Risk Assessment Method for Edge Intelligence Control Platforms Based on Hybrid Game Theory

Risk assessment techniques have been widely used in edge intelligent control platforms from security protection to decision optimization. In recent years, many researchers have applied machine learning or deep learning to risk assessment in edge intelligent control platforms. However, these approaches face significant challenges, including high computational resource requirements, high data dependency, and poor interpretability. In this paper, we propose a risk assessment method for edge intelligent control platforms based on hybrid game theory, which constructs a two-dimensional security risk assessment framework across cyber and physical domains. The attack tree model is first utilized to meticulously outline the potential attack paths and integrate the mixed-strategy game between attackers and defenders at the leaf nodes. Then, through game-theoretic analysis, the payoff functions of both parties are established, Nash equilibrium is determined to predict strategic choices, and the fuzzy analytic hierarchy process (FAHP) is combined with the CRITIC weighting method to quantify the risk associated with each node in the attack tree model. Finally, the risk values of the root nodes are aggregated to assess the overall security level of the platform. This approach effectively simulates real-world adversarial interactions and improves the accuracy and practicality of risk assessment.