Towards a Risk Sensitivity and Game-Theoretic Approach of Stochastic Fault-Tolerant Systems

From a statistical performance perspective, a risk-averse baseline control strategy for the linear-quadratic class of stochastic fault-tolerant control systems in presence of cognitive actuator tampers, structured and environmental uncertainties is obtained by a means of synchronous and adaptive game-theoretic framework. In addition, a radically new view of stochastic fault-tolerant systems design underpinned by mean-risk aware performance indexes is proposed to cater the changing needs for performance risk aversion and reliability despite of adversarial and stochastic elements from operational environments. Last but not least, a tool of computing various stationary higher-order statistics associated with the generalized chi-squared costs or losses has a strong impact on the feasibility of saddle-point equilibrium control strategies for both resilient controller and cognitive actuator tamper.