Defending a parallel system against a strategic attacker with redundancy, protection and disinformation

Abstract

Motivated by the defense of a power grid control center against cyber-attacks, this research considers the optimal allocation of defense resources for defending a parallel system consisting of identical elements. To enhance the system survivability, the defender distributes its limited resources to provide redundancy, disinform the attacker, and protect system elements. Correspondingly, the attacker allocates its limited resources to intelligence actions and impact contests. The defender determines the optimal number of genuine elements and false elements to deploy and the optimal amount of resources used for disinformation actions. The attacker seeks to maximize the system damage by taking the defender's strategy as given and allocating its resources optimally into intelligence actions and impact contests. Two types of parallel systems are considered: (1) where the system damage is defined as the probability of system failure, and (2) where the system damage is defined as the unsupplied demand of the system. Due to the complexity of the optimization model, we analyze the optimal defense strategy under various parameter settings via numerical methods. The results provide useful insights on how to defend such a system under different situations.