Crossing of eigenvalues: measure type estimates (Q2455602)

From MaRDI portal
Revision as of 19:22, 4 August 2023 by Importer (talk | contribs) (‎Created a new Item)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
scientific article
Language Label Description Also known as
English
Crossing of eigenvalues: measure type estimates
scientific article

    Statements

    Crossing of eigenvalues: measure type estimates (English)
    0 references
    0 references
    0 references
    0 references
    25 October 2007
    0 references
    Let \(\Omega\) be a bounded, sufficiently smooth domain, \(\varrho\) a positive function defined on \(\Omega\), and consider the non-homogeneous elliptic operator \(\mathcal{W}(\varrho)\psi = \mathrm{div}(\varrho\nabla\psi)\) with suitable a Neumann boundary condition \(\partial_n \psi=0\) on \(\partial\Omega\). There is an orthonormal basis of eigenvectors in \(L^2(\Omega)\) and the eigenvalue sequence \((\lambda_n)_n\) satisfies \(\lim_{n\to\infty} | \lambda_n| =\infty\). The aim of the paper is to obtain measure type estimates for the neighborhoods of the resonant set. This purpose requires careful studies of the eigenvalues and eigenvectors' derivatives with respect to \(\varrho\). More precisely, starting with the manifold \(\Sigma_{j,k,\ell}=\{\varrho: \lambda_j(\varrho)+\lambda_k(\varrho)=\lambda_\ell(\varrho)\}\), one is interested in its neighborhood \(\Sigma_{j,k,\ell}^{s,\varepsilon}=\{\varrho: d_s(\varrho, \Sigma_{j,k,\ell})<\varepsilon\}\), where the distance \(d_s\) is taken in the Besov spaces \(B^s(\Omega)=B^{s,\infty}_2(\Omega)\). This space is defined by the restriction to \(\Omega\) from the space of all \(u\in L_2\), which can be represented as \(u=\sum_n u_n\psi_n\) with coefficients \((n^s u_n)_n \in \ell_\infty\) and \((\psi_n)_n\) being a fixed orthonormal basis. Theorem~1.1 gives sufficient conditions such that \(\mu_{s,R}(\Sigma_{j,k,\ell}^{s,\varepsilon})< c \varepsilon\) in the case of simple eigenvalues. Here the measure \(\mu_{s,R}\) is introduced as a tensor product of the Lebesgue measure \(\mu\) on \(\mathbb{R}\) by \(\mu_{s,R}=\bigotimes_{n\geq 1} \left(\frac{n^s \mu}{2R}\right)\), where \(R>0\) is the radius of the ball \(B_s(R)= \{(u_n)_n: | u_n| \leq R n^{-s}\}\) in \(B^s(\Omega)\). It is normalized by \( \mu_{s,R}(B_s(R))=1\) and depends on the choice of the orthonormal basis \((\psi_n)_n\). Theorem~1.2 deals with the analogous situation in case of quasi-double eigenvalues. Sections~2 and 3 contain the proofs of Theorems~1.1 and 1.2, where the first step is always a corresponding result for quasi-diagonalization. There are further assertions on the regularity of the eigenvectors. The paper ends with some discussion of the multiple eigenvalue situation.
    0 references
    non-homogeneous elliptic operator
    0 references
    wave propagation
    0 references
    crossing eigenvalues
    0 references

    Identifiers