A context-based logic for distributed knowledge representation and reasoning

This paper is concerned with providing a logic, called Distributed First Order Logic (DFOL), for the formalization of distributed knowledge representation and reasoning systems. In such systems knowledge is contained in a set of heterogeneous subsystems. Each subsystem represents, using a possibly different language, partial knowledge about a subset of the whole domain, it is able to reason about such a knowledge, and it is able to exchange knowledge with other subsystems via query answering. Our approach is to represent each subsystem as a context, each context having its own language, a set of basic facts describing what is “explicitly known” by the subsystem, and a set of inference rules representing the reasoning capabilities of the subsystem. Knowledge exchange is represented by two different relations on contexts: the former on the languages (query mapping) and the latter on the domains (answer mapping) of different contexts. DFOL is based on a semantics for contextual reasoning, called Local Models Semantics, which allows to model contexts having different languages, basic knowledge, and reasoning capabilities, as well as relations between contexts. An axiomatization of DFOL is also presented.