 |
International Society of Science and Applied Technologies |
| Mission Risk Control Strategy via Joint Optimization of Information Acquisition and Parameter Learning | ||||
| Author | Li Yang
|
|||
| Co-Author(s) |
|
|||
| Abstract | Conventional mission risk control policies typically rely on fixed inspection intervals and deterministic degradation models. These approaches, however, fail to account for system uncertainties, which can result in delayed failure detection and suboptimal decision-making. To address these challenges, we introduce an innovative dynamic decision framework that combines non-periodic information acquisition with Bayesian parameter learning. This novel framework mitigates the significant limitations of traditional models by allowing decisionmakers to select optimal time intervals for information acquisition, thereby minimizing unnecessary inspections and ensuring system safety. The proposed framework incorporates a real-time parameter learning mechanism that adapts to the uncertainty in system degradation, and we demonstrate that nonperiodic information acquisition does not compromise the accuracy of parameter estimation. We thoroughly explore the model's structural characteristics, proving the monotonicity of the value function and the existence of a mission abort control threshold. These insights are crucial for developing a thresholdbased control policy that dynamically balances the need for information acquisition and the decision to abort a mission. To implement this policy efficiently, we propose a computationally lightweight algorithm that minimizes decision model complexity, ensuring both timely and precise decision-making.
|
|||
| Keywords | Risk control, mission reliability, Bayesian parameter learning, mission abort | |||
| Article #: RQD2025-364 | ||||
Proceedings of 30th ISSAT International Conference on Reliability & Quality in Design |