Model-theoretic semantics and tractable algorithm for CNF-BCP

  • Authors:
  • Rahul Roy-Chowdhury;Mukesh Dalal

  • Affiliations:
  • Columbia University, Department of Computer Science, New York, NY;Columbia University, Department of Computer Science, New York, NY

  • Venue:
  • AAAI'97/IAAI'97 Proceedings of the fourteenth national conference on artificial intelligence and ninth conference on Innovative applications of artificial intelligence
  • Year:
  • 1997

Quantified Score

Hi-index 0.00

Visualization

Abstract

CNF-BCP is a well-known propositional reasoner that extends clausal Boolean Constraint Propagation (BCP) to non-clausal theories. Although BCP has efficient linear-time implementations, CNF-BCP requires clausal form transformation that sometimes leads to an exponential increase in the size of a theory. We present a new quadratic-time reasoner, RFP, that infers exactly the same literals as CNF-BCP. Although CNF-BCP has been specified only syntactically, we present a simple model-theoretic semantics for RFP. We also present a convergent term-rewriting system for RFP that is suitable for reasoning with knowledge bases that are built incrementally. Potential applications of RFP include logical truth-maintenance systems and general-purpose knowledge representation systems.