Adaptive Finite Element Approximation of Hyperbolic Problems
We review some recent developments concerning the a-posteriori error analysis of h- and hp-version finite element approximations to hyperbolic problems. The error bounds stem from an error representation formula which equates the error in an output functional of interest to the inner product of the...
| Main Authors: | , |
|---|---|
| Format: | Report |
| Published: |
Unspecified
2002
|
| Summary: | We review some recent developments concerning the a-posteriori error analysis of h- and hp-version finite element approximations to hyperbolic problems. The error bounds stem from an error representation formula which equates the error in an output functional of interest to the inner product of the finite element residual with the solution of a dual (adjoint) problem whose data is the density function of the target functional. Type I a-posteriori error bounds are derived which, unlike the cruder Type II bounds, retain the dual solution in the bound as a local weight-function. The relevance of Type I a-posteriori bounds is argued by showing that the local size of the error in a hyperbolic problem may be only very weakly correlated to the local size of the residual; consequently, adaptive refinement algorithms based on the size of the local residual alone can be ineffective. The sharpness of Type I a-posteriori error bounds is demonstrated on both structured and adaptively refined meshes. |
|---|