Game-theoretic resilience analysis of Cyber-Physical Systems

Abstract

We investigate the resilience of cyber physical systems by modeling the interaction between provider and attacker as a simultaneous game that incorporates cyber and physical spaces. Both the provider and attacker aim to maximize their individual utility, which is determined by a trade-off between target revenue and investment cost. The system resilience function is formulated as a power-form product of the survival probabilities of cyber and physical spaces, each with a corresponding correlation coefficient. The contest success functions based on the reinforcement and attack levels are used to estimate the survival probabilities of cyber and physical spaces. We present the provider strategies based on the Nash equilibrium of the game, and analyze the sensitivities with respect to cyber and physical correlation coefficients, target revenues and costs. The results show that these correlation coefficients affect the cyber and physical reinforcement strategies, and also provide new insights into the system resilience.