Dynamic Domain Abstraction Through Meta-diagnosis

One of the most powerful tools designers have at their disposal is abstraction. By abstracting from the detailed properties of a system, the complexity of the overall design task becomes manageable. Unfortunately, faults in a system need not obey the neat abstraction levels of the designer. This paper presents an approach for identifying the abstraction level which is as simple as possible yet sufficient to address the task at hand. The approach chooses the desired abstraction level through applying model-based diagnosis at the meta-level, i.e., to the abstraction assumptions themselves.

[1]  Claudia Böttcher,et al.  No Faults in Structure? How to Diagnose Hidden Interactions , 1995, IJCAI.

[2]  Brian C. Williams,et al.  Diagnosing Multiple Faults , 1987, Artif. Intell..

[3]  A. V. Gemund,et al.  Diagnosing Intermittent Faults , 2008 .

[4]  W. Hamscher,et al.  XDE: diagnosing devices with hierarchic structure and known component failure modes , 1990, Sixth Conference on Artificial Intelligence for Applications.

[5]  P. Pandurang Nayak,et al.  Automated Modeling of Physical Systems , 1995, Lecture Notes in Computer Science.

[6]  Johan de Kleer,et al.  Modeling When Connections Are the Problem , 2007, IJCAI.

[7]  Brian Falkenhainer,et al.  Compositional Modeling: Finding the Right Model for the Job , 1991, Artif. Intell..

[8]  P. Pandurang Nayak.  Automated Modelling of Physical Systems , 1995 .

[9]  Raymond Reiter,et al.  Characterizing Diagnoses and Systems , 1992, Artif. Intell..

[10]  Pietro Torasso,et al.  Automatic Abstraction in Component-Based Diagnosis Driven by System Observability , 2003, IJCAI.

[11]  Luca Chittaro,et al.  Hierarchical model-based diagnosis based on structural abstraction , 2004, Artif. Intell..

[12]  de Kleer-Williams,et al.  A Theory of Model Simplification and Abstraction for Diagnosis Draft , 2003 .

[13]  Johan de Kleer,et al.  Characterizing Non-Intermittent Faults , 1991, AAAI.

[14]  Sanjaya Addanki,et al.  Reasoning About Assumptions in Graphs of Models , 1989, IJCAI.

[15]  Matthew L. Ginsberg,et al.  Readings in Nonmonotonic Reasoning , 1987, AAAI 1987.

[16]  Peter Struss,et al.  Task-dependent qualitative domain abstraction , 2005, Artif. Intell..

[17]  Mikel E. Goldblatt.  When to (and When Not To) Utilize Quality Six Sigma Tools in Water Treatment Problem Solving , 2007 .

[18]  Peter Struss.  What's in SD?: Towards a theory of modeling for diagnosis , 1992 .