Authors:
Kiyoshi Akama
1
and
Ekawit Nantajeewarawat
2
Affiliations:
1
Information Initiative Center and Hokkaido University, Japan
;
2
Sirindhorn International Institute of Technology and Thammasat University, Thailand
Keyword(s):
Model-Intersection Problem, Extended Clause, Function Variable, Equivalent Transformation.
Related
Ontology
Subjects/Areas/Topics:
Artificial Intelligence
;
Knowledge Engineering and Ontology Development
;
Knowledge Representation
;
Knowledge-Based Systems
;
Symbolic Systems
Abstract:
Built-in constraint atoms play a very important role in knowledge representation and are indispensable for practical applications. It is very natural to use built-in constraint atoms together with user-defined atoms when formalizing logical problems using first-order formulas. In the presence of built-in constraint atoms, however, the conventional Skolemization in general preserves neither the satisfiability nor the logical meaning of a given first-order formula, motivating us to step outside the conventional Skolemization and the usual space of first- order formulas. We propose general solutions for proof problems and query-answering (QA) problems on first-order formulas possibly with built-in constraint atoms. We map, by using new meaning-preserving Skolemization, all proof problems and all QA problems, preserving their answers, into a new class of model-intersection (MI) problems on an extended clause space, where clauses are in a sense ``higher-order''
since they may contain no
t only built-in constraint atoms but also function variables. We propose a general schema for solving this class of MI problems by equivalent transformation (ET), where problems are solved by repeated simplification using ET rules. The correctness of this solution schema is shown. Since MI problems in this paper form a very large class of logical problems, this theory is also useful for inventing solutions for many classes of logical problems.
(More)