Approximation of the exponential of a diffusion operator with multiscale coefficients (Q2258218)

From MaRDI portal
scientific article
Language Label Description Also known as
English
Approximation of the exponential of a diffusion operator with multiscale coefficients
scientific article

    Statements

    Approximation of the exponential of a diffusion operator with multiscale coefficients (English)
    0 references
    0 references
    3 March 2015
    0 references
    The article under review discusses certain operator norm estimates for multiscale homogenization problems for the heat equation. More precisely, the author derives the estimate \((*) \quad\quad\quad \| e^{-tA_\epsilon}-e^{-tA_0}\|_{L\big(L^2\big(\mathbb{R}^d\big)\big)}\leq c_0 \sqrt{\epsilon}, \quad t\geq 0,\quad \epsilon\in (0,1), \) where \(\|\cdot\|_{L\big(L^2\big(\mathbb{R}^d\big)\big)}\) refers to the operator norm in \(L\big(L^2\big(\mathbb{R}^d\big)\big)\), \((e^{-tA_\epsilon})_{t\geq 0}\) and \((e^{-tA_0})_{t\geq 0}\) denote the semigroups generated by the respective differential expressions \(-A_\epsilon = \text{div}\) \(a_\epsilon\text{grad}\) and \(-A_0=\text{div}\) \(a_0\, \text{grad}\) on \(\mathbb{R}^d,\) which -- as usual -- are properly defined by the respective quadratic forms. The coefficient \(a_0\) is derived via (standard) reiterated homogenization formulae involving \(a: ([-1/2,1/2)^d)^2\to \mathbb{R}^{d\times d}\) being bounded below and above as well as satisfying a Lipschitz estimate with respect to the first variable with corresponding constant uniformly in the second variable; \(a\), periodically extended, also gives \(a_\epsilon(x):=a\big(\frac{x}{\epsilon},\frac{x}{\epsilon^2}\big)\). The main result of the present paper is estimate \((*)\) with \(c_0\) depending on the spatial dimension \(d\), the boundedness constants of \(a\), as well as its Lipschitz constant. In Section 3, the author reduces \((*)\) to a corresponding estimate for the norm \(\|\cdot\|_{L\big(H^1\big(\mathbb{R}^d\big),L^2\big(\mathbb{R}^d\big)\big)}\). The main technical tool (Lemma 3) presented in Section 4 for obtaining this estimate discusses the so-called shifted first approximation. In turn, the proof of Lemma 3 is sketched in Section 7. The paper is a good read for those being experts in the field. The generalizations to the case of multiscale homogenization with more than \(2\) different scales is sketched in the concluding section.
    0 references
    0 references
    homogenization
    0 references
    operator-type estimates
    0 references
    locally periodic and multiscale coefficients
    0 references
    shift parameters
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references