A generalized LSQR algorithm
Lothar Reichel, Qiang Ye · Numerical Linear Algebra with Applications · 2008
Abstract LSQR is a popular iterative method for the solution of large linear system of equations and least‐squares problems. This paper presents a generalization of LSQR that allows the choice of an arbitrary initial vector for the solution subspace. Computed examples illustrate the benefit of being able to choose this vector. Copyright © 2008 John Wiley & Sons, Ltd.