The reasoning of uncertainty information in an expert system

An expert system needs to deal with uncertainty information. In reality, however, a human being is not confident about how much he knows about the opposite of the information that he knows: that is, there is an unknown which should be incorporated in uncertainty processing. This paper therefore proposes the three values of y(A), n(A) and u(A) to fully represent the uncertainty information. It then thoroughly discusses how the proposed three values are logically reasoned in a rule-based system based on the multivalue logic relations