A uniqueness theorem for inverse problems in quasilinear anisotropic media. II (Q6066514)

From MaRDI portal
scientific article; zbMATH DE number 7777225
Language Label Description Also known as
English
A uniqueness theorem for inverse problems in quasilinear anisotropic media. II
scientific article; zbMATH DE number 7777225

    Statements

    A uniqueness theorem for inverse problems in quasilinear anisotropic media. II (English)
    0 references
    0 references
    0 references
    0 references
    13 December 2023
    0 references
    The forward problem is a conductivity-type equation, with an anisotropic conductivity \(A(x,u) = \gamma(x,u)A(x)\), \(A\) is positive definite, \(\gamma\) positive and both \(C^{2,\alpha}\). The inverse problem the article considers is recovering \(\gamma\) from the Dirichlet-to-Neumann (DN) map. The authors are out to improve their results in [ibid. 16, No. 6, 1571--1577 (2022; Zbl 1512.35664)], which assume a smooth domain and operator, whereas here only \(C^{2,\alpha}\) smoothness of the domain and the conductivity are required. The results apply in all dimensions \(d \ge 2\). The article [\textit{Z. Sun} and \textit{G. Uhlmann}, Am. J. Math. 119, No. 4, 771--797 (1997; Zbl 0886.35176)] establishes that there is an obstruction to injectivity: the operator is invariant with respect to diffeomorphisms that keep the boundary fixed. The first result is conditional on a conjecture in the linear case: if the only obstruction to injectivity in the linear case is the diffeomorphism that holds the boundary fixed, and if the anisotropic components of the conductivities agree, then equal DN maps imply that the solution-dependent components \(\gamma\) of the conductivities are also equal up to the obstruction. The second result replaces the reliance on the conjecture with the assumption that \(A^{-1}\) is conformally equivalent to the Euclidian metric. It generalizes the isotropic result in [\textit{Z. Sun}, Math. Z. 221, No. 2, 293--305 (1996; Zbl 0843.35137)]. The first result's proof uses linearization to reduce the problem to showing that a certain diffeomorphism is the identity map. To use a result of \textit{J. Lelong-Ferrand} [Mem. Cl. Sci., Collect. Octavo, Acad. R. Belg. 39, No. 5, 3--44 (1971; Zbl 0215.50902)] the authors need a compact manifold. To extend the diffeomorphism with sufficient regularity they use a result from [\textit{T. Liimatainen} and \textit{M. Salo}, Math. Res. Lett. 21, No. 2, 341--361 (2014; Zbl 1298.30020)]. The proof of the second result relies on a more straightforward calculation.
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    Dirichlet to Neumann map
    0 references
    quasilinear elliptic equation
    0 references
    linearization
    0 references
    Riemannian manifold
    0 references
    conformal diffeomorphism
    0 references
    0 references