Abstract

One of the primary areas of application of Artificial Intelligence is diagnosis. Diagnosis from first principles is a diagnostic technique which uses knowledge of the designed structure and function of a device to determine the possible causes of the malfunction.

This work builds on the foundation of a theory of diagnosis by implementing and extending the theory. A correction to the algorithm which defines the theory is presented. The theory is extended for multiple sets of observations of the system and measurement data.

A fundamental problem in diagnosis is selecting the measurement which will be of the most benefit in reducing the number of competing diagnoses for a system. A heuristic which selects a component whose measurement is likely to be beneficial in isolating the actual diagnosis is also presented.

Department(s)

Computer Science

Comments

This report is substantially the Ph.D. dissertation of the first author, completed December 1988.

Report Number

CSc-88-10

Document Type

Technical Report

Document Version

Final Version

File Type

text

Language(s)

English

Rights

© 1988 University of Missouri--Rolla, All rights reserved.

Publication Date

December 1988

Share

 
COinS