This book addresses the development of a new generation of systems, Knowledge Base Management Systems (KBMS), specially constructed for effective and efficient management of knowledge bases, with the aim of filling part of the technological gap between artificial intelligence and databases. The book first investigates in detail the design process, the architecture, and the working methods of knowledge based systems (KS) in order to point out the key characteristics of the field as well as its current limitations, which serve then as basis for an exact formulation of KS requirements. An analysis and evaluation of other approaches (e.g., conventional DBS, non-standard DBS, coupling expert systems and DBS) to knowledge management is given. The book shows that in developing KBMS, the experience obtained by the investigation of each of these approaches is extremely important since they provide the basic concepts for building KBMS. The approaches should not be viewed as complete and final but as part of research work towards KBMS. A novel architecture is described for KBMS which integrates the functionality, flexibility and modeling power of DBS and AI. The main part of the book deals with all important architectural problems of KBMS: methods for knowledge representation with special emphasis on abstraction concepts, language proposals, and concepts for performance improvement. The book is based on practical experience accumulated over five years of successful research in coupling existing expert systems and DBS, extending DBS with deductive capabilities, and above all in the design and implementation of a KBMS prototype. Thus the book''s proposals are illustrated with detailed descriptions of their realization in existing systems or prototypes.

InhaltsangabeKnowledge and human problem solving.- A survey of knowledge-based systems.- Approaches towards KB management.- KBMS architectural issues.- The KBMS prototype KRISYS.- Modelling knowledge with KRISYS.- Conclusions.