The Moore-Penrose inverse of a companion matrix (Q426080): Difference between revisions

Let \(R\) be a ring with identity and endowed with an involution. Let \({\mathcal M}_{m \times n} (R)\) denote the set of all matrices with \(m\) rows and \(n\) columns, with entries coming from \(R\). Let \(*\) be the involution on \({\mathcal M}_{m \times n} (R)\) induced by the involution on \(R\). An element \(A \in {\mathcal M}_{m \times n} (R)\) is said to be Moore-Penrose invertible if there is \(X \in {\mathcal M}_{n \times m} (R)\) such that \(AXA=A; XAX=X; (AX)^*=AX\) and \((XA)^*=XA\). Such an \(X\) need not exist, in general. If it exists, it can be shown that it is unique and in that case we call it the Moore-Penrose inverse of \(A\) and denote it by \(A^{\dagger}\). \(A \in {\mathcal M}_{n \times n} (R)\) is said to be group invertible if there exists \(X \in {\mathcal M}_{n \times n} (R)\) such that \(AXA=A; XAX=X\) and \(AX=XA\). Again, such an \(X\) need not exist; if it exists it is unique and is denoted by \(A^{\#}\). Consider the \((n+1) \times (n+1)\) companion matrix \(M\) given by \(M=\left( \begin{matrix} 0 & a \\ I_n & {\mathbf b} \\ \end{matrix} \right)\), where \(a \in R\) and \({\mathbf b} \in R^n\). About fifteen years ago, characterizations for matrices \(M\) that are group invertible were presented by \textit{R. Puystjens} and \textit{R. E. Hartwig} [Linear Multilinear Algebra 43, No.1-3, 137--150 (1997; Zbl 0890.15003)]. In the present work, the author presents necessary and sufficient conditions for the existence of the Moore Penrose inverse of \(M\). These conditions are expressed in terms of certain elements of the ring \(R\) being units (a unit in a ring \(R\) is an invertible element in \(R\)).