COUPLING EXPERT SYSTEMS WITH DATABASE MANAGEMENT SYSTEMS

Abstract

The combined use of Database Management Systems (DBMS) and Artificial Intelligence-based Expert Systems (ES) is potentially very valuable for modern business applications. The large body of facts usually required in business information systems can be made available to an ES through an existing commercial DBMS. Furthermore, the DBMS itself can be used more intelligently and operated more efficiently if enhanced with ES features. However, the implementation of a DBMS-ES cooperation is very difficult. We explore practical benefits of the cooperative use of DBMS and ES, as well as the research challenges it presents. Strategies for providing data from a DBMS to an ES are given; complementary strategies for providing intelligence from an ES to a DBMS are also presented. Finally, we discuss architechural issues such as degree of coupling, and combination with quantitative methods. As an illustration, a research effort at New York University to integrate a logic-based business ES with a relational DBMS is described.