Repository: Freie Universität Berlin, Math Department

Direct and Iterative Methods for Numerical Homogenization

Kornhuber, R. and Podlesny, J. and Yserentant, H. (2017) Direct and Iterative Methods for Numerical Homogenization. In: Domain Decomposition Methods in Science and Engineering. Lecture Notes in Computational Science and Engineering, XXIII (116). SpringerLink, pp. 217-225. ISBN 978-3-319-52389-7

[img] PDF
Restricted to Repository staff only

250kB

Official URL: http://dx.doi.org/10.1007/978-3-319-52389-7_21

Abstract

Elliptic problems with oscillating coefficients can be approximated up to arbitrary accuracy by using sufficiently fine meshes, i.e., by resolving the fine scale. Well-known multiscale finite elements [5, 9] can be regarded as direct numerical homogenization methods in the sense that they provide approximations of the corresponding (unfeasibly) large linear systems by much smaller systems while preserving the fine-grid discretization accuracy (model reduction). As an alternative, we present iterative numerical homogenization methods that provide approximations up to fine-grid discretization accuracy and discuss differences and commonalities.

Item Type:Book Section
Subjects:Mathematical and Computer Sciences > Mathematics > Applied Mathematics
Divisions:Department of Mathematics and Computer Science > Institute of Mathematics
ID Code:1792
Deposited By: Ulrike Eickers
Deposited On:15 Feb 2016 12:23
Last Modified:22 Sep 2017 11:36

Repository Staff Only: item control page