Stokhos
Development
|
Nonlinear, inverse stochastic Galerkin ModelEvaluator. More...
#include <Stokhos_SGInverseModelEvaluator.hpp>
Public Member Functions | |
SGInverseModelEvaluator (const Teuchos::RCP< EpetraExt::ModelEvaluator > &me, const Teuchos::Array< int > &sg_p_index_map, const Teuchos::Array< int > &sg_g_index_map, const Teuchos::Array< Teuchos::RCP< const Epetra_Map > > &base_g_maps) | |
Overridden from EpetraExt::ModelEvaluator . | |
Teuchos::RCP< const Epetra_Map > | get_x_map () const |
Return solution vector map. | |
Teuchos::RCP< const Epetra_Map > | get_f_map () const |
Return residual vector map. | |
Teuchos::RCP< const Epetra_Map > | get_p_map (int l) const |
Return parameter vector map. | |
Teuchos::RCP< const Epetra_Map > | get_g_map (int l) const |
Return response map. | |
Teuchos::RCP< const Teuchos::Array< std::string > > | get_p_names (int l) const |
Return array of parameter names. | |
Teuchos::RCP< const Epetra_Vector > | get_p_init (int l) const |
Return initial parameters. | |
InArgs | createInArgs () const |
Create InArgs. | |
OutArgs | createOutArgs () const |
Create OutArgs. | |
void | evalModel (const InArgs &inArgs, const OutArgs &outArgs) const |
Evaluate model on InArgs. | |
Protected Attributes | |
Teuchos::RCP < EpetraExt::ModelEvaluator > | me |
Underlying model evaluator. | |
Teuchos::Array< int > | sg_p_index_map |
Mapping between stochastic block parameters and sg parameters. | |
Teuchos::Array< int > | sg_g_index_map |
Mapping between stochastic block responses and sg responses. | |
Teuchos::Array< Teuchos::RCP < const Epetra_Map > > | base_g_maps |
Base maps of block g vectors. | |
int | num_p |
Number of parameters. | |
int | num_g |
Number of responses. | |
int | num_p_sg |
Number of stochastic parameter vectors. | |
int | num_g_sg |
Number of stochastic response vectors. | |
Nonlinear, inverse stochastic Galerkin ModelEvaluator.
SGInverseModelEvaluator is an implementation of EpetraExt::ModelEvaluator that does the inverse of SGModelEvalutor, namely it takes SG versions of the p InArgs and g and dg/dp OutArgs, and converts them to block vectors that are passed to the underlying model evaluator. This allows block nonlinear problems to appear to SG problems.