| Abstract
|
Model-based diagnosis centers on the use o … Model-based diagnosis centers on the use of a behavioral model of asystem to infer diagnoses of anomalous behavior. For model-baseddiagnosis techniques to become practical, some serious problems in themodeling of uncertainty and in the tractability of uncertaintymanagement have to be addressed. These questions include: How can wetractably generate diagnoses in large systems? Where do the priorprobabilities of component failure come from when modeling a system?How do we tractably compute low-cost repair strategies? How can we dodiagnosis even if only partial descriptions of device operation areavailable? This dissertation seeks to bring model-based diagnosiscloser to being a viable technology by addressing these problems.We develop a set of tractable algorithms and modeling techniques thataddress each of the problems introduced above. Our approachsynthesizes the techniques used in model-based diagnosis andtechniques from the field of Bayesian networks. iques from the field of Bayesian networks.
|
