We formulate a resource-planning game between an attacker and a defender of a network control system. We consider the network to be operating in closed-loop with a linear quadratic regulator (LQR). We construct a general-sum, twoplayer, mixed strategy game, where the attacker attempts to destroy communication equipment of some nodes, and thereby render the LQR feedback gain matrix to be sparse, leading to degradation of closed-loop performance. The defender, on the other hand, aims to prevent this loss. Both players trade their control performance objectives for the cost of their actions. A Mixed Strategy Nash Equilibrium (MSNE) of the game represents the allocation of the players' respective resources for attacking or protecting the network nodes. We analyze the dependence of a MSNE on the relative budgets of the players as well as on the important network nodes that must be preserved to achieve a desirable LQR performance. MSNE is computed using nonlinear programming. Results are validated using the New England power system model, and it is shown that reliable defense is feasible unless the cost of attack is very low or much smaller than the cost of protection per generator.
We develop investment approaches to secure electric power systems against load attacks that may cause voltage instability. The attacker attempts to alter the reactive power setpoints of the loads covertly and intelligently to reduce the voltage stability margin of the grid. The defender, or the system operator, aims to compensate for this reduction by retuning the reactive power dispatch of control devices such as shunt capacitor banks. The question is: how much financial investment should the attacker and the defender plan for to succeed in their respective objectives? To address this question, we formulate a cost-based Stackelberg game, where the defender is aware of the attacker’s budget, and a robust-defense sequential algorithm for the realistic case when the defender is not fully informed about the attacker’s resources. We demonstrate that these methods operate reliably under time-varying load uncertainties. To provide scalability to large-scale power system models, we develop a genetic algorithm where both players evolve their candidate solutions in opposite directions simultaneously. Finally, the proposed methods are validated using IEEE prototype models, demonstrating that reliable and robust defense is feasible unless the defender’s resources are severely limited relative to the attacker’s resources.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.