems, products, and services. Consequently, there is a need for reliable functionality of these complex systems under various scenarios, from physical failures due to aging, through to cyber attacks. Indeed, the development of effective strategies to restore disrupted infrastructure systems continues to be a major challenge. Hitherto, there have been an increasing number of papers evaluating cyber-physical infrastructures, yet a comprehensive review focusing on mathematical modeling and different optimization methods is still lacking. Thus, this review paper appraises the literature on optimization techniques for CPS facing disruption, to synthesize key findings on the current methods in this domain. A total of 108 relevant research papers are reviewed following an extensive assessment of all major scientific databases. The main mathematical modeling practices and optimization methods are identified for both deterministic and stochastic formulations, categorizing them based on the solution approach (exact, heuristic, meta-heuristic), objective function, and network size. We also perform keyword clustering and bibliographic coupling analyses to summarize the current research trends. Future research needs in terms of the scalability of optimization algorithms are discussed. Overall, there is a need to shift towards more scalable optimization solution algorithms, empowered by data-driven methods and machine learning, to provide reliable decision-support systems for decision-makers and practitioners.