Research On The Information Leakage Mechanism In The Terrorist Plot Detect Game

Since the "9·11" incident,terrorist organizations,such as The Islamic State(ISIS),"al-Qaida" organizations and the Taliban,tend to send more terrorists to launch coordinated attacks around the world.In China,terrorist organizations such as "East Turkic" was responsible for several terrorist attacks in Xinjiang,Tibet,and Kumming,which caused a heavy loss and is a huge threat to the living and property of civilian.In order to suppress the terrorism,many countries invested a large amount of budget in counterterrorism.However,a main challenge is that how to distribute security resources in an effective way that maximize the protection of the target from the terrorists threat and minimize the attack effect.The game-theoretic security resource allocation mechanism has provided an alternative framework to address the challenges.In the past decades,many game-theoretic based applications have been successfully deployed and destroyed many attacks in the real world.However,the “Snowden event” shows that,the intelligence agencies,such as FBI,still face the challenges from confidential information leakage by hackers or insight threat.The intelligence agencies have to consider the threat from information leakage in the strategy generation stage,but previous studies about monitoring strategy generation fail to consider the information leakage.To address this problem,this paper makes two major contributions.Firstly,this paper presents a Terrorist Plot Detect Game model with the consideration of Information Leakage(TPDIL),TPDIL develops a new Stackelberg game model for the security agency to generate optimal monitoring strategy to maximize the minimax payoff.While the player's strategy space grows in an exponential way,which restricts the problem size.Secondly,this thesis provides a Double Oracle algorithm for TPDIL(DO-TPDIL).At first,DO-TPDIL search the strategy by the polynomial algorithm for effective calculation,and only use the precise algorithm when the formal algorithm fails to find one.The experimental results show that the approach can obtain robust strategies against information leakage with high feasibility and efficiency.Comparing to the “brute force” way,our solution can improve the problem size efficiency.Our solution also performs better than some base-line solutions,the performance also stays stable with different attacker sizes.This thesis also discusses the effects of different leakage scenarios and parameters to the defender utility.