Linear maps preserving the Lorentz spectrum of \(3 \times 3\) matrices (Q6197641): Difference between revisions
From MaRDI portal
Added link to MaRDI item. |
ReferenceBot (talk | contribs) Changed an Item |
||
Property / cites work | |||
Property / cites work: Second-order cone programming / rank | |||
Normal rank | |||
Property / cites work | |||
Property / cites work: Linear maps preserving Pareto eigenvalues / rank | |||
Normal rank | |||
Property / cites work | |||
Property / cites work: Generalized inverses. Theory and applications. / rank | |||
Normal rank | |||
Property / cites work | |||
Property / cites work: Linear maps preserving the Lorentz-cone spectrum in certain subspaces of \(M_n\) / rank | |||
Normal rank | |||
Property / cites work | |||
Property / cites work: On eigenvalues induced by a cone constraint. / rank | |||
Normal rank | |||
Property / cites work | |||
Property / cites work: On spectral maps induced by convex cones / rank | |||
Normal rank |
Latest revision as of 12:45, 28 August 2024
scientific article; zbMATH DE number 7821007
Language | Label | Description | Also known as |
---|---|---|---|
English | Linear maps preserving the Lorentz spectrum of \(3 \times 3\) matrices |
scientific article; zbMATH DE number 7821007 |
Statements
Linear maps preserving the Lorentz spectrum of \(3 \times 3\) matrices (English)
0 references
20 March 2024
0 references
Let \(A\) be a real \(n\times n\) matrix. Consider the Lorentz cone \(\mathcal K_n\) \[ \mathcal K_n=\left\{\left[\begin{smallmatrix}\xi\\ \mu\\ \end{smallmatrix}\right]\colon\; \xi\in\mathbb R^{n-1},\:\mu\in\mathbb R,\: ||\xi||_2\leqslant\mu\right\}. \] The eigenvalue complementarity problem for the Lorentz cone concerns finding \(\lambda\in\mathbb R\) and nonzero \(x\in\mathbb R^n\) such that \[ x\in\mathcal K_n,\quad (A-\lambda I)x\in\mathcal K_n,\quad x^T(A-\lambda I)x=0. \] We say that any \(\lambda\) satisfying the above conditions is a Lorentz eigenvalue of \(A\), whereas any \(x\) associated to it, is a Lorentz eigenvector. The set consisting of all Lorentz eigenvalues of a fixed \(A\) is called the Lorentz spectrum of \(A\), or shortly L-spectrum of \(A\), and is denoted by \(\sigma_L(A)\). The collection of points of \(\sigma_L(A)\) that are interior (resp., boundary) points of \(\mathcal K_n\) is called the interior (resp., boundary) Lorentz spectrum of \(A\), and is denoted \(\sigma_{int}(A)\) (\(\sigma_{bd}(A)\)). The authors consider linear preservers of L-spectra of matrices. The paper is kind of continuation of the paper [\textit{M. I. Bueno} et al., Banach J. Math. Anal. 15, No. 3, Paper No. 58, 20 p. (2021; Zbl 1472.15039)]. The main result states that if \(\phi\colon M_3\to M_3\) is a linear preserver of the Lorentz spectrum, then there exists an orthogonal matrix \(Q\in M_2\) such that \[ \phi(A)=\begin{bmatrix}Q&0\\ 0&1\\ \end{bmatrix}A\begin{bmatrix}Q^T&0\\ 0&1\\ \end{bmatrix}. \] Moreover, \[\sigma_{int}(A)=\sigma_{int}\left(\phi(A)\right) \quad\textrm{and}\quad \sigma_{bd}(A)=\sigma_{bd}\left(\phi(A)\right). \]
0 references
Lorentz cone
0 references
Lorentz eigenvalues
0 references
linear preservers
0 references
\(3 \times 3\) matrices
0 references