Convergence and optimality of adaptive nonconforming finite element methods for nonsymmetric and indefinite problems
- Recently an adaptive nonconforming finite element method (ANFEM) has been developed by Carstensen and Hoppe. In this paper, we extend their result to nonsymmetric and indefinite problems. The main tools in our analysis are a posteriori error estimators and a quasi-orthogonality property. In this case, we need to overcome two main difficulties: one stems from the nonconformity of the finite element space, the other is how to handle the effect of a nonsymmetric and indefinite bilinear form. Two ANFEM algorithms (ANFEM I, ANFEM II) are proposed for the lowest order Crouzeix-Raviart element. It is shown that both ANFEM algorithms are a contraction for the sum of the energy error and a scaled volume term between two consecutive adaptive loops. Moreover, optimality in the sense of optimal algorithmic complexity can be shown for ANFEM II. The results of numerical experiments confirm the theoretical findings.
Author: | Huangxin Chen, Xuejun Xu, Ronald H. W. HoppeORCiDGND |
---|---|
URN: | urn:nbn:de:bvb:384-opus4-5040 |
Frontdoor URL | https://opus.bibliothek.uni-augsburg.de/opus4/636 |
Series (Serial Number): | Preprints des Instituts für Mathematik der Universität Augsburg (2008-11) |
Type: | Preprint |
Language: | English |
Publishing Institution: | Universität Augsburg |
Contributing Corporation: | LSEC, Chines Academy of Sciences, Beijing, University of Houston |
Release Date: | 2008/02/29 |
Tag: | adaptive finite element methods; convergence; optimality; nonsymmetric problems |
GND-Keyword: | Finite-Elemente-Methode; Konvergenz; Optimum |
Institutes: | Mathematisch-Naturwissenschaftlich-Technische Fakultät |
Mathematisch-Naturwissenschaftlich-Technische Fakultät / Institut für Mathematik | |
Mathematisch-Naturwissenschaftlich-Technische Fakultät / Institut für Mathematik / Lehrstuhl für Numerische Mathematik | |
Dewey Decimal Classification: | 5 Naturwissenschaften und Mathematik / 51 Mathematik / 510 Mathematik |