Zero-determinant strategy in stochastic Stackelberg asymmetric security game

Area of Science:

• Game Theory
• Artificial Intelligence
• Cybersecurity

Background:

• The strong Stackelberg equilibrium (SSE) is a common defender strategy in asymmetric security games.
• Attackers may exhibit bounded rationality, deviating from purely optimal strategies.
• Existing SSE strategies may falter against such boundedly rational or stubborn attackers.

Purpose of the Study:

• To introduce and analyze the unilateral-control zero-determinate (ZD) strategy for defenders.
• To evaluate the performance of the ZD strategy against boundedly rational attackers.
• To compare the ZD strategy's effectiveness against the SSE strategy.

Main Methods:

• Verification of the existence of ZD strategies for the defender.
• Comparative analysis of ZD and SSE strategies against a mixed best response (BR) and stubborn attacker strategy.
• Mathematical modeling of stochastic Stackelberg asymmetric security games.

Main Results:

• The existence of ZD strategies for the defender is confirmed.
• The ZD strategy demonstrates bounded loss compared to SSE when the attacker's strategy is near the BR strategy.
• The ZD strategy outperforms SSE when the attacker's strategy is closer to a stubborn approach.

Conclusions:

• The ZD strategy provides a robust defense against boundedly rational attackers in security games.
• ZD offers a viable alternative to SSE, particularly when attacker behavior deviates significantly from pure best response.
• This research contributes to developing more adaptive and effective defense mechanisms in security game theory.