An iterative algorithm for the least Frobenius norm Hermitian and generalized skew Hamiltonian solutions of the generalized coupled Sylvester-conjugate matrix equations (Q1656677): Difference between revisions

From MaRDI portal
Import240304020342 (talk | contribs)
Set profile property.
Set OpenAlex properties.
Property / full work available at URL
 
Property / full work available at URL: https://doi.org/10.1007/s11075-017-0423-9 / rank
 
Normal rank
Property / OpenAlex ID
 
Property / OpenAlex ID: W2762155494 / rank
 
Normal rank

Revision as of 00:21, 20 March 2024

scientific article
Language Label Description Also known as
English
An iterative algorithm for the least Frobenius norm Hermitian and generalized skew Hamiltonian solutions of the generalized coupled Sylvester-conjugate matrix equations
scientific article

    Statements

    An iterative algorithm for the least Frobenius norm Hermitian and generalized skew Hamiltonian solutions of the generalized coupled Sylvester-conjugate matrix equations (English)
    0 references
    0 references
    0 references
    10 August 2018
    0 references
    Consider a set of complex matrices \(E_{i}\) (\(m\times r\) matrices), \(A_{ij}\), \(C_{ij}\) (both \(m\times n\)) and \(B_{ij}\), \(D_{ij}\) (both \(n\times r\)) for \(i=1,\dots,s\) and \(j=1,\dots,\ell\). The linear equations \[\sum_{j=1}^{\ell }\left( A_{ij}X_{j}B_{ij}+C_{ij}\bar{X}_{j}D_{ij}\right) =E_{i}, \tag{*}\] with \(i=1,\dots,s\), in the unknown \(n\times n\) matrices \(X_{j}\) (\(j=1,\dots,\ell\)) form a set of generalized coupled linear matrix equations. If this set of equations has at least one solution, then the solution with least Frobenius norm is unique. Let \(J\) be a fixed real \(n\times n\) orthogonal matrix which is also skew-symmetric (necessarily this implies that \(n\) must be even). Then a complex \(n\times n\) matrix \(X\) is called generalized skew Hamiltonian with respect to \(J\) if \(JXJ=-X^{H}\). The class of all \(X\) which are both Hermitian and generalized skew Hamiltonian with respect to \(J\) is denoted by \(HSHC^{n\times n}\). The task of the present paper is to describe an iterative method for solving (*) using a modified conjugate gradient method. The solution depends on the initial values of \((X_{1},\dots,X_{\ell})\). It is shown that, if equations (*) are consistent, \(X_{i}\in HSHC^{n\times n}\), and exact arithmetic is used, then an exact solution to (*) is obtained in at most \(smr+1\) iterative steps (\(s,m,r\) are numbers defined in the paper). This solution also lies in \(HSHC^{n\times n}\). Furthermore, there is a simple way to choose the initial \(X_{i}\) so that the solution has the least Frobenius norm of any solution in \(HSHC^{n\times n}\). The paper concludes with some numerical examples which indicate that the described method is computationally better to the algorithm given in [\textit{M. Hajarian}, Math. Methods Appl. Sci. 37, No. 17, 2782--2798 (2014; Zbl 1323.65041)].
    0 references
    0 references
    generalized Sylvester-conjugate equations
    0 references
    generalized skew Hamiltonian matrix
    0 references
    least norm solution
    0 references

    Identifiers