Comparative Analysis of Iterative Row-Action Methods for Solving Linear Systems

The Kaczmarz method and its randomized variations are efficient iterative algorithms for solving large-scale linear systems, especially when dealing with overdetermined and inconsistent systems. Various row sampling schemes can outperform the original and Randomized Kaczmarz methods for consistent systems, while the Conjugate Gradient method for Least-Squares problems overcomes all Kaczmarz variations for inconsistent systems.