Elliptic equations in divergence form with partially BMO coefficients (Q982277)

From MaRDI portal
scientific article
Language Label Description Also known as
English
Elliptic equations in divergence form with partially BMO coefficients
scientific article

    Statements

    Elliptic equations in divergence form with partially BMO coefficients (English)
    0 references
    0 references
    0 references
    6 July 2010
    0 references
    The authors study the solvability of elliptic operators in divergence form \[ {\mathcal L}u=(a^{ij}u_{x^i}+a^ju)_{x^j}+b^iu_{x^i}+cu\tag{1} \] in Sobolev spaces with rough leading coefficients. First of all, it is assumed that all the coefficients are bounded and measurable, and \(a^{ij}\) are uniformly elliptic. Note that it is not assumed that the matrix \(\{a^{ij}\}\) is symmetric. The cases of the whole space, a half-space, and a Lipschitz domain with a small Lipschitz constant are treated in the paper. For example, in the case of the whole space, it is assumed that the coefficients \(a^{ij}\) are only measurable in \(x^1\) direction and BMO in the other directions with locally small BMO semi-norms. For equations in a Lipschitz domain, additionally it is assumed that \(a^{ij}\) have small BMO semi-norms in a neighborhood of the boundary of the domain. All rigorous restrictions on the coefficients are presented in the paper. Under these assumptions, the authors establish unique \(W^1_p\) solvability of (1). They present a unified approach of both the Dirichlet boundary problem and the conormal derivative problem in a half-space and in a bounded Lipschitz domain. The authors also investigate elliptic equations in Sobolev spaces with mixed norms under the same assumptions on the coefficients. Let us formulate, with more details, only one result of the paper, for the Dirichlet problem on a bounded domain. The following problem is considered \[ \begin{cases} {\mathcal L}u=\operatorname{div}g+f &\text{in }\Omega,\\ u=0 &\text{on }\partial\Omega.\end{cases}\tag{2} \] Let \(p\in (1,\infty)\) and \(\Omega\) be a bounded domain in \(\mathbb R^d\). It is assumed that \(a^i u_{x^i}+c\leq 0\) in \(\Omega\) in the weak sense. Then, under some suitable assumptions on the coefficients of (1) and the boundary \(\partial\Omega\), for any \(f,\;g=(g_1,\dots,g_d)\in L_p(\Omega)\), there exists a unique \(u\in W^1_p(\Omega)\) satisfying (2). Moreover, the estimate \[ \|u\|_{W^1_p(\Omega)}\leq N\|f\|_{L_p(\Omega)}+N\|g\|_{L_p(\Omega)} \] holds, where the constant \(N\) is independent of \(f, g\), and \(u\). In the whole space and a half-space, the equation has the form \({\mathcal L}u-\lambda u= \operatorname{div}g+f\). Then the estimate is made precise with respect to \(\lambda\), too.
    0 references
    0 references
    elliptic equations
    0 references
    divergence form
    0 references
    BMO coefficients
    0 references
    Lipschitz domain
    0 references
    Dirichlet problem
    0 references
    conormal derivative problem
    0 references
    mixed norms
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references

    Identifiers

    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references