Knowledge Representation and Processing in Relational Data Base

Abstract

This paper is a comparative study of the set operators in rule‐based production languages and relational query languages. This paper illustrates the use of structured query language (SQL) subselect conditions IN, NOT IN, EXISTS, and NOT EXISTS for knowledge processing. The method described in this paper may be effective for expanding existing relational data‐base management system applications to knowledge management systems and may reduce the need for algebraic or rule‐based programming to embed expertise. As an example, a set of knowledge tables that contain the relationships between in‐situ soil improvement techniques, relative cost, soil types, engineering properties, and the benefits and limitations of these improvement techniques are described. The operations for manipulating the tables to select and rank soil improvement techniques are written as SQL queries. The queries are compared to a set of expert system language rules that will accomplish the task. The results indicate the potential use of engineering knowledge tables for developing design aids. Forms of the queries described in this paper can be used by practitioners, data‐base developers and expert system programmers to expand the potential use of new and existing data bases.