No label defined (Q4530268)

From MaRDI portal

!

This is the item page for this Wikibase entity, intended for internal use and editing purposes. Please use the normal view instead:

scientific article; zbMATH DE number 1748338
Language Label Description Also known as
default for all languages
No label defined
    English
    No label defined
    scientific article; zbMATH DE number 1748338

      Statements

      0 references
      11 June 2002
      0 references
      Stokes equations
      0 references
      This paper studies the optimal control problem for a mixture of two viscous fluids (with viscosities \(\mu_1\) and \(\mu_2\), resp.) which occupy variable regions \(\Omega_1^\varepsilon\) and \(\Omega_2^\varepsilon\) of a given bounded domain \(\Omega\) in \(\mathbb{R}^2\). Both regions are separated by a highly oscillating interface of period \(\varepsilon\), where \(\varepsilon\) is a small parameter which tends to zero. Given a set of admissible controls \(U_{\text{ad}}\) in \(L^2(\Omega)^{n+1}\) and a function \(\vec\theta\) therein, the state equation governing the blending process is \[ \left\{\begin{aligned} -&\text{div}(2\mu^\varepsilon e(\vec u^\varepsilon))+ \nabla p^\varepsilon=\vec f \chi_{\Omega_1^\varepsilon}+ \vec f \chi_{\Omega_2^\varepsilon} + \vec\theta \quad\text{in } \Omega,\\ &\operatorname {div} \vec u^\varepsilon = 0 \quad \text{in } \Omega, \qquad \vec u^\varepsilon = 0\quad \text{on } \partial\Omega, \end{aligned}\right. \tag{1} \] with \(\vec u^\varepsilon\) being the velocity of the mixture and \(p^\varepsilon\) its pressure. Here, \(e(\vec u^\varepsilon)\) is the stress tensor and \(\chi_{\Omega_i^\varepsilon}\) denotes the characteristic function of \(\Omega_i^\varepsilon\), \(i=1,2\). The viscosity \(\mu^\varepsilon\) of the mixture is given by \(\mu^\varepsilon \equiv \mu_1\chi_{\Omega_1^\varepsilon} + \mu_2\chi_{\Omega_2^\varepsilon}.\) For \(\varepsilon>0\) given, the optimal control associated to (1) is the function \(\vec\theta^\varepsilon_*\) in \(U_{\text{ad}}\) which minimizes the cost functional: \[ J_\varepsilon(\vec\theta) = \frac 12 \int_\Omega e(\vec u^\varepsilon):e(\vec u^\varepsilon) + \alpha\int_{\Omega} |\vec\theta|^2, \] where \(\alpha\) is a given constant. This paper proves existence of a limiting (homogenized) control \(\vec\theta_*\) in \(L^2(\Omega)^{n+1}\) and obtains a full characterization of it. More precisely, it is proved that \(\vec\theta_*\) satisfies the optimality condition \[ J_0(\vec\theta_*) \equiv \frac 12 \int_\Omega B e(\vec\theta_*):e(\vec\theta_*) + \alpha\int_{\Omega} |\vec\theta|^2 = \min_{\vec\theta\in U_{ad}} J_0(\vec\theta), \] where \(J_0\) is the homogenized cost functional. It is further proved that there is convergence of the minimal costs. These convergence results are based on a homogenization technique introduced by \textit{L. Baffico} and \textit{C. Conca} [J. Math. Anal. Appl. 233, 659-680 (1999; Zbl 0926.74096)] for the stationary Stokes equations. Clear and rigorous proofs are provided.
      0 references

      Identifiers