Worst-case time bounds for coloring and satisfiability problems

  • Authors:
  • Tomás Feder;Rajeev Motwani

  • Affiliations:
  • Department of Computer Science, Gates 4B, Stanford University, Stanford, CA;Department of Computer Science, Gates 4B, Stanford University, Stanford, CA

  • Venue:
  • Journal of Algorithms
  • Year:
  • 2002

Quantified Score

Hi-index 0.00

Visualization

Abstract

We consider worst case time bounds for certain NP-complete problems. In particular, we consider the (k, 2)-satisfiability problem which includes as special cases some canonical problems such as graph coloring and satisfiability. For the (k, 2)-satisfiability problem, we present a randomized algorithm that runs in time O*((k!)n/k)2. This bound is equivalent to O((k/ck)n) with ck increasing to the asymptotic limit e. For k ≥ 11, we improve upon the O((0.4518k)n) randomized bound of Eppstein [Proceedings of the 12th Annual ACM-SIAM Symposium on Discrete Algorithms, pp. 329-337]. A special case of (k, 2)- satisfiability is k-colorability; here we achieve the above time bound for a slightly larger ck that has the same asymptotic behavior.