Eduardo M. Garau,Pedro Morin,Carlos Zuppa
Abstract:In this article we prove convergence of adaptive finite element methods for second order elliptic eigenvalue problems. We consider Lagrange finite elements of any degree and prove convergence for simple as well as multiple eigenvalues under a minimal refinement of marked elements, for all reasonable marking strategies, and starting from any initial triangulation.
What problem does this paper attempt to address?
The problem that this paper attempts to solve is the convergence of the Adaptive Finite Element Methods (AFEM) when solving second - order elliptic eigenvalue problems. Specifically, the author aims to prove that for second - order elliptic eigenvalue problems, AFEM can converge, and this convergence applies to simple eigenvalues and multiple eigenvalues, as well as all reasonable marking strategies, and can start from any initial triangulation.
### Specific Problem Description
Consider the following continuous eigenvalue problem:
\[
-\nabla \cdot (A \nabla u)=\lambda B u \quad \text{in } \Omega, \quad u = 0 \quad \text{on } \partial \Omega,
\]
where \(A\) and \(B\) are matrix functions and scalar functions satisfying certain assumptions, and \(\Omega\) is a bounded open set. The goal of this problem is to find the eigenvalue \(\lambda\in\mathbb{R}\) and the corresponding eigenfunction \(u\neq0\) such that the above equation holds.
### Steps of the Adaptive Finite Element Method
The Adaptive Finite Element Method approximates eigenvalues and eigenfunctions through the following cycle:
1. **Solve**: Solve the finite element solution on the current mesh.
2. **Estimate**: Calculate the a posteriori error estimator.
3. **Mark**: Mark the elements to be subdivided according to the error estimator.
4. **Refine**: Subdivide the current mesh to generate a new, finer mesh.
### Main Contributions of the Paper
1. **Convergence Proof**: The paper proves the convergence of the Adaptive Finite Element Method when solving second - order elliptic eigenvalue problems, which is an area that has not been fully studied before.
2. **Wide Applicability**: The proof applies not only to simple eigenvalues but also to multiple eigenvalues; it applies not only to specific marking strategies (such as Dörfler marking) but also to all reasonable marking strategies.
3. **Relaxed Requirements for the Initial Mesh**: The initial mesh does not need to be fine enough; as long as it can capture the discontinuity of \(A\), convergence can be guaranteed.
### Main Results
The main results of the paper are as follows:
- For the discrete eigenvalues \(\lambda_k\) and eigenfunctions \(u_k\) obtained by the adaptive algorithm, there exists an eigenvalue \(\lambda\) of the continuous problem such that
\[
\lim_{k\rightarrow\infty}\lambda_k=\lambda \quad \text{and} \quad \lim_{k\rightarrow\infty}\text{dist}_{H^1_0(\Omega)}(u_k, M(\lambda)) = 0,
\]
where \(M(\lambda)\) represents the set of all eigenfunctions corresponding to the eigenvalue \(\lambda\).
Through these results, the paper provides a solid theoretical basis for the application of the Adaptive Finite Element Method in eigenvalue problems.