Homogenization of the oscillating Dirichlet boundary condition in general domains (Q2445879): Difference between revisions
From MaRDI portal
Latest revision as of 09:29, 8 July 2024
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | Homogenization of the oscillating Dirichlet boundary condition in general domains |
scientific article |
Statements
Homogenization of the oscillating Dirichlet boundary condition in general domains (English)
0 references
15 April 2014
0 references
The author describes a homogenization result for a fully nonlinear elliptic equation \(F_{\varepsilon }(D^{2}u^{\varepsilon },x,\frac{x}{\varepsilon })=0\) posed in a bounded and smooth domain \(\Omega \) of \(\mathbb{R}^{n}\) with the oscillating Dirichlet boundary condition \(u^{\varepsilon }(x)=g(x,\frac{x}{ \varepsilon })\) on \(\partial \Omega \). Here \((M,x,y)\rightarrow F_{\varepsilon }(M,x,y)\) is supposed to be uniformly elliptic, \(g\in C^{\alpha }(\mathbb{R}^{n}\times \mathbb{R}^{n})\) for some \(\alpha \in (0,1)\), and \(F_{\varepsilon }\) and \(g\) are \(\mathbb{Z}\)-periodic with respect to their variable \(y\). The author assumes that \(\Omega \) has \(C^{2}\) boundary and satisfies a strict \(\gamma \) exterior cone condition for some \(\gamma \in (0,1)\). Furthermore the set \(\Gamma =\{x\in \partial \Omega :\nu _{x}\in \mathbb{RZ}^{n}\}\) is supposed to have Hausdorff dimension \(\mathcal{H} ^{\beta _{0}}(\Gamma )=0\) for some \(\beta _{0}\) lying in an interval whose bounds depend on data of the problem. The main result of the paper proves that the viscosity solution \(u^{\varepsilon }\) of the fully nonlinear problem converges locally uniformly in \(\Omega \) to the unique solution \( \overline{u}\) given by Perron's method of the limit problem \(\overline{F} (D^{2}\overline{u},x)=0\) in \(\Omega \) with the boundary condition \(\overline{ u}(x)=\overline{g}(x)\) on \(\partial \Omega \setminus \Gamma \). Here \( \overline{F}\) is the homogenized operator associated to \(F_{\varepsilon }\), \(\overline{g}(x)=\overline{\mu }(g(x,\cdot ),F^{x},\nu _{x})\) where \( \overline{\mu }\) is given through the condition \(\sup_{x_{0}\in \mathbb{R} ^{n}}\sup_{y\in P(x_{0},\nu )}\left| v_{x_{0},\nu }(y+R\nu )-\overline{ \mu }\right| \rightarrow 0\) as \(R\rightarrow \infty \) with \(P(x_{0},\nu )=\{y:(y-x_{0})\cdot \nu >0\}\) and \(v_{x_{0},\nu }\) is the viscosity solution of the cell problem \(F(D^{2}v,y)=0\) in \(P(x_{0},\nu )\) with the boundary condition \(v(y)=\psi (y)\) on \(\partial P(x_{0},\nu )\). The author thus extends previous results such as the one by \textit{G. Barles} and \textit{E. Mironescu} [Asymptotic Anal. 82, No. 3--4, 187--200 (2013; Zbl 1278.35087)]. For the proof, the author mainly quotes ideas from this paper and properties of the solution of the cell problem that he proves.
0 references
fully nonlinear elliptic equations
0 references
oscillating boundary data
0 references
homogenization
0 references
boundary layers
0 references
cell problem
0 references
viscosity solution
0 references
0 references
0 references