Knowledge Representation And Processing Based On Infinite-Valued Model Of Medium Logic

How to describe and process negative knowledge is a basic problem in information science. In chapter 2, the paper make an introduction of concept essence of negative knowledge, ascertain five kinds negative relation (CDC, CFC, ODC, OFC and ROM) between distinct knowledge and fuzzy knowledge. According to five negative kinds negative relations, we analysis the capacity of some popular logic theories for negative knowledge representation and processing, and we find that most logic theories can't deal with the negative relations perfectly and exactly. On the other hand, medium logic is a logic theory which reflects five kinds of negative relations, and its interpretation of infinite valued provides a foundation for knowledge processing in different domains. Based on the interpretation of infinite valued for medium logic, the paper makes some research about negative knowledge representation and processing.In chapter 3, the paper makes a presentation about classical description logic ALC and fuzzy description logic FALC. As a fragment of predicate logic, ALC can't depict the negative relation between distinct knowledge, while FALC isn't provided with the faculty of present negative relation between fuzzy knowledge accurately. The paper introduces fuzzy negation～C and opposite negation╕C into basic description logic ALC. It proposes a new kind of description logic MALC with the ability to deal fuzzy information. And then, with the extension of interpret function, it provides an improved semantics for MALC which based on model of infinite valued of medium proposition logic and tableau-based algorithm for MALC.In chapter 4, the paper study on the answer set programming, and it illuminates that classical answer set programming and fuzzy answer set programming have some limitations which can't explain the classical negation and negation as failure. We deal classical negation in ASP as contradictory negation, while negation by default in ASP is treated with opposite negation. The Literals in FASP are true partly, and then the interpretation satisfies rules and program partly. For literals, we propose not only the interpretation of literals and its negations. And for interpretation consisted of literals, three functions CI, Sr and SΠare provided, which respectively measure the consistent degree of interpretations and degree of rules and program satisfied by interpretations. In the end, it defines the new answer set.In chapter 5, the paper studies the approximate reasoning of medium logic. The paper analyses the negative relation between fuzzy knowledge, which is described by medium logic and its interpretation of infinite valued for medium logic. The paper provides a new arithmetic which expands the CRI arithmetic. The semantic match degree and its formula which include semantic similarity and semantic distance are put forward. Further, the paper proposes an approximate reasoning approach based on the measures of the semantic match degree.