Abstract:

Smart manufacturing employs advanced technologies like the Internet of Things, AI, and automation to optimize productivity, achieve operational flexibility, and streamline processes. In an era of hyperconnectivity, there is also an emergence of advanced cyber risk challenges due to the increased attack dimensions and evolving ecosystems. Smart manufacturing contexts lack the appropriate situational adaptability with cyber risk assessment mechanisms, leading to issues such as dealing with uncertainty, vagueness, non-linearity, and real-time risk assessment. These shortcomings enable the emergence of suboptimum risk evaluation with time lag response processes. To address these gaps, this paper presents a new framework called Bayesian-HFLNet, which stands for Bayesian and Hybrid Fuzzy Logic Network. This framework integrates Bayesian networks for probabilistic reasoning and with hybrid fuzzy logic’s capability to handle uncertainty and vague data. The Bayesian component models the conditional dependency structure among the existing network nodes which are the identified risks, while the fuzzy logic layer provides the means to address vagueness and ambiguity of some risk indicators through rule-based reasoning. This construction is designed for dynamic evaluations of cybersecurity risks in smart manufacturing. It processes data from the industrial network, sensor, and control systems to provide dynamic and continuous risk evaluation. Furthermore, Bayesian-HFLNet utilizes fuzzy rules to redefine threat clarity, ensuring effective identification and control recommendation. The results indicate improved accuracy of risk identification and detection when Bayesian-HFLNet is applied to dynamic scenarios compared with static cases, enhancing adaptability and decision-making capabilities.The outcomes of the experiments show an improvement in managing uncertainty, the rapid response to evolving dangers, and the overall defense posture of the system.This positions the framework as a solid tool for sustaining cybersecurity in intricate smart manufacturing systems.