A.A. Sorochinsky

There is a large number of specialists, who are involved in the process of describing and analyzing the multiscale system. Each specialist works only with a part of this description, which is stored locally in an object-characteristic database (OCDB). OCDB bases on expert key concepts. Fuzzy query language can be used to build databases and for data mining. A fuzzy set is characterized by: - the range of possible values; - the type of membership function; - the parameters of membership function. There are various types of membership functions curves. The most widely membership functions are: triangular, trapezoidal, Gaussian and g-bell. It is necessary convert a fuzzy query into a sandart SQL to perform the transaction. It is necessary specify a threshold for the query to get those elements, whose degree of membership is not less than the specified number (α ? level). Query results generates after calculating of the query match index (QMI). The value calculating depends on logical operations, which were used. The steps of fuzzy query algorithm: Input: fuzzy query 1. Converting fuzzy query to a fuzzy SQL-like query 2. Converting fuzzy SQL-like query to a standard SQL query 3. Sending a standard SQL query to the database 4. Getting an answer from the database 5. Calculation QMI for each position of the answer 6. Ranging QMI Output: Query output OCDB manages to store and process both numerical and hierarchical descriptions at different semantic levels (with varying degrees of generalization). The using of fuzzy logic in queries provides subspecialists to select the necessary objects or characteristics through certain common characteristics, in terms of natural (spoken) language.