Golub–Kahan vs. Monte Carlo: a comparison of bidiagonlization and a randomized SVD method for the solution of linear discrete ill-posed problems

Abstract Randomized methods can be competitive for the solution of problems with a large matrix low rank. They also have been applied successfully to large-scale linear discrete ill-posed by Tikhonov regularization (Xiang and Zou in Inverse Probl 29:085008, 2013). This entails computation an approximation partial singular value decomposition A that is numerical The present paper compares randomized method Krylov subspace based on Golub–Kahan bidiagonalization respect accuracy computing time discusses characteristics make them well suited method.