Cooperative Fault Tolerant Control of Security Games

This paper investigates the cooperative fault tolerant control of security games. In security game, an attacker will choose a path with maximally successful probability to compromise assets, and defender can adjust the resource allocation properly to minimize the successful probability. However, some nodes with malicious faults will reduce or improve the successful probability of defense, which causes greater losses for defender. Based on the proposed algorithm, the case of self tolerance is analyzed first by comparing the optimal solutions before and after faults. Furthermore, an algorithm of cooperative fault tolerant control is proposed to reduce the inadmissible cost by adjusting the defense function of healthy edges.