scm_offline.m File Reference

scm_offline_data = scm_offline(model, detailed_data, M_train, D_train) More...

Go to the source code of this file.

Functions
function scm_offline_data =	scm_offline (model, detailed_data, M_train, D_train)
	scm_offline_data = scm_offline(model, detailed_data, M_train, D_train) More...
function y =	scmoffline>scm_shifted_evaluation (model, x, q, a, Q, M, H)
function [ current_Coefficients , coeffs ] =	scmoffline>Theta (model)

Detailed Description

scm_offline_data = scm_offline(model, detailed_data, M_train, D_train)

Definition in file scm_offline.m.

Function Documentation

function scm_offline_data = scm_offline	(		model,
			detailed_data,
			M_train,
			D_train
	)

scm_offline_data = scm_offline(model, detailed_data, M_train, D_train)

function computing the offline quantities for the SCM.

Input

• model : the model

- detailed_data

• M_train(optional): A training set for the SCM-greedy can be given as Input

• D_train(optional): A set which influences the constraints in the online-Phase.

  fields of the model can be set in the model to configurate the

offline-phase of the SCM. If they are not set standard values are taken

• model.scm_M_alpha : natural number influencing the number of constraints used out of the set C in the online-phase.
• model.scm_M_plus : natural number influencing the number of constraints used out of the set D in the online-phase.
• model.scm_eps_tol : stopping condition for the included greedy-algorithmn. Stops when the error indicator is lower than this value.
• model.scm_size_C : stopping condition for the included greedy-algorithmn. Stops when the the desired size of C is reached. (in general the size of C should scale up with the complexety of the problem (Q and size or parameter space)
• model.scm_desired_constant: 1 for the coercive case (i.e. SCM produces lower and upper bounds to the coercivity constant) 2 for the inf-sup stable case (i.e. SCM produces lower and upper bounds to the inf-sup constant)

Output (the structure scm_offline_data containing the following fields)

• sigma_plus : the upper values of the bounding box, required in the online-phase
• sigma_minus : the lower values of the bounding box, required in the online-phase
• D : the set of parameters for the constraints J(mu,y) >= 0, required in the online-phase
• Theta_D : the coefficient vectors of the parameters in D, required in the online-phase
• C : the set of parameters for the constraints J(mu,y) >= alpha(mu), required in the online-phase
• Theta_C : the coefficient vectors of the parameters in C, required in the online-phase
• P_train : the parameter training set (for additional information)
• alpha_C : constains the exact stability constant (coercivity or inf-sup) for every parameter in C, required in the online-phase
• ystern_C : constains the vector y*(mu) for every mu in C, required in the online-phase
• max_err_seq : the error sequence of the greedy algorithmn (for additional information)
• exit_on_tol : 1 if the greedy stopped because of reaching the tolerance, 0 else
• exit_on_emptyTrain : 1 if the greedy stopped because all the parameters in the trainings set are used in C, 0 else
• exit_on_size_C : 1 if the greedy stopped because the desired size of C is reached, 0 else

Parameters

model	model
detailed_data	detailed data
M_train	M train
D_train	D train

Return values

scm_offline_data

scm offline data

Required fields of model:

• verbose —  flag indicating the verbosity level of informative output
• scm_M_alpha —  scm M alpha
• scm_eps_tol —  scm eps tol
• scm_size_C —  scm size C
• scm_desired_constant —  scm desired constant
• scm_M_plus —  scm M plus
• rb_problem_type —  rb problem type
• get_inner_product_matrix —  function W=f(model_data) returning the mass matrix \(W\) for inner product computation \(\langle u,v \rangle = u^t W v\).
• nt —  number of time steps for evolution discretizations
• Q_E —  Q E
• mu_ranges —  matrix storing the admissable ranges for parameter vector components. Each two dimensional column vector represents a range for a component specified by model.mu_names.
• decomp_mode —  flag indicating the operation mode of the function:
  • 0 (complete) : no affine parameter dependence or decomposition is performed.
  • 1 (components) : for each output argument a cell array of output matrices is returned representing the \(q\)-th component independent of the parameters given in mu_names.
  • 2 (coefficients) : returns a vector where each coordinate represents the \(q\)-the coefficient \(\sigma_{\cdot}^{q}(\mu)\) dependent on the parameters given in mu_names.
• operators —  operators

Required fields of detailed_data:

• L_E_comp —  L E comp
• L_I_comp —  L I comp
• A_comp —  A comp
• df_info.h10_inner_product_matrix —  df info.h10 inner product matrix
• df_info.dirichlet_gids —  df info.dirichlet gids

Generated fields of scm_offline_data:

• M_alpha —  M alpha
• M_plus —  M plus
• eps_tol —  eps tol
• size_C —  size C
• desired_constant —  desired constant
• Options —  Options
• sigma_plus —  sigma plus
• sigma_minus —  sigma minus
• D —  D
• Theta_D —  Theta D
• P_train —  P train
• C —  C
• Theta_C —  Theta C
• alpha_C —  alpha C
• ystern_C —  ystern C
• isfeasable —  isfeasable
• max_err_seq —  max err seq
• exit_on_tol —  exit on tol
• exit_on_emptyTrain —  exit on emptyTrain
• exit_on_size_C —  exit on size C

Definition at line 17 of file scm_offline.m.

Here is the call graph for this function:

Here is the caller graph for this function: