On homogenization of the Landau-Lifshitz equation with rapidly oscillating material coefficient (Q2129659)

The article addresses homogenization of the Landau-Lifshitz equation \[ \frac{\partial \mathbf{m}^\varepsilon}{\partial t}(x,t) = -\mathbf{m}^\varepsilon(x,t)\times \mathcal{L}^\varepsilon \mathbf{m}^\varepsilon(x,t) -\alpha\times \mathbf{m}^\varepsilon(x,t)\times \mathbf{m}^\varepsilon(x,t)\times \mathcal{L}^\varepsilon \mathbf{m}^\varepsilon(x,t), \] where \(\mathcal{L}^\varepsilon \mathbf{m} =\nabla\cdot(a^\varepsilon\nabla\mathbf{m})\) with an \(\varepsilon\)-periodic material coefficient, namely \[a^\varepsilon(x)=a(x\varepsilon^{-1}),\ a(\cdot)\text{ is a periodic function}.\] There are several articles dealing with the homogenization of this equation. The homogenized equation has the form \[ \frac{\partial \mathbf{m}_0}{\partial t}(x,t) = -\mathbf{m}_0(x,t)\times \overline{\mathcal{L}} \mathbf{m}_0(x,t) -\alpha\times \mathbf{m}_0(x,t)\times \mathbf{m}_0(x,t)\times \overline{\mathcal{L}} \mathbf{m}_0(x,t), \] where \(\overline{\mathcal{L}} \mathbf{m} =\nabla\cdot(\nabla\mathbf{m}\mathbf{A}^H)\) with some constant homogenized matrix \(\mathbf{A}^H\). The main contribution of the current paper are the estimates for the difference between the solutions \(\mathbf{m}_\varepsilon\) and \(\mathbf{m}_0\) with arbitrary many correction terms. The rates for the convergence in terms of \(\varepsilon\) in Sobolev norms for dimensions \(n\in\{1,2,3\}\) are obtained; these rates depend on the length of the time interval considered. In short, the main result reads as follows: under some regularity assumptions on the material coefficient \(a(x)\) and the solutions \(\mathbf{m}^\varepsilon\) and \(\mathbf{m}_0\), one has the estimate \[ \|\mathbf{m}^\varepsilon(\cdot, t) - \tilde{\mathbf{m}}^\varepsilon_J\|_{H^q}\leq C\varepsilon^\gamma, \quad t\in [0,T \varepsilon^\sigma],\ \sigma\in [0,2], \] Here \begin{itemize} \item \( \tilde{\mathbf{m}}^\varepsilon_J(x, t)=\mathbf{m}_0+\sum_{j=1}^J\varepsilon^j\mathbf{m}_j(x,x\varepsilon^{-1},t,t\varepsilon^{-2}) \), where \(\mathbf{m}_1\), \(\mathbf{m}_2\),\dots are some higher order correctors, \item \(q\) and \(\gamma\) depend on \(\sigma\), \(J\) and the regularity of \(\mathbf{m}^\varepsilon\) and \(\mathbf{m}_0\). \end{itemize} In particular, for \(\sigma=0\) (fixed time interval) one has \[ \|\mathbf{m}^\varepsilon(\cdot, t) - \tilde{\mathbf{m}}^\varepsilon_J\|_{L^2}\leq C\varepsilon, \quad \|\mathbf{m}^\varepsilon(\cdot, t) - \tilde{\mathbf{m}}^\varepsilon_J\|_{H^1}\leq C, \] for \(0\le t\le T\) and \(0\le J \le 2\).