Attack Prediction Model Based on Repeated Game in Ad Hoc Network

Cao Qi-yinga · Jisuanji gongcheng · 2012

Most Intrusion Detection System(IDS) are weak in real-time response attacker's action,so it can not predict attacker's action well.Aiming at the problem,this paper puts forward an attack prediction model between attacker and IDS based on repeated game.It establishes the stage game model with the two players and gives the Nash equilibrium,gets the unique sub-game perfect Nash equilibrium when the game plays infinity times and chalks up the probability of every strategy of two players in stage 1~ stage t according to the Quantal Response Equilibrium(QRE) model with the Gambit simulation.The model is simulated in Ad Hoc network by using the GloMoSim,whose results show that compared with Nash equilibrium model,QRE equilibrium model gains better prediction effect.

Read the paper · More papers on PaperTik