Hadwiger's conjecture is decidable

  • Authors:

  • Ken-ichi Kawarabayashi;Bruce Reed

  • Affiliations:

  • National Institute of Informatics}, Tokyo, Japan;McGill University}, Motreal, PQ, Canada

  • Venue:
  • Proceedings of the forty-first annual ACM symposium on Theory of computing
  • Year:
  • 2009

Quantified Score

Hi-index 0.00

Visualization

Abstract

The famous Hadwiger's conjecture asserts that every graph with no Kt-minor is (t-1)-colorable. The case t=5 is known to be equivalent to the Four Color Theorem by Wagner, and the case t=6 is settled by Robertson, Seymour and Thomas. So far the cases t ≥ 7 are wide open. In this paper, we prove the following two theorems: There is an O(n2) algorithm to decide whether or not a given graph G satisfies Hadwiger's conjecture for the case t. Every minimal counterexample to Hadwiger's conjecture for the case t has at most f(t) vertices for some explicit bound f(t). The bound f(t) is at most pppt, where p=101010t. Our proofs for both results use the well-known result by Thomassen [46] for 5-list-coloring planar graphs, together with some results (but not the decomposition theorem) of Graph Minors in [36]. Concerning the first result, we prove the following stronger theorem: For a given graph G and any fixed t, there is an O(n2) algorithm to output one of the following: a (t-1)-coloring of G, or a Kt-minor of G, or a minor H of G of order at most f(t) such that H does not have a Kt-minor nor is (t-1)-colorable. The last conclusion implies that H is a counterexample to Hadwiger's conjecture with at most f(t) vertices for the case t. The time complexity of the algorithm matches the best known algorithms for 4-coloring planar graphs (the Four Color Theorem), due to Appel and Hakken, and Robertson, Sanders, Seymour and Thomas, respectively. Let us observe that when t=5, the algorithm gives rise to an algorithm for the Four Color Theorem. The second theorem follows from our structure theorem, which has the following corollary: Every minimal counterexample G to Hadwiger's conjecture for the case t either has at most f(t) vertices, or has a vertex set Z of order at most t-5 such that G-Z is planar. It follows from the Four Color Theorem that the second assertion does not happen to any minimal counterexample to Hadwiger's conjecture for the case t. Thus in constant time, we can decide Hadwiger's conjecture for the case t.