Date of Award


Document Type


Degree Name

Master of Science


Department of Electrical and Computer Engineering

First Advisor

Gregg Gunsch, PhD


There are numerous methods used for troubleshooting devices. Each method has certain domains, knowledge requirements, and assumptions required for it to perform well. However, oftentimes no one method by itself is sufficient to completely solve a troubleshooting problem. Therefore, an architecture is required to control the combined use of many problem solving methods. The combination of multiple problem solving methods makes the troubleshooting process more robust in terms of device domains that can be dealt with and quality of diagnoses produced. Troubleshooting has two tasks: diagnosis and problem resolution. This research provides an architecture that allows dynamic method selection during diagnosis. Dynamic method selection factors the current state of the diagnosis process along with other method parameters to determine which method to use to advance the diagnosis process. The architecture was developed by combining themes from diagnosis research that focused on dynamic multimethod diagnosis and its control. This work has produced several results. It provides an architecture to organize the methods and a basis for making control decisions concerning method use during diagnosis. It identifies a generous number of methods useful to perform diagnosis. It identifies the knowledge these methods require.

AFIT Designator


DTIC Accession Number