The algorithmic complexity of k-domatic partition of graphs

  • Authors:
  • Hongyu Liang

  • Affiliations:
  • Institute for Interdisciplinary Information Sciences, Tsinghua University, Beijing, China

  • Venue:
  • TCS'12 Proceedings of the 7th IFIP TC 1/WG 202 international conference on Theoretical Computer Science
  • Year:
  • 2012

Quantified Score

Hi-index 0.00

Visualization

Abstract

Let G=(V,E) be a simple undirected graph, and k be a positive integer. A k-dominating set of G is a set of vertices S⊆V satisfying that every vertex in V∖S is adjacent to at least k vertices in S. A k-domatic partition of G is a partition of V into k-dominating sets. The k-domatic number of G is the maximum number of k-dominating sets contained in a k-domatic partition of G. In this paper we study the k-domatic number from both algorithmic complexity and graph theoretic points of view. We prove that it is $\mathcal{NP}$-complete to decide whether the k-domatic number of a bipartite graph is at least 3, and present a polynomial time algorithm that approximates the k-domatic number of a graph of order n within a factor of $(\frac{1}{k}+o(1))\ln n$, generalizing the (1+o(1))ln n approximation for the 1-domatic number given in [5]. In addition, we determine the exact values of the k-domatic number of some particular classes of graphs.