The skew-symmetric ortho-symmetric solutions of the matrix equations \(A^*XA=D\) (Q2911210)

A paradigm from linear algebra associates to \(A^{T}Ax=b\) the least squares method (regarded as a minimization problem). In practice, (e.g. signals), such approach may be obsolete. For the recovering of a linear system due to incomplete data or revising given data, the authors introduce \(A^{\ast}XA=D\).NEWLINENEWLINE{Problem 1.} Given \(A\in C^{n\times m}\), \(D\in C^{m\times m}\), find a skew-symmetric ortho-symmetric matrix \(X\) such that \(A^{\ast}XA=D\).NEWLINENEWLINEFollowing the above paradigm, next the authors consider the optimal approximation associated with the first problem:NEWLINENEWLINE{Problem 2.} Find an admissible \(\widehat{X}\), when \(\widetilde{X}\in C^{m\times n}\) such that \(\left\| \widetilde {X}-\widehat{X}\right\| =\inf_{X\;\text{admissible }}\left\| \widetilde{X}-X\right\| \), where the admissible matrices are the solution of Problem 1.NEWLINENEWLINEIf Problem 1 is solvable then Problem 2 has a unique solution.