Information Technology Reference
In-Depth Information
Chapter 4
Qualitative Reasoning
Qualitative reasoning starting from description of the physical system structure
and life system structure, deduces the behavior description in order to predict the
systematic behavior and provide the reason. It explains the systematic behavior
through local structural rules among the systematic parts, i.e. the transition of
behavior of part state only relates to directly adjacent part.
4.1 Introduction
The qualitative reasoning theory in artificial intelligence originates from the
study on physical phenomenon. The early works are usually focused on some
physical course, such as dynamics, hydrodynamics, thermal current et al. In 1952,
Simmons proposed the causality of qualitative analysis. In 1977, Rieger
published the paper (Rieger,1977) on simulation of causality. In 1984, Vol. 24 of
"Artificial Intelligence" magazine published a special edition of qualitative
reasoning, which included the laying foundation papers of de Kleer, Forbus and
Kuipers. This indicated that the qualitative reasoning was becoming mature. In
1986, Iwasaki and Simon published paper titled "Causality in Device
Behavior"(Iwasaki & Simon,1986). Over the past ten years, these basic
approaches have played an important role in the research and application of
qualitative reasoning, and make it become one of the most fruitful fields in
artificial intelligence. In 1993, Vol. 59 of "Artificial Intelligence" magazine
published a group of papers, reviewing these laying foundation works
Different qualitative reasoning approaches have been proposed according to
the different structural description of physical systems. Common approaches are
qualitative model approach by de Kleer (de Kleer,1984), qualitative process
approach by Forbus(Forbus,1984) and qualitative simulation approach by
147
