Expansion of weighted pseudoinverse matrices with positive definite weights into matrix power products. iterative methods (Q2458108)

Let \(A\in \mathbb{R}^{m\times n}, X \in \mathbb{R}^{n\times m}, B\in \mathbb{R}^{m\times m}\) and \(C\in \mathbb{R}^{n\times n}\) be symmetric positive definite matrices. Then the weighted pseudoinverse matrix of \(A\) is defined as the unique matrix \(X\) satisfying \(AXA=A, XAX=X, (BAX)^t=BAX, (CXA)^t=CXA\) where \(t\) denotes the transpose operation. The authors study an expansion of weighted pseudoinverse matrices with positive definite weights into infinite matrix power products with negative exponents and arbitrary positive parameters. The influence of the parameters on the convergence rate is also investigated. They apply the expansion to analyze iterative methods for evaluating weighted pseudoinverse matrices and weighted normal pseudoinverse matrices with positive definite weighted and also solving constrained least-squares problems.