| In the real world, information is often vague or ambiguous. However, in most commercial Database Management Systems (DBMS), it is impossible or inconvenient to describe such fuzzy information. Indubitably, it is an obvious shortage of traditional DBMS.Fuzzy database is an important and effective way for storing and managing incomplete information. Relational database has a long history and developed maturely. Thus, fuzzy relational database is a promising research direction. Designing and exploiting a Fuzzy Relational Database Management System (FRDBMS) is the only way or the most efficient way for the implementation of fuzzy relational database.In this paper, at first, the definition of fuzzy database is given. The latest development and research of fuzzy database have been summarized and the existing problems and solutions of fuzzy database are also pointed out.Secondly, relational fuzzy database modeling theory is discussed, and a two-level fuzzy database management system model is proposed, that is customer-oriented level and system-oriented level. In customer-oriented level, data models and relation models of traditional database are being extended in order to describe and store fuzzy information and data. In system-oriented level, a metadata base is designed and created in order to use and manage fuzzy information and data.Thirdly, the metadata auto-generation process is discussed. By use of this function, users can manipulate fuzzy databases as conveniently as traditional databases.Fourthly, several kinds of operations in the FRDBMS are defined. They are fuzzy algebraic operations, fuzzy logical operations, fuzzy relational operations, fuzzy arithmetical operations and fuzzy query operations. Some of these operations are programmed as system functions.At last, a simple FRDBMS is set up. Users can create fuzzy database upon it and achieve the store and part management of fuzzy data. |
PDF Full Text Request