A note on parallel approximate pseudoinverse matrix techniques for solving linear least squares problems

摘要

A new parallel generic approximate sparse pseudoinverse matrix technique using a decoupled column-wise approach, based on modified row-threshold incomplete QR factorization techniques, is proposed. The explicit preconditioned conjugate gradient least squares method in conjunction with the new parallel generic approximate sparse pseudoinverse matrix technique is used for solving linear least square problems. Numerical results indicating the applicability and effectiveness of the proposed parallel generic approximate sparse pseudoinverse matrix techniques for solving various model problems are given. (C) 2020 Elsevier B.V. All rights reserved.