INTERFACING DATABASES AND KNOWLEDGE BASED SYSTEMS FOR STRUCTURAL ENGINEERING APPLICATIONS (ARTIFICIAL INTELLIGENCE

Database management systems are important components of existing integrated computer-aided design systems. Knowledge-based systems are beginning to be applied to a broad range of engineering problems. A powerful, adaptable interface between these components is necessary to build an integrated structural engineering computing environment.;The thesis examines the basic issues involved in interfacing knowledge-based systems (KBS) with database management systems (DBMS) and presents a conceptual model for a knowledge-based approach to interfacing these two types of systems. In the model, each KBS and DBMS contains knowledge describing its data structures and data access methods. The data spaces of the components are integrated into a single global data space. A knowledge-based interface processor responds to requests from KBSs by sending requests to the available DBMSs. Between a component (KBS or DBMS) and the interface, a request undergoes a two-step translation: syntactic translation between the component's request language and the interface request language, and semantic translation between the context of the component's data representation and the context of the global data representation.;The conceptual model is implemented in a prototype system, KADBASE. KADBASE is a flexible, knowledge-based interface in which multiple expert systems and multiple databases can communicate as independent, self-descriptive components within an integrated, distributed engineering computing system. The architecture of KADBASE consists of knowledge-based system interfaces (KBSI) for each KBS, knowledge-based database interfaces (KBDBI) for each DBMS, and a network data access manager (NDAM). The NDAM accepts requests from the KBSs through their KBSIs, locates the required data in the available DBMSs, and sends requests to those DBMSs through their KBDBIs. The KBSIs and KBDBIs are responsible for syntactically and semantically translating requests from the component's request language and data space into the global request language and data space.;The thesis presents examples for three knowledge-based systems to demonstrate the use of KADBASE in typical structural engineering design applications. The three applications are a structural section design system, a building cost estimator, and a steel frame joint designer.