Randomized implicitly restarted Arnoldi method for the non-symmetric eigenvalue problem (Q6902500)

This interesting paper studies a randomized implicitly restarted Arnoldi method for the non-symmetric eigenvalue problem. The context of the problem is the following. Consider the eigenvalue/eigenvector equation \(Au=\lambda u\) where \(A\in \mathbb R^{n\times n}\), \(u\in \mathbb R^n\), \(||u||=1\) and \(\lambda \in \mathbb C\). As is well known, the problem of finding the eigenvalues and/or eigenvectors of a given linear transformation has many applications in science and engineering.\N\NThis paper deals with the computation of a small subset of eigenpairs of a matrix \(A\) that may be large, sparse and nonsymmetric. The authors introduce a randomized algorithm for solving the nonsymmetric eigenvalue problem referred to as a randomized Implicitly Restarted Arnoldi method. Their method uses a sketch-orthogonal basis during the Arnoldi process while maintaining the Arnoldi relation and using in addition, a restarting scheme to focus on a certain part of the spectrum.\N\NThe paper is well written with a good set of references.