On a formula of Thompson and McEnteggert for the adjugate matrix (Q2052175): Difference between revisions

From MaRDI portal
Import240304020342 (talk | contribs)
Set profile property.
Importer (talk | contribs)
Changed an Item
 
(One intermediate revision by one other user not shown)
Property / OpenAlex ID
 
Property / OpenAlex ID: W3208832534 / rank
 
Normal rank
Property / arXiv ID
 
Property / arXiv ID: 2103.13739 / rank
 
Normal rank

Latest revision as of 01:22, 19 April 2024

scientific article
Language Label Description Also known as
English
On a formula of Thompson and McEnteggert for the adjugate matrix
scientific article

    Statements

    On a formula of Thompson and McEnteggert for the adjugate matrix (English)
    0 references
    0 references
    0 references
    25 November 2021
    0 references
    Let \(A\in\mathcal{R}^{n\times n}\), where \(\mathcal{R}\) is a commutative ring with identity. Let \(\lambda_0\), \(v\), \(w\), and \(p\) be its eigenvalue, a corresponding right eigenvector, a corresponding left eigenvector, and the characteristic polynomial, respectively. The Thompson-McEnteggert formula (TM in what follows) states that \[ w^Tv\,\mathrm{Adj}\,(\lambda_0I-A)=p'(\lambda_0)vw^T. \] It was found by \textit{R. C. Thompson} and \textit{P. McEnteggert} in [Linear Algebra Appl. 1, 211--243 (1968; Zbl 0224.15012)], for the case \(\mathcal{R}=\mathbb{C}\) and \(A\) Hermitian. If \(\lambda_0\) is multiple and \(w^Tv=0\), then the TM formula is meaningless. If \(\lambda_0\) is simple, then it is meaningful. Assuming that \(\mathcal{R}\) is a field, the authors provide a generalized formula (denoted by gTM), that is meaningful for multiple \(\lambda_0\) too. If \(A\) is nonsingular, they also express the elementary divisors of \(\mathrm{Adj}\,(A)\) in terms of those of \(A\). Let \(A\in\mathbb{C}^{n\times n}\) be Hermitian. Let \(\lambda_i\), \(v_i\), and \(v_{ij}\) be its \(i\)-th largest eigenvalue, a corresponding unit eigenvector, and its \(j\)-th component, respectively. Omitting the \(j\)-th row and column from \(A\), let \(\mu_{jk}\) be the \(k\)-th largest eigenvalue of the matrix so obtained. The ``classical eigenvector-eigenvalue identity'' (denoted by EE), \[ |v_{ij}|^2\prod_{\substack{k=1\\ k\ne i}}^n(\lambda_i-\lambda_k)=\prod_{k=1}^n(\lambda_i-\mu_{jk}), \quad i,j=1,\dots,n, \] follows from the TM formula. So, the EE formula is meaningful if all eigenvalues of \(A\) are simple. Using the gTM formula, the authors generalize the EE formula to an identity that is meaningful also if \(A\) has defective eigenvalues with geometric multiplicity 1. The authors also present other applications of the TM and gTM formulas.
    0 references
    0 references
    adjugate
    0 references
    eigenvalues
    0 references
    eigenvectors
    0 references
    elementary divisors
    0 references
    rank-one matrices
    0 references
    0 references
    0 references