A Locally Constrained Parallel Activation Model for Diagnostic Reasoning
Abstract
A general purpose competition-based parallel activation paradigm for diagnostic reasoning has been formulated. To facilitate the task of formulating and testing this paradigm, a parallel activation model generator had also to be developed. Using the network specification language provided by this generator, a knowledge base for diagnosing faults in a prototype chemical processing plant was built to test the viability of the proposed approach as a practical diagnostic paradigm. Diagnosis of failures in process plants has been attempted in the past using conventional AI methodologies, which have raised several practical issues which need to be resolved before a viable automated tool can be built. (kr)
Document Details
- Document Type
- Technical Report
- Publication Date
- Oct 01, 1989
- Accession Number
- ADA217808
Entities
People
- Sanjeev B. Ahuja
Organizations
- University of Maryland, Baltimore