The CIFF Proof Procedure for Abductive Logic Programming with Constraints: Definition, Implementation and a Web Application

Abduction has found broad application as a powerful tool for hypothetical reasoning with incomplete knowledge, which can be handled by labeling some pieces of information as abducibles, i.e. as possible hypotheses that can be assumed to hold, provided that they are consistent with the given knowledge base. Attempts to make the abductive reasoning an effective computational tool gave rise to Abductive Logic Programming (ALP) which combines abduction with standard logic programming. A number of so-called proof procedures for ALP have been proposed in the literature, e.g. the IFF procedure, the Kakas and Mancarella procedure and the SLDNFA procedure, which rely upon extensions of different semantics for logic programming. ALP has also been integrated with Constraint Logic Programming (CLP), in order to combine abductive reasoning with an arithmetic tool for constraint solving. In recent years, many proof procedures for abductive logic programming with constraints have been proposed, including ACLP and the A-System which have been applied to many fields, e.g. multi-agent systems, scheduling, integration of information. This dissertation describes the development of a new abductive proof procedure with constraints, namely the CIFF proof procedure. The description is both at the theoretical level, giving a formal definition and a soundness result with respect to the three-valued completion semantics, and at the implementative level with the implemented CIFF System 4.0 as a Prolog meta-interpreter. The main contributions of the CIFF proof procedure are the advances in the expressiveness of the framework with respect to other frameworks for abductive logic programming with constraints, and the overall computational performances of the implemented system. The second part of the dissertation presents a novel application of the CIFF proof procedure as the computational engine of a tool, the CIFFWEB system, for checking and (possibly) repairing faulty web sites. Indeed, the exponential growth of the WWW raises the question of maintaining and automatically repairing web sites, in particular when the designers of these sites require them to exhibit certain properties at both structural and data level. The capability of maintaining and repairing web sites is also important to ensure the success of the Semantic Web vision. As the Semantic Web relies upon the definition and the maintenance of consistent data schemas (XML/XMLSchema, RDF/RDFSchema, OWL and so on), tools for reasoning over such schemas (and possibly extending the reasoning to multiple web pages) show great promise. The CIFFWEB system is such a tool which allows to verify and to repair XML web sites instances, against sets of requirements which have to be fulfilled, through abductive reasoning. We define an expressive characterization of rules for checking and repairing web sites' errors and we do a formal mapping of a fragment of a well known XML query language, namely Xcerpt, to abductive logic programs suitable to fed as input to the CIFF proof procedure. Finally, the CIFF proof procedure detects the errors and possibly suggests modifications to the XML instances to repair them. The soundness of this process is directly inherited from the soundness of CIFF