The inverse obstacle problem for nonlinear inclusions (Q6994470)

The authors consider the following nonlinear PDE:\N\[\N\begin{cases} \N\nabla\cdot(\gamma_{\mathrm{BG}}(x)\nabla u_{A}(x))=0 & \text{in }\Omega\setminus A, \\\N\nabla\cdot\left(\gamma_{\mathrm{NL}}(x,|\nabla u_{A}(x)|)\nabla u_{A}(x)\right)=0 & \text{in }A, \\\Nu_{A}(x) = f(x) & \text{on }\partial\Omega, \\\N\gamma_{\mathrm{BG}}(x)\partial_{\nu}u_{A}(x^{+}) = \gamma_{\mathrm{NL}}\left(x,|\nabla u_{A}(x^{-})|\right)\partial_{\nu}u_{A}(x^{-}) & \text{on }\partial A, \\\Nu_{A}(x^{+}) = u_{A}(x^{-}) & \text{on }\partial A, \N\end{cases}\N\]\Nwhere \(\Omega\) is a bounded domain in \(\mathbb{R}^{n}\) with \(n\ge 2\), filled with a material characterized by \(\gamma_{\mathrm{BG}}\), and \(A\) is a region properly contained in \(\Omega\), occupied by an anomaly described by \(\gamma_{\mathrm{NL}}\). Here, for each \(x\in\partial A\), the traces \(u_{A}(x^{\pm})\) and the normal derivatives \(\partial_{\nu}u_{A}(x^{\pm})\) can be formally understood as\N\[\N\begin{aligned} \N& u_{A}(x^{+}) = \lim_{\Omega\setminus A\ni y\rightarrow x} u_{A}(y), && \partial_{\nu}u_{A}(x^{+}) = \lim_{\Omega\setminus A\ni y\rightarrow x} \nu(x)\cdot\nabla u_{A}(y), \\\N& u_{A}(x^{-}) = \lim_{A\ni y\rightarrow x} u_{A}(y), && \partial_{\nu}u_{A}(x^{-}) = \lim_{A\ni y\rightarrow x} \nu(x)\cdot\nabla u_{A}(y), \N\end{aligned}\N\]\Nwhere \(\nu=\nu(x)\) is the unit outward normal vector at each \(x\in\partial A\).\N\NThe paper studies a nonlinear inverse obstacle problem: reconstructing the anomaly \(A\) from suitable boundary measurements on \(\partial\Omega\), specifically using the average Dirichlet-to-Neumann operator, under the assumption that the material properties \(\gamma_{\mathrm{BG}}\) and \(\gamma_{\mathrm{NL}}\) are known. The authors employ a non-iterative reconstruction method based on the monotonicity principle.