Improvements of the theorem of Duchet and Meyniel on Hadwiger's conjecture (Q2565692)

From MaRDI portal
scientific article
Language Label Description Also known as
English
Improvements of the theorem of Duchet and Meyniel on Hadwiger's conjecture
scientific article

    Statements

    Improvements of the theorem of Duchet and Meyniel on Hadwiger's conjecture (English)
    0 references
    0 references
    0 references
    0 references
    28 September 2005
    0 references
    The well-known Hadwiger's conjecture states that every graph \(G\) has the complete graph \(K_{\chi(G)}\) as a minor where \(\chi(G)\) denotes the chromatic number of \(G\). \textit{D. R. Woodall} [J. Graph Theory 11, 197--204 (1987; Zbl 0672.05027)] conjectured that every graph \(G\) with \(n\geq 1\) vertices has \(K_{\lceil \frac{1}{\alpha(G)}n\rceil}\) as a minor where \(\alpha(G)\) denotes the independence number of \(G\). As \(\chi(G)\alpha(G)\geq n\), Hadwiger's conjecture implies Woodall's conjecture. In case \(\alpha(G)=2\), however, the two conjectures are equivalent [\textit{M. D. Plummer, M. Stiebitz} and \textit{B. Toft}, Discuss. Math., Graph Theory 23, 333-363 (2003; Zbl 1053.05052)]. In this context, a theorem of \textit{P. Duchet} and \textit{H. Meyniel} [Ann. Discrete Math. 13, 71--73 (1982; Zbl 0522.05060)] states that every graph \(G\) has a \(K_{\lceil \frac{1}{2\alpha(G)-1}n\rceil}\) as a minor. In this paper, the authors improve the theorem above as follows: (1) Every graph \(G\) with \(n\) vertices, clique number \(w(G)\) and independence number \(\alpha(G)\geq 2\) has \(K_{\lceil \frac{1}{2\alpha(G)-1}(n+\omega(G))\rceil}\) as a minor, (2) Every graph \(G\) with \(n\) vertices and independence number \(\alpha(G)\geq 3\) has \(K_{\lceil \frac{1}{(4\alpha(G)-3)(2\alpha(G)-1)}(n(4\alpha(G)-2))\rceil}\) as a minor, and (3) Let \(G\) be a graph with \(n\) vertices and \(\alpha(G)=2\). If \(G\) is \((n-4)\)-connected, then \(G\) has \(K_{\lceil \frac{1}{2}n\rceil}\) as a minor. For general graphs \(G\) with \(\alpha(G)=2\), a problem due to Paul Seymour still remains open. It reads: Is it possible to obtain a larger constant than \(\frac{1}{3}\) in the theorem of Duchet and Meyniel?
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    independence number
    0 references
    connected matching
    0 references
    0 references