The existence of uniform hypergraphs for which the interpolation property of complete coloring fails

From MaRDI portal
Publication:2065902



Abstract: In 1967 Harary, Hedetniemi, and Prins showed that every graph G admits a complete t-coloring for every t with chi(G)letlepsi(G), where chi(G) denotes the chromatic number of G and psi(G) denotes the achromatic number of G which is the maximum number r for which G admits a complete r-coloring. Recently, Edwards and Rzc a.{z}ewski (2020) showed that this result fails for hypergraphs by proving that for every integer k with kge9, there exists a k-uniform hypergraph H with a complete chi(H)-coloring and a complete psi(H)-coloring, but no complete t-coloring for some t with chi(H)<t<psi(H). They also asked whether there would exist such an example for 3-uniform hypergraphs and posed another problem to strengthen their result. In this paper, we generalize their result to all cases k with kge3 and settle their problems by giving several kinds of 3-uniform hypergraphs. In particular, we disprove a recent conjecture due to Matsumoto and the third author (2020) who suggested a special family of 3-uniform hypergraph to satisfy the desired interpolation property.


The authors study the so-called interpolation property for uniform hypergraphs, which concerns complete (vertex) colorings of uniform hypergraphs. The main result of the paper states that for any \(k \geq 3\), there exist \(k\)-uniform hypergraphs that do not have the interpolation property in a strong sense (defined below). A \(k\)-uniform hypergraph \(H\) is a hypergraph whose edges all have cardinality \(k\). A \(t\)-coloring of a hypergraph \(H\) is an assignment of colors to the vertices of \(H\), where the colors come from a collection of \(t\) colors. A coloring is proper if any two vertices that are in an edge receive distinct colors. The chromatic number of \(H\), denoted by \(\chi(H)\), is the minimum number of colors needed to properly color \(H\). A coloring of a \(k\)-uniform hypergraph \(H\) is complete if is proper and every subset of \(k\)-colors appears on at least one edge of \(H\). The achromatic number of \(H\), denoted by \(\psi(H)\) is the largest integer \(t\) such that there is a complete \(t\)-coloring of \(H\). It is known that every graph \(G\) has a complete coloring. A result of \textit{F. Harary} et al. [Port. Math. 26, 453--462 (1967; Zbl 0187.20903)] states that that if \(G\) is a graph, then \(G\) admits a \(t\)-coloring for every \(t\) such that \(\chi(G) \leq t \leq \psi(G)\). In other words, every graph has the interpolation property. On the other hand, \textit{K. Edwards} and \textit{P. Rzążewski} [Discrete Math. 343, No. 2, Article ID 111673, 12 p. (2020; Zbl 1429.05065)] showed that, for all \(k \geq 9\), there exist hypergraphs \(H\) which have complete \(\chi(H)\)-coloring and \(\psi(H)\)-colorings, but no complete \(t\)-coloring for some \(\chi(H) \leq t \leq \psi(H)\). The authors strengthen the result of Edwards and Rzążewski [loc. cit.], and show that for any \(k \geq 3\), there exist hypergraphs \(H\) which have complete \(\chi(H)\)-coloring and \(\psi(H)\)-colorings, but no complete \(t\)-coloring for any \(\chi(H) \leq t \leq \psi(H)\).





Describes a project that uses

Uses Software






This page was built for publication: The existence of uniform hypergraphs for which the interpolation property of complete coloring fails

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2065902)