The role of term symmetry in E-completion procedures

  • Authors:
  • Ralph W. Wilkerson;Blayne E. Mayfield

  • Affiliations:
  • Department of Computer Science, University of Missouri-Rolla, Rolla, MO;Department of Computer Science, Oklahoma State University, Stillwater, OK

  • Venue:
  • CSC '90 Proceedings of the 1990 ACM annual conference on Cooperation
  • Year:
  • 1990

Quantified Score

Hi-index 0.00

Visualization

Abstract

A major portion of the work and time involved in completing an incomplete set of reductions using an E-completion procedure such as the one described by Knuth and Bendix or its extension to associative-commutative equational theories as described by Peterson and Stickel is spent calculating critical pairs and subsequently testing them for confluence and coherence. A pruning technique which removes from consideration those critical pairs that represent redundant or superfluous information can make a marked difference in the run time and efficiency of an E-completion procedure to which it is applied. In this paper, a technique is proposed for removing critical pairs from consideration at various points before, during, or after their formation. This method is based on the property of term symmetry, which will be defined and explored with respect to E-unification and E-completion procedures. Informally, term symmetry exists between two terms when one can be transformed into the other through variable renaming. By identifying and eliminating various forms of term symmetry which arise between syntactic structures such as, reductions, critical pairs, subterms, and unifiers, it is possible to derive an E-completion procedure that produces the same results without processing these symmetric redundancies.