Attack-Defense Differential Game Model for Network Defense Strategy Selection

Abstract

The existing game-theoretic approaches for network security problems mostly use the static game or the multi-stage dynamic game. However, these researches can not meet the timeliness requirment to analyze the network attack and defense. It is better to regard the attack and defense as a dynamic and real-time process, in which way the rapidity and continuity of network confrontation can be described more precisely. Referring to the epidemic model SIR, we formulated the novel model NIRM to analyze the evolution of network security states. Based on the mentioned above, the attack-defense differential game model was constructed by introducing the differential game theory. Then we figured out the solution of saddle-point strategies in the game. By analyzing the game equilibrium, the algorithm of optimal defense strategies selection in the real-time confrontation was designed, which is more targeted and has greater timeliness. Finally by simulation experiments, we demonstrated the validity of the model and method proposed in this paper, and drew some instructive conclusions on network defense deployment.