Queries with Multivalued Logic-Based Semantics for Imperfect Information Fusion

Abstract

The paper is a study of a querying approach to the integration of imperfect information, using the framework offered by the multivalued logics based on the algebraic concept of bilattice. In particular, the queries are seen as integrating rules built with operations provided by the connectives of bilattices. An important aspect arising in the context of query evaluation is its efficiency, therefore we equally tackle the problem of query equivalence, which traditionally constitutes one of the central problems in query optimization. The paper provides results that characterize the query containment and equivalence problems syntactically, which naturally lead to algorithms for testing for equivalence of queries integrating imperfect information in the context of bilattices. Finally we study the complexity of query equivalence and the data complexity of query evaluation.