Alembic: an efficient algorithm for CNF preprocessing

  • Authors:
  • Hyojung Han;Fabio Somenzi

  • Affiliations:
  • University of Colorado at Boulder;University of Colorado at Boulder

  • Venue:
  • Proceedings of the 44th annual Design Automation Conference
  • Year:
  • 2007

Quantified Score

Hi-index 0.00

Visualization

Abstract

Satisfiability (SAT) solvers often benefit from a preprocessing of the formula to be decided. For formulae in conjunctive normal form (CNF), subsumed clauses may be removed or partial resolution may be applied. Preprocessing aims at simplifying the formula and at increasing the deductive power of the solver. These two objectives are sometimes competing. We present a new algorithm that combines simplification and increase of deductive power and we show its effectiveness in speeding up SAT solvers.