A Probabilistic Approach to Hierarchical Model-based Diagnosis
This work addresses fault diagnosis in complex systems, but it appears incremental as it extends existing probabilistic methods to hierarchical models.
The paper tackles the problem of isolating faults in systems using model-based diagnosis by developing a fully probabilistic approach that supports hierarchical models, resulting in computational gains through a modified Bayesian network inference algorithm.
Model-based diagnosis reasons backwards from a functional schematic of a system to isolate faults given observations of anomalous behavior. We develop a fully probabilistic approach to model based diagnosis and extend it to support hierarchical models. Our scheme translates the functional schematic into a Bayesian network and diagnostic inference takes place in the Bayesian network. A Bayesian network diagnostic inference algorithm is modified to take advantage of the hierarchy to give computational gains.