Patch dynamics with buffers for homogenization problems (Q2490313): Difference between revisions

An important class of problems exhibits a smooth behaviour on macroscopic space and time scales, while only a microscopic evolution law is known. For such time-dependent multi-scale problems, an ``equation-free'' framework is proposed, of which patch dynamics is an essential component. Patch dynamics is designed to perform numerical simulations of an unavailable macroscopic equation on macroscopic time and length scales; it uses appropriately initialized simulations of the available microscopic model in a number of small boxes (patches), which cover only a fraction of the space-time domain. Precisely, in this article, the patch dynamics scheme is described for multi-scale initial-boundary-value problems of the form \[ \partial_t u_{\varepsilon}(x,t) = \partial_x(a(x/\varepsilon)\partial_x u_{\varepsilon}(x,t)), \quad \text{in} \quad [0,T) \times [0,1], \] subject to the following initial and boundary conditions: \[ u_{\varepsilon}(x,0) = u^0(x) \in L^2([0,1]), \quad u_{\varepsilon}(0,t) = u_{\varepsilon}(1,t) = 0, \] where \(a(y) = a(x/\epsilon)\) is uniformly elliptic and periodic in \(y\) and \(\epsilon\) is a small parameter. This scheme approximates an unavailable effective equation over macroscopic time and length scales, when only a microscopic evolution law is given; it only uses appropriately initialized simulations of the microscopic model over small subsets (patches) of the space-time domain. It is shown that it is possible to use arbitrary boundary conditions for these patches, provided that suitably large buffer regions ``shield'' the boundary artefacts from the interior of the patches. An error estimate for a model homogenization problem with Dirichlet boundary conditions is obtained. To illustrate the method, a set of numerical example is examined, which include a nonlinear reaction-diffusion equation and the Kuramoto-Sivashinsky equation.