Some Promstk~ of Knowledge Information Inferred by Two Inter+ive

I. INTRODUCTION A new concept of beliefs called interactiue uxrension which are derivable from two interconnected default reasoning rystems is proposed. and its properties are discussed. Default reasoning proposed by R. Reiter is suitable for incomplete knowledge reasoning in artificial intellJgence, logic programming and deductive database, because it can draw plausible conclusions from the incomplete axioms. Since the conclusions can be invalidated when the partial world description is supplemented by new information, the logic based on the reasoning is a nonmonotonic logic When two default reasoning system are interconnected and exchange their beliefs each other, each system has a set of beliefs finally dy the default reasoning about the originally owned beliefs and the exchanged beliefs. A pair of acceptable se& of the beliefs for each system is called interactiue extension of the systems. When a large scale default reasoning system (deductive database) is comtructed by a consolidation of many small scale default reasoning systems, the beliefs of each system should he increased by the default reasoning and the interactions with other systems. Jn order to analyze the properties of the set of final beliefs acquired each system, the 'proposed interactive extension of the default systems h an indispensable concept. Permission to copy without fee all or part of tiis material is granted provided fhar the copies are not made or distributed for direct commercial advantage. Ihe DASFAA copyright notice and the title of the pubIication and its date appear, and notice is given that copying is by permission of the Organiu'ng Committee of the International Symposium on Database Syslcms for Advanced Applications. TO copy otherwise, or fo republish, requires a fee and /or special permission fmm the Organizing Committee. It has been known that artificial intelligence. logic programming and deductive d&abase; are Interconnected in the area of logical reasoning CU * classical logical reasoning is always monotonic, Le., the set of theorems increases monotonically with the set of axioms. The classical logic formalizing the reasoning, ag., first order logic, is descriptive& universal: Any-g that c+n be expressed in any language can be expressed in %he lo& The classical logic is., however, insufficient to solve 'effectively the sorts of problems in artificial intelligence, logic pro&sming; and deductive database. For example, such as the frame problem In artJficl&l; intelJigen& a descrlptlon using the &s;&al; logic may be extremely untieldy. Moreover, since a goOa deal of what we know about a …