Attack-Defense Game Model with Multi-Type Attackers Considering Information Dilemma.

Area of Science:

• Network security
• Game theory applications
• Infrastructure protection

Background:

• Infrastructure networks are vital but vulnerable to attacks causing damage and disruption.
• Existing defense models often overlook attacker capabilities, leading to passive defender strategies.
• Information asymmetry in network defense is a critical factor influencing attack-prevention success.

Purpose of the Study:

• To develop a game theory model for infrastructure network protection that incorporates proficient attackers.
• To analyze the impact of information gaps and attacker types on defense strategies.
• To provide a framework for optimizing defender actions against diverse threats.

Main Methods:

• Introduction of the link-hiding rule and the information dilemma.
• Development of a Bayesian static game model for attack-defense scenarios with multiple attacker types.
• Application of Harsanyi transformation and bi-matrix game for model solution.
• Experimental validation using a scale-free network model.

Main Results:

• Defenders gain advantages by strategically hiding network links against normal attackers.
• Accurate estimation of attacker distribution significantly improves defender outcomes.
• Ignoring proficient attackers or misestimating their distribution leads to defender losses.
• The model demonstrates the importance of considering attacker sophistication in network defense.

Conclusions:

• Game theory offers a robust framework for infrastructure protection against sophisticated adversaries.
• Dynamic defense strategies, including information management, are essential for network security.
• Accurate threat assessment and adaptive strategies are key to mitigating risks in vulnerable networks.