Other Scholarly Content


Qualitative reasoning about engineering artifacts using qualitative process theory Public Deposited

Downloadable Content

Download PDF


Attribute NameValues
  • Over the past few years Qualitative Reasoning about physical systems has emerged strongly as an important area in Artificial Intelligence. Many qualitative reasoning formalisms have been developed and applied to varied domains, namely, hydraulic systems, electrical systems, industrial control etc. This paper examines Qualitative Process Theory [Forbus 84] formalism in view of its abilities and limitations in reasoning about physical systems. The main motivations of this endeavor are - to study the nature of problems involved in reasoning about complex physical systems - to investigate the applicability of Qualitative Process Theory formalism for reasoning about engineering artifacts - to extend the original QP theory to deal with complex physical systems. For this purpose a DC Shunt Generator is modelled using Qualitative Process Theory. The generator is modelled at different levels of complexity. The working Of the various models is discussed and then the the inferences sanctioned by QP theory for that model is examined. The investigation brings out the nature of the tasks involved in reasoning about such complex physical systems and demonstrates the abilities and limitations of the QP theory as a formalism for physical reasoning. Closed loop systems seem to offer tough chaHenge to the investigators in qualitative reasoning. It is in reasoning about closed loop models, that the QP theory fails to reason appropriately. It will be shown how the theory its too abstract quantity space representation and its too-general dependency structure makes the theory a weak one in reasoning about complex physical systems. The quantity space representation of the QP theory is extended and a new dependency propagation scheme is proposed. The generator is modelled on the extended QP theory and the inferences sanctioned by this extended theory are compared with the behavior of the system. The organization of the rest of the paper is as follows: Section 2, covers the modelling of the DC generator using the original QP theory and lists various inference categories involved in reasoning with engineering artifacts. Section 3, deals with the extensions to the QP theory. In section 4, the DC generator is modelled using the extended QP theory and the inferences sanctioned by the extended theory models are compared with the actual behavior of the generator. Some future research avenues and relevant reference materials are included. An algorithm for extracting the qualitative dependencies from quantitative equations is given in the Appendix.
  • 1987 best estimate for issue date based on available information.
Resource Type
Date Issued
Academic Affiliation
Rights Statement
Peer Reviewed



This work has no parents.