Solving covering problems using LPR-based lower bounds

  • Authors:
  • Stan Liao;Srinivas Devadas

  • Affiliations:
  • Advanced Technology Group, Synopsys, Inc.;Department of EECS, MIT

  • Venue:
  • DAC '97 Proceedings of the 34th annual Design Automation Conference
  • Year:
  • 1997

Quantified Score

Hi-index 0.00

Visualization

Abstract

Unate and binate covering problems are a special class ofgeneral integer linear programming problems with which several problemsin logic synthesis, such as two-level logic minimization and technologymapping, are formulated. Previous branch-and-bound methodsfor exactly solving these problems use lower-bounding techniques basedon finding maximal independent sets. In this paper we examine lower-boundingtechniques based on linear programming relaxation (LPR) forthe binate covering problem. We show that a combination of traditionalreductions (essentiality and dominance) and incremental computation ofLPR-based lower bounds can exactly solve difficult covering problemsorders of magnitude faster than traditional methods.