On convergence of EVHSS iteration method for solving generalized saddle-point linear systems (Q1726426): Difference between revisions

Based on \textit{J.-L. Zhang}'s publication [Numer. Linear Algebra Appl. 25, No. 4, e2166, 14 p. (2018; Zbl 06945801)], the author proves that the iteration method is convergent unconditionally to the unique solution of the linear system \[ Ax=\left ( \begin{matrix} B&E \\ -E^{\ast} &C \\ \end{matrix} \right ) \left ( \begin{matrix} y \\ z \\ \end{matrix} \right )=\left ( \begin{matrix} f \\ g \\ \end{matrix} \right )=b, \] where \(B\) and \(C\) are Hermitian positive semidefinite matrices, and \(E\) is a rectangular matrix. With a precise technique for the spectral radius \(\rho\) of the iteration matrix the bound \(\rho<1\) is proven.