Stability of Degenerate Schrödinger Equation with Harmonic Method

2025

Read Stability of Degenerate Schrödinger Equation with Harmonic Method (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2509-05533.txt.

Introduction

A Schrödinger equation whose diffusion fades at a boundary point is studied by Khadidja Fekirini, Naima Louhibi, and Abbes Benaissa in a 2025 analysis paper. Such degenerate equations arise where a coefficient vanishes, making standard solution theory fail at exactly the point where control is applied.

The authors first establish that the problem, posed on an interval with mixed boundary conditions, is well-posed, meaning a unique solution depends sensibly on the data. They then prove stability, showing solutions decay at exponential and polynomial rates. A method based on harmonic analysis delivers the result. The authors describe the decay result as optimal for this setting.

Catalog record

YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE80 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2509-05533

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Stability of Degenerate Schrödinger Equation with Harmonic Method Abstract This paper is devoted to study the well-posedness and stability of degenerate Schrödinger equation with a boundary control acting at the degeneracy. First, we establish the well-posedness of the degenerate problem v t ​ ( x , t ) + ı ​ ( x α ​ v x ​ ( x , t ) ) x = 0 , with ​ x ∈ ( 0 , 1 ) v_{t}(x,t)+\imath(x^{\alpha}v_{x}(x,t))_{x}=0,\hbox{ with }x\in(0,1) , controlled by Dirichlet-Neumann conditions. Then, exponential and polynomial decreasing of the solution are established. This result is optimal and it is obtained using complex analysis method. Khadidja Fekirini, Naima Louhibi & Abbes Benaissa, Laboratory of Analysis and Control of PDEs, Djillali Liabes University, P. O. Box 89, Sidi Bel Abbes 22000, ALGERIA. E-mails: fekirini - [email protected] [email protected] benaissa - [email protected] AMS (MOS) Subject Classifications :35B40, 35Q41, 35P15. Key words and phrases: degenerate Schrödinger equation, exponential polynomial stability, semigroup theory, spectral analysis. 1 Introduction This paper is devoted to study the existence and stability of solutions of degenerate Schrödinger equation with a boundary control acting at the degeneracy in suitable sobolev spaces. More precisely, we consider the following system (1) \left\{\matrix{v_{t}(x,t)+\imath(x^{\alpha}v_{x}(x,t))_{x}=0\hfill&\hbox{ in }(0,1)\times(0,+\infty),\hfill&\cr(x^{\alpha}v_{x})(0,t)=\imath\rho\partial^{\tilde{\alpha},\eta}v(0,t)\hfill&\hbox{ on }(0,+\infty),\hfill&\cr v_{x}(1,t)=0\hfill&\hbox{ on }(0,+\infty),\hfill&\cr v(x,0)=v_{0}(x)\hfill&\hbox{ on }(0,1),\hfill&\cr}\right. where 0<\alpha<1,\rho>0 and the term \partial^{\tilde{\alpha},\eta} stands for the generalized fractional integral of order 0<\tilde{\alpha}\leq 1 (see [ 5 ] ), which is given by \partial^{\tilde{\alpha},\eta}w({t})=\left\{\matrix{w(t)\hfill&\hbox{ for }\tilde{\alpha}=1,\ \eta\geq 0,\hfill\cr\displaystyle\frac{1}{\Gamma(1-\tilde{\alpha})}\displaystyle\int_{0}^{{t}}({t}-s)^{-\tilde{\alpha}}e^{-{\eta}({t}-s)}w(s)\,ds,\hfill&\hbox{ for }0<\tilde{\alpha}<1,\ \eta\geq 0.\hfill\cr}\right. The controllability and stabilization of Schrödinger equations without degeneracies have attracted considerable attention over the past years. Under the so-called geometric control condition, it is proved by G. Lebeau [ 11 ] that the Schrödinger equation is exactly controllable for arbitrary short time. In [ 14 ] , Machtyngier addressed the exact controllability in H^{-1}(\Omega) , with \Omega is a bounded domain, where Dirichlet boundary condition in L^{2}(\Omega) . The approaches adapted are HUM (Hilbert Uniqueness Method) and multipliers techniques. The boundary stabilization of the Schrödinger equation has also received a lot of attention. For an introduction, see [ 10 ] , where Lasiecka and Triggiani examine solution existence, uniqueness, and uniform boundary stability at the energy level in L^{2}(\Omega) for the n-dimensional linear Schrödinger equation within a bounded open domain. This system is given by: \left\{\begin{array}[]{lr}u_{t}+i\Delta u=0&\Omega\times(0,+\infty),\\ \frac{\partial u}{\partial\nu}=iu&x\in\Gamma_{1},t\geq 0,\\ u=0&x\in\Gamma_{2},t\geq 0,\\ u(x,0)=u_{0}(x),u_{t}(x,0)=u_{1}(x)&x\in\Omega.\end{array}\right. The authors adopted semigroup theory to show the global existence of the system and thereafter determined an optimal decay result using the multiplier method. A similar study was accomplished in [ 15 ] , where iu was replaced with im(x)u_{t} . In this case, the authors proved exponential decay in both the L^{2} -norm and the H^{1} -norm by employing the same approach while imposing geometric control conditions on the boundary. In [ 9 ] , the problem treated is the following \quad\left\{\begin{array}[]{ll}\Phi_{t}(x,t)=-\imath\Phi_{xx}(x,t),&0<x<1,\ \ t>0,\\ \Phi(1,t)=u(t),&t\geq 0,\\ \Phi_{x}(0,t)=0,&t\geq 0,\\ y(t)=\Phi_{x}(1,t),&t\geq 0,\end{array}\right. where u(t) is the ”input” and y(t) is the ”output”. Two novel control designs are proposed to exponentially stabilize the system. Furthermore, S. Nicaise and S. Rebiai [ 18 ] examined the influence of time delays on the boundary and internal feedback stabilization of the multidimensional Schrödinger equation, which is usually used to model the behavior of quantum systems. Their study centred on how these time delays influence the stabilization process, focusing to provide insights into the dynamics and control of such systems across different contexts. The system is given by: \left\{\begin{array}[]{lr}y_{t}(x,t)+i\Delta y(x,t)=0,&x\in\Omega,t>0,\\ y(x,0)=y_{0}(x,t)&x\in\Omega,t>0,\\ y(x,t)=0&x\in\Gamma_{0},t>0,\\ \frac{\partial y}{\partial\nu}=i\mu_{1}y(x,t)+i\mu_{2}y(x,t-\tau),&\in\Gamma_{1},t>0,\\ y(x,t-\tau)=f_{0}(x,t-\tau)&x\in\Omega,0<t<\tau.\\ \end{array}\right. On the contrary, when the principal part is degenerate not much is known in the literature, despite that many problems that are pertinent for applications are modeled by Schrödinger equations degenerating at the boundary of the space domain. In [ 7 ] , the authors considered the following Schrödinger equation (2) \left\{\matrix{v_{t}(x,t)+\imath(x^{\alpha}v_{x}(x,t))_{x}=0\hfill&\hbox{ in }(0,1)\times(0,+\infty),\hfill&\cr v(0,t)=0\hfill&\hbox{ on }(0,+\infty),\hfill&\cr v_{t}(1,t)+v_{x}(1,t)+v(1,t)=0\hfill&\hbox{ on }(0,+\infty),\hfill&\cr v(x,0)=v_{0}(x)\hfill&\hbox{ on }(0,1),\hfill&\cr}\right. They proved that the solution decays exponentially in an appropriate energy space. Moreover, the degeneracy does not affect the decay rates of the energy. Here, the situation is different since we impose a damping at point x=0 , where the degeneracy of the elliptic operator (x^{\alpha}\partial_{x}v)_{x} holds, which turns out to be a more challenging issue. To our best knowledge, this is the first attempt to study the global decaying solutions for a degenerate Schrödinger equation under a control acting on the degenerate boundary. Moreover, the enegy method based on multiplier techniques used in [ 7 ] do not seem to be work in the case of a feedback acting at a degenerate point x=0 . In this work, we are interested in studying precisely this issue, extending the results obtained in [ 20 ] , where the authors discuss the same issue in the case of wave equations. We obtain new results on decay estimates depending on parameters \alpha and \tilde{\alpha} . This paper is organised as follows. In section 2, we give some preliminaries. In section 3, the well-posedness results of the system (\ref{P}) are given using semigroup theory. In section 4, we prove an asymptotic and polynomial decay using Borichev-Tomilov Theorem. In section 5, we prove lack of exponential stability using Rouché’s Theorem. of the obtained system (\ref{P}) and we prove an optimal decay rate. 2 Preliminaries In this section, we introduce notations, definitions and propositions that will be used later. First we introduce some weighted Sobolev spaces: H^{1}_{\alpha}(0,1)=\left\{v\in L^{2}(0,1),v\hbox{ is locally absolutely continuous in }(0,1],\ x^{\alpha/2}v_{x}\in L^{2}(0,1)\right\}, and H^{2}_{\alpha}(0,1)=\left\{v\in L^{2}(0,1),v\in H^{1}_{\alpha}(0,1),\ x^{\alpha}v_{x}\in H^{1}(0,1)\right\}, where H^{1}(0,1) represent the classical Sobolev space. We remark that H^{1}_{\alpha}(0,1) is a Hilbert space with the scalar product (u,v)_{H_{\alpha}^{1}(0,1)}=\displaystyle\int_{0}^{1}(u\overline{v}+x^{\alpha}u^{\prime}(x)\overline{v^{\prime}(x)})\,dx,\quad\forall u,v\in H_{\alpha}^{1}(0,1). Remark 2.1 we have that \matrix{v^{2}(1)&=&\displaystyle\int_{0}^{1}(xv^{2}(x))_{x}\,dx=\displaystyle\int_{0}^{1}(v^{2}(x)+2xv(x)v^{\prime}(x))\,dx\hfill\cr&\leq&2\displaystyle\int_{0}^{1}v^{2}(x)\,dx+\displaystyle\int_{0}^{1}x^{2}(v^{\prime}(x))^{2}\,dx\hfill\cr&\leq&2\displaystyle\int_{0}^{1}v^{2}(x)\,dx+\displaystyle\int_{0}^{1}x^{\alpha}(v^{\prime}(x))^{2}\,dx\leq 2\|v\|_{H_{\alpha}^{1}(0,1)}^{2}.\hfill\cr} Moreover, we have \matrix{|v(x)|&=&\left|-\displaystyle\int_{x}^{1}v^{\prime}(x)\,dx+v(1)\right|\hfill\cr&\leq&\displaystyle\int_{0}^{1}|v^{\prime}(x)|\,dx+|v(1)|\hfill\cr&\leq&\left(\displaystyle\int_{0}^{1}\frac{1}{x^{\alpha}}\,dx\right)^{1/2}\left(\displaystyle\int_{0}^{1}x^{\alpha}|v^{\prime}(x)|^{2}\,dx\right)^{1/2}+|v(1)|\hfill\cr&\leq&\left(\frac{1}{\sqrt{1-\alpha}}+\sqrt{2}\right)\|v\|_{H_{\alpha}^{1}(0,1)}.\hfill\cr} Hence H_{\alpha}^{1}(0,1)\hookrightarrow C([0,1]). 2.1 Augmented model In this section we reformulate (\ref{P}) into an augmented system. For that, we need the following proposition. Proposition 2.1 (see [ 16 ] ) Let \mu be the function: (3) \mu(\xi)=|\xi|^{(2\tilde{\alpha}-1)/2},\quad-\infty<\xi<+\infty,\ 0<\tilde{\alpha}<1. Then the relationship between the ‘input’ U and the ‘output’ O of the system (4) \partial_{t}\phi(\xi,t)+(\xi^{2}+\eta)\phi(\xi,t)-U(t)\mu(\xi)=0,\quad-\infty<\xi<+\infty,\eta\geq 0,t>0, (5) \phi(\xi,0)=0, (6) O(t)=(\pi)^{-1}\sin(\tilde{\alpha}\pi)\displaystyle\int_{-\infty}^{+\infty}\mu(\xi)\phi(\xi,t)\,d\xi, where U\in C^{0}([0,+\infty)) , is given by (7) O=I^{1-\tilde{\alpha},\eta}U, where [I^{\alpha,\eta}f](t)=\displaystyle\frac{1}{\Gamma(\tilde{\alpha})}\displaystyle\int_{0}^{t}(t-\tau)^{\tilde{\alpha}-1}e^{-\eta(t-\tau)}f(\tau)\,d\tau. Lemma 2.1 (see [ 1 ] ) If \lambda\in D_{\eta}={\bf\hbox{\sc I\hskip-7.0ptC}}\backslash]-\infty,-\eta] then \displaystyle\int_{-\infty}^{+\infty}\displaystyle\frac{\mu^{2}(\xi)}{\lambda+\eta+\xi^{2}}\,d\xi=\displaystyle\frac{\pi}{\sin\tilde{\alpha}\pi}(\lambda+\eta)^{\tilde{\alpha}-1}. Using now Proposition 2.1 and relation ( 7 ), system (\ref{P}) may be recast into the following augmented system [formula] \left\{\matrix{v_{t}(x,t)+\imath(x^{\alpha}v_{x}(x,t))_{x}=0,\hfill&\cr\phi_{t}(\xi,t)+(\xi^{2}+\eta)\phi(\xi,t)-v(0,t)\mu(\xi)=0,\hfill&-\infty<\xi<+\infty,\ \ \hfill t>0,\hfill&\cr(x^{\alpha}v_{x})(0,t)=i\zeta\displaystyle\int_{-\infty}^{+\infty}\mu(\xi)\phi(\xi,t)\,d\xi,\hfill&\hfill\cr v_{x}(1,t)=0,\hfill&\hfill\cr v(x,0)=v_{0}(x),\phi(\xi,0)=0,\hfill&\cr}\right. where \zeta=\varrho(\pi)^{-1}\sin(\tilde{\alpha}\pi) . We define the energy associated with the solution of the problem (\ref{P}) by (8) E(t)=\frac{1}{2}\int_{0}^{1}|v(x,t)|^{2}dx+\frac{\zeta}{2}\displaystyle\int_{-\infty}^{+\infty}|\phi(\xi,t)|^{2}\,d\xi. Lemma 2.2 The energy functional defined by (\ref{27}) decays as follows E^{\prime}(t)=-\zeta\displaystyle\int_{-\infty}^{+\infty}(\xi^{2}+\eta)|\phi(\xi,t)|^{2}d\xi\leq 0. Proof. Multiplying the equation of (\ref{10}) by \overline{v} , integrating over (0,1) , applying integration by parts and using boundary conditions we obtain (9) \int_{0}^{1}v_{t}\overline{v}\ dx=-\zeta\displaystyle\int_{{\bf\hbox{\sc I\hskip-2.0ptR}}}\mu(\xi)\phi(\xi)d\xi\overline{v}(0,t)+\imath\int_{0}^{1}x^{\alpha}|v_{x}|^{2}\ dx, Multiplying the second equation in ( 1 ) by \zeta\overline{\phi} and integrating over (-\infty,+\infty) , we get: (10) \ \ \ \ \displaystyle\frac{\zeta}{2}\displaystyle\frac{d}{dt}\|\phi\|^{2}_{L^{2}({\bf\hbox{\sc I\hskip-2.0ptR}})}+\zeta\displaystyle\int_{{\bf\hbox{\sc I\hskip-2.0ptR}}}(\xi^{2}+\eta)||\phi(\xi,t)|^{2}d\xi-\zeta\Re\displaystyle\int_{{\bf\hbox{\sc I\hskip-2.0ptR}}}\mu(\xi)\ \overline{\phi}(\xi,t)d\xi v(0,t)=0. We take the sum of ( 10 ) and the real part of ( 9 ), we get E^{\prime}(t)=-\zeta\displaystyle\int_{{\bf\hbox{\sc I\hskip-2.0ptR}}}(\xi^{2}+\eta)||\phi(\xi,t)|^{2}d\xi\leq 0. Remark 2.2 In the case \tilde{\alpha}=1 , we take \varrho v(0,t) instead of \varrho\partial^{\alpha,\eta}v(0,t) . We do not need to introduce an augmented system. In this case the operator {\cal A} takes the form (11) {\tilde{\cal A}}v=-i(x^{\alpha}u_{x})_{x} with domain (12) D({\tilde{\cal

Read the complete text