On quasiconvex hulls in symmetric \(2\times 2\) matrices (Q873636)

From MaRDI portal
scientific article
Language Label Description Also known as
English
On quasiconvex hulls in symmetric \(2\times 2\) matrices
scientific article

    Statements

    On quasiconvex hulls in symmetric \(2\times 2\) matrices (English)
    0 references
    29 March 2007
    0 references
    The author is interested in the study of the quasiconvec hull of compact sets of symmetric \(2\times 2\) matrices; in particular, he is interested in situations where the quasiconvex hull can be separated into smaller independent pieces. The main result is a geometric criterion for the identity \((K_1\cup K_2)^{\text{qc}}=K_1^{\text{qc}}\cup K_2^{\text{qc}}\) to be true, where \(K^{\text{qc}}\) represents the quasiconvex hull of the matrix set \(K\). The quasiconvex hull is defined for compact sets of matrices, \(K\subset{\mathbb R}^{m\times n}\), as follows \[ K^{\text{qc}}=\{\overline{\nu}: \nu\text{ is a homogeneous gradient Young measure, } \operatorname{supp}\nu \subset K\}, \] where \(\overline\nu\) is the barycenter of \(\nu\); this definition, by duality with quasiconvex functions, is equivalent to \[ K^{\text{qc}}=\{A: f(A)\leq \sup_K f \text{ for all quasiconvex }f: {\mathbb R}^{m\times n}\to{\mathbb R}\}. \] The main result of the paper is that, if \(\nu\) is a homogeneous gradient Young measure with \(\operatorname{supp}\nu\subset K_1\cup K_2 \subset {\mathbb R}_{\text{sym}}^{2\times 2}\) and there exists \(B\in {\mathbb R}_{\text{sym}}^{2\times 2}\) with \(\det B <0\) such that \[ K_1\cup K_2\subset {\mathcal U}_B {A\in{\mathbb R}_{\text{sym}}^{2\times 2}:\det(A-tB)<0, \text{ for all } t\in {\mathbb R}}, \] with \(K_1\) and \(K_2\) contained in different connected components of \({\mathcal U}_B\), then \(\text{supp}\,\nu\subset K_1\) or \(\operatorname{supp}\nu\subset K_2\). The main tool in the proof is a kind of directional maximum principle for second order elliptic equations in the plane in non-divergence form with measurable coefficients. The elliptic operator, or more precisely the measurable coefficients of the elliptic operator are chosen in an appropriate way so that the directional maximum principle can be applied to a sequence \((D^2w_j)\) generating the homogeneous gradient Young measure \(\nu\) and then leading to the result.
    0 references
    0 references
    Gradient Young measures
    0 references
    Quasiconvex Hull
    0 references
    Maximum principle
    0 references
    0 references