Modeling knowledge merging systems

With the great richness and diversity of knowledge bases available today, a central difficulty is to combine knowledge coming from several sources in order to solve complex problems that could not be handled in isolation. In this study, we focus on two important characteristics about the knowledge merging problem: inconsistency and incompleteness. The purpose of this work is to present a logical framework that enables to formally specify knowledge merging systems and which tolerates inconsistency and incompleteness. This framework is based on a functional approach which is used in this paper to guide our formal development. The major features of this approach are dual reasoning, which includes a uniform way to deal with both aspects of partial and conflicting knowledge, and incrementality, which enables to gradually build pools of knowledge and to improve incomplete answers.