A delayed splitting bottom-up procedure for model generation

  • Authors:
  • Kiyoshi Akama;Ekawit Nantajeewarawat

  • Affiliations:
  • Information Initiative Center, Hokkaido University, Hokkaido, Japan;Computer Science Program, Sirindhorn International Institute of Technology, Thammasat University, Pathumthani, Thailand

  • Venue:
  • AI'12 Proceedings of the 25th Australasian joint conference on Advances in Artificial Intelligence
  • Year:
  • 2012

Quantified Score

Hi-index 0.00

Visualization

Abstract

Meaning-preserving Skolemization is essential for development of a correct and efficient method of solving query-answering problems. It requires global existential quantifications of function variables, which in turn require an extended space of logical formulas. This paper proposes a bottom-up procedure for computing a set of models that sufficiently represents the set of all models of a given clause set in the extended formula space. Instantiations of function variables often result in generation of infinitely many models. To overcome the difficulty, a model-making pattern is introduced for representing a possibly infinite number of models, and such a pattern is split as late as possible. The proposed procedure provides a method for solving query-answering problems that include unrestricted use of universal and existential quantifications.