An inverse problem for the magnetic Schrödinger equation in infinite cylindrical domains (Q1626588)

The authors analyze an inverse problem for the Schrödinger equation \( (i\partial _{t}+\Delta _{A}+q)u=0\) posed in \((0,T)\times \Omega \) where \( \Omega =\omega \times \mathbb{R}\), \(\omega \) being a bounded and simply connected domain of \(\mathbb{R}^{2}\) with \(C^{2}\) boundary \(\partial \omega \) . The initial condition \(u(0,.)=0\) is posed in \(\Omega \) and the Dirichlet boundary condition \(u=f\) is imposed on \(\Sigma =(0,T)\times (\partial \omega \times \mathbb{R})\). Here \(\Delta _{A}\) is the Laplace operator associated to the magnetic field \(A=(a_{j})_{j}\in W^{1,\infty }(\Omega )^{3}\) through \( \Delta _{A}=\sum_{j}(\partial _{x_{j}}+ia_{j})^{2}\), \(q\in L^{\infty }(\Omega )\) and \(f\in L^{2}(\Sigma )\). The authors define the Dirichlet-to-Neumann operator as \(\Lambda _{A,q}(f)=(\partial _{\nu }+iA\cdot \nu )u\). The purpose of the paper is to prove existence and uniqueness results for the inverse problem which consists to determine the electric potential \(q\) and the gauge class of \(A\) from the Dirichlet-to-Neumann operator \(\Lambda _{A,q}\). Two magnetic fields \( A_{j}\in W^{1,\infty }(\Omega )^{3}\) are gauge equivalent if there exists \( \psi \in W^{2,\infty }(\Omega )\) such that \(\psi \mid _{\partial \omega \times \mathbb{R}}=0\) and \(A_{2}=A_{1}+\nabla \psi \). The authors first define the notion of solution to the Schrödinger problem in the transposition sense if \(u\in L^{\infty }(0,T;H^{-1}(\Omega ))\) and \( \left\langle u,F\right\rangle _{L^{\infty }(0,T;H^{-1}(\Omega )),L^{1}(0,T;H_{0}^{1}(\Omega ))}=\left\langle f,\partial _{\nu }v\right\rangle _{L^{2}(\Sigma )}\) for every \(F\in L^{1}(0,T;H_{0}^{1}(\Omega ))\) and \(v\in C(0,T;H^{1}(\Omega ))\) being the unique solution to the problem \((i\partial _{t}+\Delta _{A}+q)v=F\) in \( (0,T)\times \Omega \), \(v(T,.)=0\) in \(\Omega \) and \(v=0\) on \(\Sigma \). The authors prove that for every \(M>0\), \(A\in W^{1,\infty }(\Omega ,\mathbb{R} )^{3}\), \(q\in W^{1,\infty }(\Omega ,\mathbb{R})\) satisfying \(\left\| A\right\| _{W^{1,\infty }(\Omega ,\mathbb{R}^{3})}+\left\| q\right\| _{W^{1,\infty }(\Omega )}\leq M\), and \(f\in H_{0}^{2,1}(\Sigma ) \) the first Schrödinger problem has a unique solution \(u\in H^{1}(0,T;H^{1}(\Omega ))\) in the transposition sense which satisfies \( \left\| u\right\| _{H^{1}(0,T;H^{1}(\Omega ))}\leq C\left\| f\right\| _{H^{2,1}(\Sigma )}\) and \(\left\| \partial _{\nu }u\right\| _{L^{2}(\Sigma )}\leq C\left\| f\right\| _{H^{2,1}(\Sigma )}\) for a constant \(C\) which depends on \(M\) and on the data of the problem. The authors start proving the existence of a unique solution to the second Schrödinger problem using the framework of m-dissipative operators in an appropriate Banach space. They then prove that the linear map \(F\rightarrow \partial _{\nu }v\) is bounded from \(L^{1}(0,T;H_{0}^{1}( \Omega ))\) to \(L^{2}(\Sigma )\) and they conclude representing \(u\) in terms of a solution \(v\). For the inverse problem, the authors introduce the class of admissible magnetic potentials as \(\mathcal{A}=\{A=(a_{i});\) \( a_{i},a_{2}\in L_{x_{3}}^{\infty }(\mathbb{R},H_{0}^{2}(\omega ))\cap W^{2,\infty }(\Omega \omega );\) \(a_{3}\in C^{3}(\overline{\Omega })\), \( \sup_{x\in \Omega }\sum_{\alpha }\left\langle x_{3}\right\rangle ^{d}\left| \partial _{x}^{\alpha }a_{3}(x)\right| <\infty \) for some \(d>1\), \(\partial _{x}^{\alpha }a_{3}(x)=0\) on \(\partial \Omega \) for some \( \alpha \}\). The authors first prove a uniqueness result for the 2-form \( \mathrm{d}A=\frac{1}{2}\sum_{ij}(\partial _{x_{j}a_{i}}-\partial _{x_{i}a_{j}})\mathrm{d}x_{j}\wedge \mathrm{d}x_{i}\) and the electric potential \(q\). Under further hypotheses on the difference \(A_{1}-A_{2}\) between two magnetic fields the authors prove that \(\mathrm{d}A_{1}=\mathrm{d }A_{2}\). In both cases, they also prove estimates on the difference \(\mathrm{ d}A_{1}-\mathrm{d}A_{2}\) in terms of the difference between the two associated Dirichlet-to-Neumann operators. They finally prove estimates on the differences \(A_{1}-A_{2}\) and \(q_{1}-q_{2}\) in terms of the difference between the two associated Dirichlet-to-Neumann operators. For the proofs of these results, the authors introduce the notion of GO solution. They introduce a magnetic potential mollification and they use Fourier transform.