Iterative Solvers: Eigenvalue Problems

This lecture explores iterative solvers for eigenvalue problems, focusing on methods based on matrix-vector multiplication and preconditioning. The instructor discusses user criteria, convergence metrics, and the comparison of different eigensolvers based on theoretical and practical criteria. Various user criteria challenges are addressed, including residual tolerance and low rank approximations. The lecture also covers the choice of methods such as Restarted Krylov and Single vector methods. Additionally, it delves into the importance of convergence metrics, efficiency, and robustness in solving eigenvalue problems iteratively.