22-colorings in Kk-regular Kk-uniform Hypergraphs
22-colorings in Kk-regular Kk-uniform Hypergraphs
复制标题
DOI:
10.1016/j.ejc.2013.04.005
复制
发表时间:
2013-10
期刊:
影响因子:
--
通讯作者:
Michael A. Henning;Anders Yeo
中科院分区:
文献类型:
--
作者:
Michael A. Henning;Anders Yeo
A hypergraph is 2-colorable if there is a 2-coloring of the vertices with no monochromatic hyperedge. Let Hkdenote the class of all k-uniform k-regular hypergraphs. The Lovász Local Lemma, devised by Erdös and Lovász in 1975 to tackle the problem of hypergraph 2-colorings, implies that every hypergraph H∈Hkis 2-colorable, provided k≥9. Alon and Bregman [N. Alon, Z. Bregman, Every 8-uniform 8-regular hypergraph is 2-colorable, Graphs Combin. 4 (1988) 303–306] proved the slightly stronger result that every hypergraph H∈Hkis 2-colorable, provided k≥8. It is implicitly known in the literature that the Alon–Bregman result is true for all k≥4, as remarked by Vishwanathan [S. Vishwanathan, On 2-coloring certain k-uniform hypergraphs, J. Combin. Theory Ser. A 101 (2003) 168–172] even though we have not seen it explicitly proved. For completeness, we provide a short proof of this result. As remarked by Alon and Bregman the result is not true when k=3, as may be seen by considering the Fano plane. Our main result in this paper is a strengthening of the above results. For this purpose, we define a set X of vertices in a hypergraph H to be a free set in H if we can 2-color V(H)∖X such that every edge in H receives at least one vertex of each color. Equivalently, X is a free set in H if it is the complement of two disjoint transversals in H. For every k≥13, we prove that every hypergraph H∈Hkof order n has a free set of size at least n/5. For any ϵ where 0<ϵ<1 and for sufficiently large k, we prove that every hypergraph H∈Hkof order n has a free set of size at least ckn, where ck=1−6(1+ϵ)ln(k)/k, and so ck→1 as k→∞. As an application, we show that the total restrained domination number of a graph on n vertices with sufficiently large minimum degree k is at most 12(1−ck)n, which significantly improves the best known bound of 12n+1.