Abstract:

It is necessary and urgent to set up an effective and economic early warning network to defend the critical targets in a harsh environment by detecting the intruders in a short time. A direct graph model based on the attack graph is established to descript the intruding plots and potential deployment locations of detectors. In order to deal with the configuration of the fixed and mobile monitors, a stochastic mathematical model is presented to minimize the early warning time and deployment cost while satisfying the detection probability threshold. Therefore, a simulation based optimization algorithm is proposed to solve the configuration problems. The simulation results show that the optimization algorithm can obtain a better performance gain compared with the random or hierarchical deployment methods.