Exact internal controllability of Maxwell's equations (Q5917979)

From MaRDI portal
scientific article; zbMATH DE number 1406318
Language Label Description Also known as
English
Exact internal controllability of Maxwell's equations
scientific article; zbMATH DE number 1406318

    Statements

    Exact internal controllability of Maxwell's equations (English)
    0 references
    0 references
    0 references
    11 January 2001
    0 references
    The aim of this paper is to give two results on the exact controllability of the following Maxwell equations with locally distributed control: \[ \begin{cases} E'- \nabla \times H= \chi_{G(t)} (x)u,\;H'+ \nabla\times E= 0, &\text{in }\Omega\times \mathbb{R}^+,\\ \nabla\cdot E= \nabla\cdot H=0, &\text{in }\Omega\times \mathbb{R}^+,\\ \nu \times E= 0, &\text{on }\Gamma\times \mathbb{R}^+,\\ E(0)= E_0,\;H(0)= H_0, &\text{in }\Omega, \end{cases} \tag{1} \] where \(\Omega\) is bounded, open and connected in \(\mathbb{R}^3\) with a \(C^1\) boundary \(\partial\Omega= \Gamma\), \(G(t)\) is a subdomain of \(\Omega\), \(\nabla\times = \text{curl}\) (= rot), \(\nabla \cdot= \operatorname {div}\), \(E\) and \(H\) represent the electric and magnetic fields, respectively, \((E(x,t),H(x,t))\) is the state and \((u(x,t),G(t))\) is the control. First of all the author considers the case when the location and the shape of the controller \(G(t)\) is fixed (i.e. \(G(t)\equiv \omega\), \(\omega\) is a subdomain of \(\Omega\), \(\omega\neq \Omega\)). Under the hypotheses that \(\omega\) is a neighborhood of \(\Gamma\) and \(\Omega\) is star-shaped with respect to some fixed point \(x_0\in \mathbb{R}^3\), i.e. \[ (x-x_0) \nu(x)\geq 0\quad\text{for all }x\in \Gamma, \] it is shown that if \(T> T_1\triangleq 2\max_{x\in \Omega\setminus \omega}|x|\), then for a given initial distribution \((E_0, H_0)\in J(\Omega)\times \widehat{J}(\Omega)\), there exists \(u\in L^2(0,T; J(\Omega))\) such that the solution of (1) satisfies \(E(T)= 0\), \(H(T)= 0\) in \(\Omega\). Here \(J(\Omega)\) denotes the closure in \(L^2(\Omega)\) of \(\{\psi\); \(\psi\in C^\infty (\overline{\Omega})\), \(\nabla \cdot \psi= 0\}\) and \(\widehat{J} (\Omega)\) denotes the closure in \(L^2(\Omega)\) of \(\{\psi\); \(\psi\in C_0^\infty (\Omega)\), \(\nabla \cdot \psi=0\}\). Next the rapid exact controllability of (1) is studied in the case when \(G(t)\in {\mathcal G}\), where \({\mathcal G}\) is a family of set-valued functions \(G: [0,\infty)\to 2^\Omega\). It is supposed that any \(G(.)\in{\mathcal G}\) is continuous with respect to the Hausdorff metric and for any given \(T>0\) there exists \(G(.)\in{\mathcal G}\) which satisfies a certain inclusion. Under these conditions the author shows that for any given time duration \(T>0\), Maxwell's equations (1) are exactly controllable in \(J(\Omega) \times \widetilde{J} (\Omega)\) by choosing control in \(L^2 (0,T; J(\Omega))\). The multiplier technique is used as the main tool for investigation.
    0 references
    0 references
    0 references
    0 references
    0 references
    exact controllability
    0 references
    Maxwell equations
    0 references
    set-valued functions
    0 references
    multiplier technique
    0 references
    0 references