J. Korean Math. Soc. 2018; 55(6): 1529-1540
Online first article August 10, 2018 Printed November 1, 2018
https://doi.org/10.4134/JKMS.j170809
Copyright © The Korean Mathematical Society.
Hyun-Min Kim, Young-jin Kim, Jie Meng
Pusan National University, National Institute for Mathematical Sciences, Pusan National University
The matrix equation $X^p + {A^*}XA=Q$ has been studied to find the positive definite solution in several researches. In this paper, we consider fixed-point iteration and Newton's method for finding the matrix $p$-th root. From these two considerations, we will use the Newton-Schulz algorithm (N.S.A). We will show the residual relation and the local convergence of the fixed-point iteration. The local convergence guarantees the convergence of N.S.A. We also show numerical experiments and easily check that the N.S. algorithm reduce the CPU-time significantly.
Keywords: fixed-point iteration, Newton's method, Newton-Schulz algorithm, local convergence
MSC numbers: 65H10
2020; 57(5): 1119-1133
2016; 53(4): 781-793
2014; 51(6): 1155-1175
2014; 51(5): 955-970
© 2022. The Korean Mathematical Society. Powered by INFOrang Co., Ltd