A game theoretic analysis of intrusion detection in access control systems

We present a game-theoretic analysis of intrusion detection in access control systems. A security game between the attacker and the intrusion detection system is investigated both in finite and continuous-kernel versions, where in the latter case players are associated with specific cost functions. The distributed virtual sensor network based on software agents with imperfect detection capabilities is also captured within the model introduced. This model is then extended to take the dynamic characteristics of the sensor network into account. Properties of the resulting dynamic system and repeated games between the players are discussed both analytically and numerically.