Holistic Assurance Framework
AUTHOR: Dr. Lance Sherry
GEORGE MASON UNIVERSITY
One of the challenges in designing and operating systems composed of interacting components is validating that the emergent behavior of the system does not cause one or more components to migrate, over time, into a hazardous operating state. Research on several modern airline accidents exhibit the characteristics of Interaction Accidents – no component failed, but the interaction of components resulted in a hazardous state.
Due to the dependence on time, emergent behavior cannot be evaluated by analysis of the design. In theory, it can be evaluated by Digital-Twin agent-based simulations. However, running these simulations to uncover rare event emergent hazardous state is prohibitive due to:
(1) the combinatorics of initial states of each of the components, and
(2) the combinatorics of the time dimension (i.e., small variations in timing can result in very different outcomes).
Deep Learning Neural Networks (DLNNs) have shown promise to capture the underlying combinatoric behavior as well as compress the time dimension. This report demonstrates the application of DLNN to identify emergent behavior from components with hybrid moded/continuous behavior that plays out over time. DLNNs were trained and tested for three systems with increasing behavioral complexity. The DLNNs were able to accurately represent the time-dependent behavior for which they were trained/tested. The DLNNs were also able to learn and predict emergent behavior for behaviors that were not included in the training/testing data (up to 63% of the missing cases).
These results suggest that DLNNs could be used to supplement MBSE/Digital-Twins to increase the operational state-space coverage for System Validation Testing. The implications of these results are discussed.