The div-curl lemma ``trente ans après'': an extension and an application to the \(G\)-convergence of unbounded monotone operators (Q1019133)

From MaRDI portal
scientific article
Language Label Description Also known as
English
The div-curl lemma ``trente ans après'': an extension and an application to the \(G\)-convergence of unbounded monotone operators
scientific article

    Statements

    The div-curl lemma ``trente ans après'': an extension and an application to the \(G\)-convergence of unbounded monotone operators (English)
    0 references
    0 references
    0 references
    0 references
    28 May 2009
    0 references
    The standard (or one of the standard) statement regarding the result known as ``compensated compactness'' or ``div-curl lemma'' is the following: given an open subset of \(\Omega\subset {\mathbb R}^n\) and two sequences \(\{v_h\}_{h \in {\mathbb N}}, \{w_h\}_{h \in {\mathbb N}} \subset (L^2(\Omega))^n\) weakly converging in \((L^2(\Omega))^n\) respectively to \(v\) and \(w\) such that \[ \{\text{div} \, v_h \}_h \text{ compact in } H^{-1}(\Omega), \qquad \{\text{curl} \, w_h \}_h \text{ compact in } (H^{-1}(\Omega))^{n \times n} \] then \(v_h \cdot w_h\) is weakly compact in \({\mathcal D}'(\Omega)\) and \(\lim_h v_h \cdot w_h = v \cdot w\). This result was born first to study \(G\)-convergence and homogenization of elliptic equations in divergence form, but then it found other applications. In the present paper the authors consider a case where the sequence \(v_h \cdot w_h\) is not necessarily a sequence in \(L^1(\Omega)\), but it may be a distribution. The situation considered is the following: \[ v_h \in L^p(\Omega), \;w_h \in L^q(\Omega), \;1 \leq p,q < +\infty, \;1 \leq \tfrac{1}{p} + \tfrac{1}{q} \leq 1 + \tfrac{1}{n} \] with (\(p'\) and \(q'\) the conjugate esponents of \(p\) and \(q\)) \[ \{\text{div} \, v_h \}_h \text{ compact in } W^{-1,q'}(\Omega), \qquad \{\text{curl}\, w_h \}_h \text{ compact in } (W^{-1,p'}(\Omega))^{n \times n}. \] The limit of \(v_h \cdot w_h\) in \({\mathcal D}'(\Omega)\), up to a subsequence, may be of the form \[ v \cdot w + \sum_{k=1}^{\infty} \text{div} (r_k \delta_{x_k}) \] with \(r_k \in {\mathbb R}^n\), \(x_k \in \Omega\). In the last section the authors give an application to \(G\)-convergence for nonlinear operators of the form \(u \mapsto - \text{div} ( a(x, \nabla u))\) whose model example may be \(a(x,\xi) = |A(x) \xi|^{n-2} A^T(x) A(x) \xi\), \(A\) matrix whose entries are \(L^{\infty}(\Omega)\) and \(A^T A \geq I_n\).
    0 references
    0 references
    0 references
    0 references
    0 references
    compensated compactness
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references