Game-theoretic analysis of attack and defense in cyber-physical network infrastructures

Critical infrastructures rely on cyber and physical components that are both subject to natural, incidental or intentional degradations. Game theory has been used in studying the strategic interactions between attackers and defenders for critical infrastructure protection, but has not been extensively used in complex cyber-physical networks. This paper fills the gap by modeling the probabilities of successful attacks in both cyber and physical spaces as functions of the number of components that are attacked and defended. The results show that the attack effort would first increase then decrease in (a) defense effort, (b) the probability of successful attack on each component, (c) the number of minimum required functioning resources, and (d) the maximum number of available resources. Comparing simultaneous and sequential games, our results show that the defender performs better when she moves first. Our research provides some novel insights into the survival of such infrastructures and optimal resource allocation under various costs and target valuations that players may have.