Qualitative Reasoning for
Engineering Problem Solving
Sponsor: Computer Science Division, Office of Naval Research
Principal Investigator: Kenneth D. Forbus
Project Summary: This project explores the use of qualitative physics to integrate qualitative, quantitative, visual, and teleological knowledge in solving engineering problems. Qualitative physics provides a framework for organizing and integrating other kinds of engineering knowledge. It captures the kinds of knowledge engineers use when they are able to answer simple questions directly based on common sense, and when they figure out what specialized knowledge should be used in answering harder questions. Qualitative knowledge is crucial in interpreting quantitative results; for example, a boiler design that presumes an impossibly high operating temperature must be modified.
We have made substantial progress on developing formal models
for quantitative thermodynamics domain knowledge that support
automatic engineering analyses, on new ontologies for reasoning
about fluids, on new methods for visual reasoning about diagrams,
and on teleological representations plus an algorithm for
deriving function from structure and behavior. The best way to
summarize our current goals is that, in addition to continuing
these developments, we plan to integrate them into the first
broad-scale domain theory that incorporates qualitative,
quantitative, teleological, and spatial knowledge, organized
around multiple ontologies, that can be used for a variety of
engineering tasks. We want to develop our ideas and experimental
prototypes to the point where we can build systems that can
- Allow a person to draw (using a special-purpose graphical editor that provides domain-specific annotations) a pneumatic/hydraulic device and provide in a formal language a partial description of its purpose.
- Analyze the annotated diagram to extract the structural description normally provided by hand to qualitative reasoning systems.
- Generate a model for the system based on the question asked, using a reasonable perspective and level of abstraction.
- Apply the most appropriate reasoning technique (e.g., qualitative analysis, visual reasoning, symbolic algebra, and/or numerical computation) to the model to answer the question.
Back to Qualitative Spatial Reasoning page | Back to Self-Explanatory Simulators page | Back to Qualitative Process Theory page | Back to Compositional Modeling page | Back to Previous Projects page | Back to QRG Home Page