Core Concepts
The paper studies the parametric semilinear elliptic eigenvalue problem with an affine dependence on the parameters, focusing on the analyticity of the ground eigenpairs and their mixed derivatives. This allows for efficient uncertainty quantification using quasi-Monte Carlo methods.
Abstract
The paper investigates the properties of the ground state of a parametric semilinear elliptic eigenvalue problem with an affine dependence on the parameters. The key results are:
The ground eigenpair (λ(y), u(y)) is shown to be separately complex analytic with respect to the parameters y. This allows for taking arbitrarily high order derivatives of the eigenpair.
The smallest eigenvalue λ(y) is shown to be uniformly bounded and strictly positive, and there exists a uniform positive gap between the smallest eigenvalues of the related linear operators O(y) and T(y).
Using these properties, the paper derives upper bounds on the mixed derivatives of the ground eigenpair and the ground energy that have the same form as the recent results for the linear eigenvalue problem.
As an application, the paper considers the parameters as uniformly distributed random variables and estimates the expectation of the eigenpairs and the ground energy using a randomly shifted quasi-Monte Carlo lattice rule, showing a dimension-independent error bound.
The analysis involves new techniques to handle the nonlinearity, including the use of the implicit function theorem and careful investigation of the nonlinear eigenvalue problem.
Stats
The following sentences contain key metrics or figures:
The smallest eigenvalue λ(y) is uniformly bounded away from 0.
There exists a constant CT > 0 such that λT(y) - λ(y) ≥ CT for all y ∈ U.