Moving Target Defense Decision-Making Method: A Dynamic Markov Differential Game Model

Today most of the moving target defense decision-making methods are based on models of a discrete dynamic game. To more accurately study network attack-defense strategies against continuous confrontations, we analyze offensive and defensive behavior from a dynamic perspective. We propose a moving target defense decision-making method based on a model of a dynamic Markov differential game. We implement dynamic analysis and deduction of multi-stage continuous attack and defense confrontations for scenarios of continuous real-time network attack-defense. We take into account the influence of random factors and changes of the network system in the gaming process, combine differential gaming with the Markov decision-making method, and construct models of attack-defense games. We propose a solution for game equilibrium based on an objective function designed according to the total discounted payoff of the offensive and defensive game and the analysis of the characteristics of multi-staged game equilibrium. On this basis an optimal strategy selection method is designed. We apply and verify the game model and the defense strategy selection algorithm by using the moving target defense technique. We conduct simulations to verify the effectiveness and feasibility of the model and algorithm.