Multilevel Monte Carlo methods for random nonlinear eigenvalue problems (Q7409151)

This paper proposes a multilevel Monte Carlo estimator that computes statistical quantities of interest from solutions to parameter-dependent nonlinear eigenvalue problems by introducing new maps that remain well-defined in the presence of crossings and bifurcations, without tracking eigenvalues or eigenvectors across samples or discretization levels. This approach is demonstrated using quasi-Monte Carlo on a random domain Laplace Dirichlet problem.