Computing Reviews

Model-Based Diagnosis or Reasoning from First Principles
Peischl B., Wotawa F. IEEE Intelligent Systems & Their Applications18(3):32-37,2003.Type:Article
Date Reviewed: 11/06/03

Diagnosis of malfunctions based on reasoning from first principles is discussed in this good research paper. However, the term “or” in the title is misleading: Model-based diagnosis can be done to detect design flaws or operational malfunctions, and can be based on simulation or on symbolic processing of a diagnosis model of the design. The paper would also benefit from a brief discussion of related works.

In the paper, basic terms are clearly defined, and the malfunction diagnosis problem, as well as the proposed solution, is well explained with simple circuits. Specifically, a consistency-based approach where the reasoning is from causes to effects is explained, and the benefits of hierarchical modeling are clarified with an example. Successful applications in software debugging, equipment configuration, vehicle diagnosis, space probes, and some other areas are elaborated.

Diagnostic problems are important, and reliability of the solution techniques is challenging. The reliability issue is not mentioned or tackled in the paper. For example, seven sentences in first-order logic are provided for the diagnosis model of the circuit without any explanation about why their completeness is important. Especially for large models, the assurance of the completeness (and also the redundancy) of the rules is of primordial importance.

Reviewer:  Tuncer I. Ören Review #: CR128518 (0403-0353)

Reproduction in whole or in part without permission is prohibited.   Copyright 2024 ComputingReviews.com™
Terms of Use
| Privacy Policy