Randomly coloring planar graphs with fewer colors than the maximum degree

  • Authors:
  • Thomas P. Hayes;Juan C. Vera;Eric Vigoda

  • Affiliations:
  • Toyota Technological Institute at Chicago, Chicago, IL;Georgia Institue of Technology: College of Computing, Atlanta, GA;Georgia Institue of Technology: College of Computing, Atlanta, GA

  • Venue:
  • Proceedings of the thirty-ninth annual ACM symposium on Theory of computing
  • Year:
  • 2007

Quantified Score

Hi-index 0.00

Visualization

Abstract

We study Markov chains for randomly sampling k-colorings of a graph with maximum degree δ. Our main result is a polynomial upper bound on the mixing time of the single-site update chain knownas the Glauber dynamics for planar graphs when k=Ω(δ/logδ). Our results can be partially extended to the more general case where the maximum eigenvalue of the adjacency matrix of the graphis at most δ1-ε, for fixed ε 0. The main challenge when k ≤ δ + 1 is the possibility of "frozen" vertices, that is, vertices for which only one coloris possible, conditioned on the colors of its neighbors. Indeed, when δ = O(1), even a typical coloring canhave a constant fraction of the vertices frozen.Our proofs rely on recent advances in techniquesfor bounding mixing time using "local uniformity" properties.