FUB Combining Production Rules and Ontologies: Tight and Loose Coupling

From ONTORULE Show Case

The main goal of this work is to combine production rules and ontologies. The aim of such a combination is to use an ontology to describe the vocabulary used in the rules. The language we consider for the description of such ontologies is the W3C standard OWL, which is a declarative, logical language, based on Description Logics family of knowledge representation languages. In contrast, the language of production rules has an operational semantics, i.e., the semantics is given by the algorithm used to execute the rules. The integration of two knowledge representation languages with such different semantics requires a solid theoretical foundation in order to understand the implications of the combination – both semantical and operational – on a deep level.

This work bridges the gap between the semantics of production rules and ontologies.

• Loose Coupling. This approach to the combination is characterized by the fact that the semantics of the production rules and the ontologies are decoupled: the interaction between the two semantics is based on entailment. First-Order Logic (FOL) and Description Logics (DLs) have been used to describe an ontology and a working memory. Different kinds of semantics of a condition satisfiability (patterns matching) and an action performance using some known results of research (for example, in the area of DL ontologies) have been developed. In particular, two different semantics of actions, changing a working memory are presented and discussed. One of them, the formula-based, is syntactical and based on a direct changing of a working memory by adding and removing of some facts, while the second one, called Possible Models Approach, is based on changing of common models of an ontology and a working memory, considering these models as a possible states-of-affairs of the application. On the basis of algorithms, this work also describes a procedure for computing the result of an action, i.e. a new working memory for some DLs from a DL−Lite family.

• Tight Coupling. An integrated semantics and an axiomatization of this semantics in fixpoint logic (FPL). Because of the integrated nature of this semantics, it encompasses more aspects of the production system: for example, the conflict- resolution strategy is an integral part of the definition of the semantics, as well as the axiomatization, although both have been constructed in such a way as to allow easy integration of other conflict-resolution strategies. The axiomatization may be used to analyze combinations and check certain properties, such as termination. A notable difference here with traditional production systems (PS), is that the set of predicates in the PS is divided between description logic (DL) predicates, i.e. predicates which may occur in the ontology, and production system predicates. In this way we interpret DL predicates under OWA, and PS predicates under CWA.

Corresponding results are reported in Deliverable 3.3.