Multi-Leader Multi-Follower Stackelberg Game Based Defense Strategies Optimization for Interdependent Systems

Many critical infrastructures today exhibit complex interdependencies, which facilitate cross system attacks. Therefore, intrusion intent and path prediction for interdependent systems are crucial. This article proposes multi-leader multi-follower Stackelberg game model utilizing attack graphs to depict potential attack paths and defense strategies, aimed at optimizing defense strategies decisions for interdependent systems. To address potential conflicts of interest in collaborative decision-making among multiple defenders, we introduce two game theory algorithm variants that balance individual and collective utility. The experimental results show that game theory based algorithms exhibit significant advantages in both runtime performance and solution quality, due to their mutual adjustment in multiple iterations, which is more in line with the stability requirements of practical applications.