Hybrid risk control of mission-critical systems under degradation uncertainties: integrated maintenance-termination optimization

Malfunctions of safety–critical systems during mission implementation frequently lead to severe consequences and significant losses. To reduce the likelihood of system malfunction and enhance mission reliability, it is essential to implement effective risk control policies leveraging real-time degradation data. This study delves into a mission-critical system that experiences heterogenous degradation process throughout the mission execution. We account for the degradation heterogeneity among components by updating distributions of unknown degradation control parameters leveraging equidistantly sampled degradation signals. Utilizing parameter learning outcomes, this research establishes an adaptive control framework for maintenance restoration and mission termination planning, with the objective of minimizing the expected expense associated with maintenance, mission failure, and system malfunction. We establish a series of structural properties regarding the cost value function, and verify the existence of control-limits along with monotonic behaviors in directing informed maintenance and mission termination decisions. Furthermore, we conduct sensitivity analyses to evaluate how variations in core degradation/cost parameters impact the strategy optimality. Comparative experiments on ocean-going radar systems are performed to highlight model applicability in risk-aware loss control.