Abstract: "Expert systems using experimental knowledge are widely used, but they cannot solve unexpected problems and it is difficult to acquire experimental knowledge from human experts. One approach to solving the problem is to use qualitative reasoning that uses deep knowledge. Qualitative r...

Full description

Bibliographic Details

Format:	Book
Language:	English
Published:	Tokyo, Japan 1992
Series:	Shin-Sedai-Konpyūta-Gijutsu-Kaihatsu-Kikō <Tōkyō>: ICOT technical report 791
Subjects:	Electronic circuit design Knowledge representation (Information theory)
Summary:	Abstract: "Expert systems using experimental knowledge are widely used, but they cannot solve unexpected problems and it is difficult to acquire experimental knowledge from human experts. One approach to solving the problem is to use qualitative reasoning that uses deep knowledge. Qualitative reasoning can estimate the value of quantities that change continuously during dynamic behaviors, but the cost of executing qualitative reasoning for large scale systems become huge. One approach to solving this problem is to use qualitative reasoning for the only parts that qualitative reasoning is necessary. In this paper, we try to use qualitative reasoning for the parts that deal with analog treatments and to use event reasoning for the parts that deal with digital treatments. Event reasoning is usually faster than qualitative reasoning because it does not evaluate continuous changes of quantities. In qualitative reasoning, on the other hand, even discontinuous changes are dealt with as continuous changes. Here we apply this approach to circuits consisting of a DTL circuit and two d-flipflops. The two d-flipflops are regarded as digital circuits and, as such, their internal structures are ignored. The DTL circuit is dealt with as an analog circuit consisting of a transistor and some resistors and diodes."
Physical Description:	30 S.