On nonlinear preservers of weak matrix majorization (Q2460752)
From MaRDI portal
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | On nonlinear preservers of weak matrix majorization |
scientific article |
Statements
On nonlinear preservers of weak matrix majorization (English)
0 references
13 November 2007
0 references
Consider the space \(M_{nm}\) of real \(n\times m\) matrices. For two matrices \(X,Y\in M_{nm}\) we say \(X\) is \textit{weakly matrix majorized} by \(Y\) (and write \(X\prec Y\)) if there exists a row stochastic matrix \(R\) such that \(X=RY\). We also write \(X\sim Y\) if \(X\prec Y\) and \(Y\prec X\). A mapping \(T:M_{nm} \to M_{nm}\) is called a \textit{strong preserver} of \(\prec\) if \(TX\prec TY\) \(\Leftrightarrow\) \(X\prec Y\). Two strong preservers \(T_{1}\) and \(T_{2}\) are equivalent if \(T_{1}X\sim T_{2}X\) for all \(X\in M_{nm}.\) For each \(m\)-vector \(x\) let \(x^{(n)}\in M_{nm}\) be the matrix with all rows equal to \(x\). If \(T\) is a strong preserver of \(\prec\) then for each \(m\)-vector \(x\), there exists an \(m\)-vector \(y\) such that \(Tx^{(n)}=y^{(n)}\), and the \textit{border operator }for \(T\) is the function \(S:\mathbb{R} ^{m}\to\mathbb{R}^{m}\) such that \(Tx^{(n)}=(Sx)^{(n)}\). The main results of this paper are the following. If \(T\) is a surjective strong preserver with \(T0=0\) and \(n\geq2\), then the border operator is a linear operator (but the conclusion may be false if \(n=1\)). For any strong preserver \(T\) there exists an invertible \(m\times m\) matrix \(K\) such that \(TX-T0\sim XK\) for all \(X\), so \(T-T0\) is equivalent to a linear strong preserver. An alternative proof is also given to a theorem of \textit{A. M. Hasani} and \textit{M. Radjablipour} [Electronic J. Linear Algebra 15, 260--268 (2005)]: If \(T\) is a linear strong preserver then there exists an \(n\times n\) permutation matrix \(P\) and an \(m\times m\) invertible matrix \(L\) such that \(TX=PXL\) for all \(X\).
0 references
nonlinear preserver
0 references
linear preserver
0 references