Rank structures preserved by the \(QR\)-algorithm: the singular case (Q818169)

In a previous paper the authors have shown that when the matrix \(A\) is nonsingular, the QR iterates share the same rank structure, which is defined through its blocks by Rank \(A_{k}(i_{k}:n,1:j_{k})\leq r_{k}\) where \(A_{k}=A-\lambda _{k}I.\) In this paper, it is shown that even when \(A\) is singular, they can still construct QR iterates that preserve the rank structure.