Adaptive Finite Element Methods for the Identification of Distributed Parameters in Partial Differential Equations

In this thesis, we develop algorihms for the solution of inverse problems by the adaptive finite element method. In particular, we consider problems where distributed coefficients in partial differential equations are to be identified from measurements of the state variable. The ingredients for efficient algorithms are error estimates for various quantities, such as the misfit or the error in the coefficient, and adaptive finite element discretization strategies. We also discuss the numerical solution of the resulting discrete equations, as well as the incorporation of inequality constraints on the sought coefficient. The efficiency of the proposed methods is verified at a number of synthetic examples.

Level thesis:degree_level

thesis.doctoral

Grantor dc:publisher

Universität Heidelberg

Year

2002

Author dc:creator

• Bangerth, Wolfgang

Contributors dc:contributor

• Rannacher, Rolf

Repository record source_url

http://www.ub.uni-heidelberg.de/archiv/2508

OAI identifier oai:identifier

oai:archiv.ub.uni-heidelberg.de:2508