Case-Based Troubleshooting Knowledge Management

AI in Equipment Maintenance and Support

Volume

Papers from the 1999 AAAI Spring Symposium

Contents

Lack of troubleshooting knowledge support combined with the time pressure to perform can adversely affect the performance on unplanned maintenance tasks. We propose that a knowledge management (KM) perspective should used to support unplanned maintenance tasks. The KM approach explicitly supports the continuous processes of creation, collection, organization, storage, and delivery of troubleshooting knowledge. We describe a set of methodologies based on the paradigm of case-based reasoning (CBR) for managing the troubleshooting knowledge. These methodologies have been implemented in a suite of software tools called SpotLight*. We demonstrate our approach with three applications in the aerospace industry. Exploratory analysis from a field trial at a regional airline shows that the tools can potentially generate cost savings of 8% on departure delays and cancellations. Issues for further research are presented.