<?xml version="1.0" encoding="utf-8" standalone="yes"?>
<item>
  <id>06019055</id>
  <dt>j</dt>
  <an>06019055</an>
  <augroup>
    <au>Sciriha, Irene</au>
    <au>da Fonseca, C.M.</au>
  </augroup>
  <ti>On the rank spread of graphs.</ti>
  <so>Linear Multilinear Algebra 60, No. 1, 73-92 (2012).</so>
  <py>2012</py>
  <pu>Taylor \& Francis, Abingdon</pu>
  <lagroup>
    <la>EN</la>
  </lagroup>
  <ccgroup>
  </ccgroup>
  <utgroup>
    <ut>$\lambda $-optimal matrix</ut>
    <ut>minimum rank</ut>
    <ut>maximum nullity</ut>
    <ut>rank spread</ut>
    <ut>core graph</ut>
  </utgroup>
  <cigroup>
  </cigroup>
  <ligroup>
    <li>doi:10.1080/03081087.2011.567389</li>
  </ligroup>
  <abgroup>
    <ab>Summary: For a simple graph $G=({\Cal V},{\Cal E})$ with vertex-set ${\Cal V}= \{1,\dots,n\}$, let ${\Cal S}(G)$ be the set of all real symmetric $n$-by-$n$ matrices whose graph is $G$. We present terminology linking established as well as new results related to the minimum rank problem, with spectral properties in graph theory. The minimum rank $\text{mr}(G)$ of $G$ is the smallest possible rank over all matrices in ${\Cal S}(G)$. The rank spread $r_v(G)$ of $G$ at a vertex $v$, defined as $\text{mr}(G)- \text{mr}(G-v)$, can take values $\varepsilon\in\{0, 1,2\}$. In general, distinct vertices in a graph may assume any of the three values. For $\varepsilon=0$ or $1$, there exist graphs with uniform $r_v(G)$ (equal to the same integer at each vertex $v$). We show that only for $\varepsilon=0$, will a single matrix ${\bold A}$ in ${\Cal S}(G)$ determine when a graph has uniform rank spread. Moreover, a graph $G$, with vertices of rank spread zero or one only, is a $\lambda$-core graph for a $\lambda$-optimal matrix ${\bold A}$ in ${\Cal S}(G)$. We also develop sufficient conditions for a vertex of rank spread zero or two and a necessary condition for a vertex of rank spread two.</ab>
    <rv></rv>
  </abgroup>
</item>