Representation of real normal \((T + H)\) matrices in the case where the skew-symmetric parts of both summands are skew-circulant matrices (Q887504)

From MaRDI portal
scientific article
Language Label Description Also known as
English
Representation of real normal \((T + H)\) matrices in the case where the skew-symmetric parts of both summands are skew-circulant matrices
scientific article

    Statements

    Representation of real normal \((T + H)\) matrices in the case where the skew-symmetric parts of both summands are skew-circulant matrices (English)
    0 references
    26 October 2015
    0 references
    The real normal \((T+H)\)-problem is the problem of describing normal matrices that can be written as a sum of a real Toeplitz matrix \(T\) and a real Hankel matrix \(H\). Let \(P_n\) be the \(n\times n\) backward identity matrix. Let \(A, B\) be arbitrary \(n\times n\) real Toeplitz matrices. Then \(BP_n\) is Hankel and \(R = A+BP_n\) is a matrix that can be written as a sum of a real Toeplitz matrix and a real Hankel matrix. If \(R\) is normal, then \(A\) and \(B\) satisfy the equation \[ A^TA - AA^T = [(A-A^T)B + B(A-A^T)]P_n. \tag{E1} \] \textit{V. N. Chugunov} [Zh. Vychisl. Mat. Mat. Fiz. 52, No. 2, 209--213 (2012; Zbl 1249.15036); translation in Comput. Math. Math. Phys. 52, No. 2, 198--202 (2012)] showed that the real normal \((T+H)\)-problem can be divided into four cases: {\parindent=6mm \begin{itemize} \item[(1)] \(A\) is symmetric. \item [(2)] \(A-A^T\) and \(B-B^T\) are both circulant matrices. \item [(3)] \(A-A^T\) and \(B-B^T\) are both skew-circulant matrices. \item [(4)] \(B\) is symmetric. \end{itemize}} In the first case, \(R\) is a solution to the matrix equation for any \(B\). The second case was solved by the author [Math. Notes 96, No. 2, 275--285 (2014; Zbl 1317.15026); translation from Mat. Zametki 96, No. 2, 294--305 (2014)]. This paper deals with the third case. Let \(A = C_1 + S_1, B = C_2 + S_2\), and \(S = S_1^T - S_1\), where \(C_1, C_2\) are circulant matrices with zero diagonal, and \(S_1, S_2\) are skew-circulant matrices. It is shown that the matrix equation ({E1}) is equivalent to the system \[ \begin{cases} (C_2+C_1P_n)S + S(C_2 + C_1P_n) = 0, \\ SS_2 = 0. \end{cases} \tag{E2} \] and that the first equation in ({E2}) is satisfied if and only if {\parindent=6mm \begin{itemize} \item[(1)] \(C_1 = C_2 = 0\), or \item [(2)] \(n\) is even, \(C_1 - \alpha I_n\) is a \((-\alpha, 0)\)-checkerboard matrix, and \(C_2\) is a \((0, \alpha)\)-checkerboard matrix. \end{itemize}} This result gives the complete solution to the third case of the real normal \((T+H)\)-problem. Finally, a construction of matrices \(A\) and \(B\) satisfying the second equation in the system ({E2}) is given.
    0 references
    0 references
    Toeplitz matrix
    0 references
    Hankel matrix
    0 references
    skew-circulant matrix
    0 references
    normal real matrix
    0 references
    lower-triangular matrix
    0 references
    checkerboard matrix
    0 references
    0 references
    0 references
    0 references