Evaluate using continuous norms. More...
#include <time_extrapolation.h>
Public Types | |
typedef ISubDiagErrorEst< TVector > | base_type |
typedef GridFunction< TDomain, TAlgebra > | grid_function_type |
typedef IGridFunctionSpace< grid_function_type > | subspace_type |
Public Member Functions | |
GridFunctionEstimator (double ref) | |
void | add (SmartPtr< subspace_type > spSubspace) |
add sub-space component | |
void | add (SmartPtr< subspace_type > spSubspace, number sigma) |
bool | update (SmartPtr< TVector > vUpdate, number alpha, SmartPtr< TVector > vFine, SmartPtr< TVector > vCoarse) |
apply w/ rel norm | |
void | set_reference_norm (number norm) |
std::string | config_string () const |
print config string | |
![]() | |
number | get_current_estimate () |
get estimate | |
void | reset_estimate () |
Protected Types | |
typedef TAlgebra::vector_type | TVector |
typedef std::pair< SmartPtr< subspace_type >, number > | weighted_obj_type |
Protected Attributes | |
number | m_refNormValue |
std::vector< weighted_obj_type > | m_spWeightedSubspaces |
![]() | |
number | m_est |
Evaluate using continuous norms.
Can provide subspaces (incl. scaling).
sum_I \sigma_I \frac{|e|_I}{|u|_I} For ref value <= 0 TOL is scaled relatively w.r.t. function norm.
|
inlinevirtual |
print config string
Reimplemented from ug::ISubDiagErrorEst< TAlgebra::vector_type >.
|
inlinevirtual |
apply w/ rel norm
Implements ug::ISubDiagErrorEst< TAlgebra::vector_type >.