step7_error_landscape.m File Reference

script generating error landscapes by computing the true error of reduced simulations vs. detailed simulations for given test parameters over differing basis sizes. More...

Go to the source code of this file.

Functions
function	step7_error_landscape ()
	script generating error landscapes by computing the true error of reduced simulations vs. detailed simulations for given test parameters over differing basis sizes. More...

Detailed Description

script generating error landscapes by computing the true error of reduced simulations vs. detailed simulations for given test parameters over differing basis sizes.

Definition in file step7_error_landscape.m.

Function Documentation

function step7_error_landscape

(

)

script generating error landscapes by computing the true error of reduced simulations vs. detailed simulations for given test parameters over differing basis sizes.

The generated cell array output can later be visualized with plot_error_landscape().

required variables that need to be set

• basis_gen
• detailed_data
• Nsize : number of reduced basis sizes to be tested.
• Msize : number of colletaral reduced basis sizes to be tested.
• csample : vector of increasing values between 0 and 1 indicating a set of reduced basis sizes to be tested for coupled plots.

See also

couple_N_and_M_by_c() for more information.

• cdescr : description text for coupled plots.
• mu_set_size : default size of Regression tests parameter set.
• params.step7_outputfile : file name where the generated output structure is saved.

optional variables that can be set

• error_plots : cell array of strings specifying which test cases shall be handled by the script. Possible choices are
  1. train_set – error landscape over N and M for training parameter set
  2. train_set_coupled – error plot over c where c is the parameter of couple_N_and_M_by_c() for training parameter set
  3. error – error landscape over N and M
  4. 'error_coupled – error plot over c where c is the parameter of couple_N_and_M_by_c()
  5. error_ei – error landscape over N and M for error between reduced and detailed_simulation with empirically interpolated operators.
  6. error_ei_coupled – error landscape over c where c is the parameter of couple_N_and_M_by_c() for error between reduced and detailed_simulation with empirically interpolated operators.
  7. error_ei_rb_proj – error landscape over N and M for error between reduced and detailed_simulation with empirically interpolated operators with a subsequent projection on the reduced basis space in each time step.
  8. error_ei_rb_proj_coupled – error landscape over c where c is the parameter of couple_N_and_M_by_c() for error between reduced and detailed_simulation with empirically interpolated operators with a subsequent projection on the reduced basis space in each time step.
  By default all error landscapes are selected.

Definition at line 17 of file step7_error_landscape.m.

Here is the call graph for this function:

Here is the caller graph for this function: