A miscellanea of qualitative and symmetry properties of the solutions to the two-phase Serrin’s problem
Lorenzo Cavallina · 2024
Read A miscellanea of qualitative and symmetry properties of the solutions to the two-phase Serrin’s problem by Lorenzo Cavallina (2024) 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-2411-00320.txt.
Description
Abstract This paper investigates the solutions to the two-phase Serrin’s problem, an overdetermined boundary value problem motivated by shape optimization. Specifically, we study the torsional rigidity of composite beams, where two distinct materials interact, and examine the properties of the optimal configurations (critical shapes) under volume constraints. We first show that such a shape optimization problem admits no local minimizers.
Catalog record
| CREATOR | Lorenzo Cavallina |
|---|---|
| YEAR | 2024 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 33 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2411-00320
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
A miscellanea of qualitative and symmetry properties of the solutions to the two-phase Serrin’s problem Lorenzo Cavallina Abstract This paper investigates the solutions to the two-phase Serrin’s problem, an overdetermined boundary value problem motivated by shape optimization. Specifically, we study the torsional rigidity of composite beams, where two distinct materials interact, and examine the properties of the optimal configurations (critical shapes) under volume constraints. We first show that such a shape optimization problem admits no local minimizers. Then, using the method of moving planes, we show that the solutions exhibit no extended or narrow branches (“tentacles”) away from the core. We then show that the outer boundary of a solution cannot exhibit flat parts and that the only configuration whose outer boundary contains a portion of a sphere is the one given by concentric balls. Finally, we establish that concentric balls are the only admissible configurations that solve the two-phase Serrin’s problem for two distinct sets of conductivity values. Key words. two-phase, overdetermined problem, shape-optimization problem, symmetry, method of moving planes, qualitative properties of the solutions AMS subject classifications. 35J15, 35N25, 35Q93. 1 Introduction Let \Omega be a bounded domain of \mathbb{R}^{N} ( N\geq 2 ) of class (at least) C^{1,\gamma} for \gamma\in(0,1] and let D\subset\overline{D}\subset\Omega be a nonempty bounded open set. Throughout this paper, we will be concerned with the following overdetermined problem (which is a generalization of the famous Serrin’s overdetermined problem [ 1 ] ): \begin{cases}-{\rm div}(\sigma\nabla u)=1\quad\text{in }\Omega,\\ u=0\quad\text{on }\partial\Omega,\\ \partial_{n}u=c\quad\text{on }\partial\Omega.\end{cases} (1.1) Here \partial_{n} denotes the normal derivative with respect to the outward unit normal vector n\in C^{\gamma}(\partial\Omega,\mathbb{R}^{N}) , c is a given real constant, and \sigma is the piece-wise constant function defined as \sigma(x):=\begin{cases}\sigma_{c}\quad\text{if }x\in D\;\text{ (core)},\\ 1\quad\text{if }x\in\Omega\setminus D\;\text{ (shell)},\end{cases} (1.2) for some given positive constant \sigma_{c}\neq 1 . Notice that, problem ( 1.1 ) is overdetermined as we are imposing two distinct boundary conditions on \partial\Omega at the same time. As a result, problem ( 1.1 ) will only admit a solution u for some suitable pairs (D,\Omega) . Throughout this paper, when no confusion arises, we will also refer to such pairs (D,\Omega) as the solutions to the overdetermined problem ( 1.1 ). When D is just an open set, one can give meaning to problem ( 1.1 ) via the following weak formulation: we say that a function u\in H_{0}^{1} solves ( 1.1 ) if it satisfies \int_{\Omega}\sigma\nabla u\cdot\nabla\varphi=\int_{\Omega}\varphi+c\int_{\partial\Omega}\varphi (1.3) for all \varphi\in H^{1}(\Omega) . Even though less immediate to see at first glance, the above formulation still yields an overdetermined problem, as we have also imposed the value of the trace of u on the boundary \partial\Omega by requiring u\in H_{0}^{1}(\Omega) . Remark 1.1 (Reformulation as a transmission problem) . Independently of the regularity of the interface \partial D , equation -{\rm{div}}(\sigma\nabla u)=1 holds in the classical sense point-wise away from the interface. That is to say, u satisfies -\sigma_{c}\Delta u=1 in D and -\Delta u=1 in \Omega\setminus\overline{D} (this follows by localizing the equation in each phase by choosing suitable test functions in the weak form ( 1.3 )). Moreover, if \partial D is sufficiently smooth, u satisfies the following transmission conditions at the interface: \left\llbracket u\right\rrbracket=0\quad\text{on }\partial D,\quad\left\llbracket\sigma\partial_{n}u\right\rrbracket=0\quad\text{on }\partial D, where \left\llbracket\cdot\right\rrbracket denotes the difference between the traces on \partial D from the outside and the inside respectively, and \partial_{n} denotes the normal derivative on \partial D . Remark 1.2 (On the regularity of \partial\Omega ) . Throughout this paper, we assume that \partial\Omega is of class C^{1,\gamma} for some \gamma\in(0,1] . Thus, by the classical Schauder boundary estimates (see [ 2 ] ), the solution u of the Dirichlet problem given by the first two lines of ( 1.1 ) is of class C^{1,\gamma} in a neighborhood of the boundary \partial\Omega . In particular, if such u also satisfies the Neumann boundary condition in ( 1.1 ) (that is, if (D,\Omega) solves the overdetermined problem ( 1.1 )), then it does so in the classical (point-wise) sense. As a result, the local regularity result [ 3 , Theorem 2] implies that \partial\Omega is an analytic surface. Overdetermined problem ( 1.1 ) has many physical interpretations. For instance, when Serrin originally introduced his overdetermined problem in [ 1 ] , corresponding to the case \sigma\equiv 1 in ( 1.1 ), he gave the following fluid dynamical interpretation: “ the tangential stress on the pipe wall is the same at all points of the wall if and only if the pipe has a circular cross section. ” Another interpretation given in [ 1 ] pertains to the linear theory of elasticity (see [ 4 ] ): “ when a solid straight bar is subject to torsion, the magnitude of the resulting traction which occurs at the surface of the bar is independent of position if and only if the bar has a circular cross section. ” The latter interpretation readily generalizes to the case described in ( 1.1 ), where the bar is a “composite”, that is, made by two materials rigidly bonded together according to the distribution given by ( 1.2 ). We remark that there is a natural reason for requiring the traction to be constant on the outer surface of the bar. Indeed, as it will be explained in further detail in section 2 , such an overdetermined condition arises as a necessary condition for the shape optimization problem of maximizing the torsional rigidity functional T_{D}(\cdot) under a volume constraint. Here T_{D}(\Omega):=\int_{\Omega}\sigma|\nabla u|^{2}=\int_{\Omega}u, where u , the so-called “ Prandtl stress function ”, is defined as the solution to the boundary value problem -{\rm div}(\sigma\nabla u)=1\quad\text{in }\Omega,\qquad u=0\quad\text{on }\partial\Omega, that is, u the unique function in H_{0}^{1}(\Omega) satisfying ( 1.3 ) for all \varphi\in H_{0}^{1}(\Omega) . We remark that such a simplified description of torsional rigidity via Prandtl stress function is physically accurate only if the bar is sufficiently long (compared to the dimension of its cross-section) and we refer the interested reader to [ 5 ] , where the authors fruitfully use this approach to study the multi-material composites found in the morphology of plant stems. Finally, we briefly mention some works where similar overdetermined problems have been considered in the context of two-phase thermal conductors [ 6 , 7 , 8 ] . In what follows, we will introduce the main results of this paper. As shown in [ 9 ] , the radial configurations (D_{0},\Omega_{0}) given by concentric balls are local maximizers for T_{D} under a volume constraint if \sigma_{c}>1 and are saddle shapes if \sigma_{c}<1 . The following theorem concerns the non-existence of local minimizers. Theorem I . Let (D,\Omega) be a solution to the overdetermined problem ( 1.1 ). Then, (D,\Omega) is not a local minimizer in the sense of Definition 2.2 . The remaining theorems concern some qualitative properties of \partial\Omega . In particular, the following theorem formalizes the intuitive notion that “ \Omega follows the shape of D without meandering too much”. Theorem II . Let (D,\Omega) be a solution to the overdetermined problem ( 1.1 ). Then \Omega “ has no tentacles ”, in the sense of Definition 3.1 . The following theorem shows that \partial\Omega generally exhibits a non-trivial shape. Theorem III . Let (D,\Omega) be a