Adaptive Markov Game Theoretic Data Fusion Approach for Cyber Network Defense

Abstract

This paper extends our game theoretic situation awareness and impact assessment approach for cyber network defense to consider the change of threat intent during cyber conflict. From a perspective of data fusion and adaptive control, we use a Markov (stochastic) game method to estimate the belief of each possible cyber attack pattern. With the consideration that the parameters in each game player's cost function is not accessible to other players, we design an adaptation scheme, based on the concept of Fictitious Play (FP), for the piecewise linearized Markov game model. A software tool is developed to demonstrate the performance of the adaptive game theoretic high level information fusion approach for cyber network defense and a simulation example shows the enhanced understating of cyber-network defense.