An Accelerated Randomized Kaczmarz Algorithm
Abstract
The randomized Kaczmarz () algorithm is a simple but powerful approach for solving consistent linear systems . This paper proposes an accelerated randomized Kaczmarz () algorithm with better convergence than the standard algorithm on ill conditioned problems. The per-iteration cost of and are similar if is dense, but is much more able to exploit sparsity in than is . To deal with the sparse case, an efficient implementation for , called , is proposed. A comparison of convergence rates and average per-iteration complexities among , , and is given, taking into account different levels of sparseness and conditioning. Comparisons with the leading deterministic algorithm --- conjugate gradient applied to the normal equations --- are also given. Finally, the analysis is validated via computational testing.
Cite
@article{arxiv.1310.2887,
title = {An Accelerated Randomized Kaczmarz Algorithm},
author = {Ji Liu and Stephen J. Wright},
journal= {arXiv preprint arXiv:1310.2887},
year = {2014}
}