Title :
XDE: diagnosing devices with hierarchic structure and known component failure modes
Author :
Hamscher, Walter
Author_Institution :
Price Waterhouse Technol. Center., Menlo Park, CA, USA
Abstract :
The Extended Diagnosis Engine (XDE), a program that incorporates hierarchic diagnosis and fault models into a model-based diagnosis framework, is discussed. XDE depends upon an assumption-based truth maintenance system (ATMS). The novel aspects of XDE are that the diagnosis descends through hierarchies, one describing the physical organization of the devices and another describing its functional organization, and fault models are only used heuristically and are not required to be exhaustive. The combinatorial explosion caused by introducing fault models is partially addressed by introducing an explicit context-switching mechanism into the ATMS used by XDE. XDE has been used in a prototype system for troubleshooting board-scale digital circuits. Experience with XDE suggests the need for research into a more flexible control structure. The program should not always try refinements before decompositions, nor decompositions before probes. If it were possible to estimate the cost and benefit of the refinements and decompositions available, this decision could be made in the same decision analytic framework used by GDE (General Diagnostic Engine) to choose probes
Keywords :
circuit analysis computing; digital circuits; failure analysis; knowledge based systems; Extended Diagnosis Engine; XDE; assumption-based truth maintenance system; board-scale digital circuits; combinatorial explosion; component failure modes; context-switching mechanism; control structure; decision analytic framework; decompositions; fault models; functional organization; hierarchic structure; model-based diagnosis framework; physical organization; probes; refinements; troubleshooting; Artificial intelligence; Circuit faults; Context modeling; Contracts; Costs; Digital circuits; Engines; Explosions; Fault diagnosis; Prototypes;
Conference_Titel :
Artificial Intelligence Applications, 1990., Sixth Conference on
Conference_Location :
Santa Barbara, CA
Print_ISBN :
0-8186-2032-3
DOI :
10.1109/CAIA.1990.89170