rkma

This repository contains the MATLAB code that generates the figures in the paper "The Randomized Kaczmarz Method with Mismatched Adjoint".

Authors:

• Dirk Lorenz (d.lorenz@tu-braunschweig.de)
• Sean Rose (seanrose949@gmail.com)
• Frank Schöpfer (frank.schoepfer@uni-oldenburg.de

Contents

Drivers (run these to generate figures):

example_convergence_perturbation.m  script to generate figure 2
example_inconsistent.m              script to generate figure 3
example_underdetermined.m           script to generate figure 4
example_ct.m                        script to generate figure 5
example_optimize_probabilities.m    script to generate figure 6

Routines called by the drivers:

rkma.m                      function for the randomized Kaczmarz
sampling.m                  function to sample according to a prob-vector

Dependencies

example_optimize_probabilities.m needs projsplx.m from http://ufdc.ufl.edu/IR00000353/.

example_ct.m needs the AIRtool package https://doi.org/10.1016/j.cam.2011.09.039

To use the writeout functionality requires matlab2tikz from https://github.com/matlab2tikz/matlab2tikz