Imperfect Information Fusion Using Rules with Bilattice Based Fixpoint Semantics

We present an approach to reasoning non-uniformly by default with uncertain, incomplete and inconsistent information using sets of rules/extended logic programs in the context of multivalued logics with a bilattice structure. A fixpoint semantics for extended logic programs used in the process of inference is described, along with its computational approach. We show how this theoretic approach is applicable to the problem of integration of imperfect information coming from multiple sources.