A family of higher-order convergent iterative methods for computing the Moore-Penrose inverse (Q426929): Difference between revisions

The paper describes an iterative method for computing the Moore-Penrose inverse that is an extension of the method by \textit{W. Li} and \textit{Z. Li} [ibid. 215, No. 9, 3433--3442 (2010; Zbl 1185.65057)]. The paper is short, well-written, and relatively clear. It contains basic concept of the method, three auxiliary lemmas, and the main theorem proving the convergence. The promising numerical experiments are performed for at most \(30\times 30\) matrices. On the other hand, not all is written in the paper. To understand some parameters of the basic iterative scheme, it is necessary to see the original paper by Li and Li. It seems also that the method is not suitable for large-scale problems, since each iteration requires multiplications of matrices -- the time consuming operation.