Greedy.TrainingSetAdaptation Class Reference

Detailed Description

greedy basis generation extension which adaptively refines the trainings parameter set.

Todo:

add reference to adaptive paper.

Definition at line 18 of file TrainingSetAdaptation.m.

Inheritance diagram for Greedy.TrainingSetAdaptation:

Collaboration diagram for Greedy.TrainingSetAdaptation:

Public Member Functions
	TrainingSetAdaptation (Greedy.Interface child)
	constructor extending a Greedy algorithm object
function IDetailedData detailed_data =	init_basis (Greedy.User.IReducedModel rmodel,ModelData model_data)
	construction of an initial reduced basis
function Greedy.User.IDetailedData detailed_data =	basis_extension (Greedy.User.IReducedModel rmodel,Greedy.User.IDetailedData detailed_data, checkpoint)
	basis extension routine
function eta =	refinement_element_indicators (Greedy.User.IReducedModel rmodel,Greedy.User.IDetailedData detailed_data,Greedy.DataTree.Detailed.ILeafNode dd_leaf)
	evaluates the refinement indicators
Public Member Functions inherited from Greedy.MetaInterface
	MetaInterface (child)
	mbgi = mbgi@.nosp@m.Gree.nosp@m.dy.In.nosp@m.terf.nosp@m.ace(child.rb_engine, child.model);
function	prepare (IReducedModel rmodel,ModelData model_data)
	initialization routine for basis extension
Public Member Functions inherited from Greedy.Interface
function Greedy.Combined summ =	horzcat ()
	combines an arbitrary number of basis generation algorithms
function detailed_data =	gen_detailed_data (IReducedModel rmodel, detailed_data,Greedy.Checkpoint checkpoint)
	main entry function construction the detailed data tree storing the reduced basis functions.
virtual function Greedy.DataTree.Detailed.INode detailed_data =	init_basis (IReducedModel rmodel,ModelData model_data)
	construction of an initial reduced basis
virtual function Greedy.DataTree.Detailed.INode detailed_data =	basis_extension (IReducedModel rmodel,Greedy.DataTree.Detailed.INode detailed_data, checkpoint)
	basis extension routine
function value =	get_generated_basis_type ()
	returns the generated_basis_type property. Why???

Public Attributes
	max_refinement_steps = inf
	maximum number of refinement steps before the algorithm stops.
	max_refinement_level = 5
	maximum number of refinement levels in the refineable trainings parameter set
	refinement_mode = "uniform"
	string specifying the method type for the adaptation of the parameter sampling. uniform or adaptive
	refinement_theta = 0.5
	ratio of elements which are refined during an adaptive refinement step
	element_indicator_mode = "nodes_cogs"
	string specifying the indicator for the training set nodes.
	element_indicator_s_max = 100
	specifies maximum number of skipped refinements in case of *_skippedrefs indicator modes.
	force_stop_at_max_refinement_level = true
	boolean flag indicating whether at the maximum refinement level the algorithm shall stop.
Public Attributes inherited from Greedy.MetaInterface
	child
	id
	generated_basis_type
Public Attributes inherited from Greedy.Interface
	generated_basis_type
	string specifying the detailed data produced by this basis generation algorithm object.
	id
	This is an id string inherited by the underlying extension algorithm object implementing a Greedy.Plugin.Interface.
boolean	enable_checkpointing = true
	control variable variable controlling whether check points shall be created after every reduced basis extension step.

Static Public Attributes
static const	info_fields
	cell array of field names to be copied to the detailed data node instance during init_basis() call

Constructor & Destructor Documentation

Greedy.TrainingSetAdaptation.TrainingSetAdaptation

(

Greedy.Interface

child

)

constructor extending a Greedy algorithm object

Parameters

child

the greedy algorithm to be extended by training set adaptation.

Definition at line 144 of file TrainingSetAdaptation.m.

Member Function Documentation

function Greedy.User.IDetailedData detailed_data = Greedy.TrainingSetAdaptation.basis_extension	(	Greedy.User.IReducedModel	rmodel,
		Greedy.User.IDetailedData	detailed_data,
			checkpoint
	)

basis extension routine

basis extension routine This method is run by the gen_detailed_data() method at the very end, and shall actually construct the reduced basis.

Parameter values

checkpoint: an object of type Greedy.Checkpoint used for checkpointing.

Parameters

rmodel	object specifying how the reduced simulations can be computed.
detailed_data	object defining the basis generation algorithm and storage for storing high dimensional data, i.e. dependent on dimension \(H\). This data is necessary for detailed simulations, construction of online matrices, reduced_data and reconstruction of reduced simulations.
checkpoint	checkpoint

Return values

detailed_data

object storing the reduced basis information in the leaf nodes and information on the reduced basis generation in every node.

copied from basisgen_refined written by

Author

Bernard Haasdonk

Definition at line 174 of file TrainingSetAdaptation.m.

function IDetailedData detailed_data = Greedy.TrainingSetAdaptation.init_basis	(	Greedy.User.IReducedModel	rmodel,
		ModelData	model_data
	)

construction of an initial reduced basis

construction of an initial reduced basis This method is called at the beginning of the gen_detailed_data() method right after a call to prepare(). It should be used to construct an initial reduced basis, e.g. approximating the initial value functions in case of evolution problems.

Parameters

rmodel	object specifying how the reduced simulations can be computed.
model_data	Matlab structure storing (possibly) high dimensional data needed by IDetailedModel.detailed_simulation().

Return values

detailed_data

object storing the reduced basis information in the leaf nodes and information on the reduced basis generation in every node.

Definition at line 157 of file TrainingSetAdaptation.m.

function eta = Greedy.TrainingSetAdaptation.refinement_element_indicators	(	Greedy.User.IReducedModel	rmodel,
		Greedy.User.IDetailedData	detailed_data,
		Greedy.DataTree.Detailed.ILeafNode	dd_leaf
	)

evaluates the refinement indicators

Parameters

rmodel	object specifying how the reduced simulations can be computed.
detailed_data	object defining the basis generation algorithm and storage for storing high dimensional data, i.e. dependent on dimension \(H\). This data is necessary for detailed simulations, construction of online matrices, reduced_data and reconstruction of reduced simulations.
dd_leaf	Detailed data node which is currently extended leaf in the detailed data tree.

Return values

eta

eta

Definition at line 276 of file TrainingSetAdaptation.m.

Member Data Documentation

Greedy.TrainingSetAdaptation.element_indicator_mode = "nodes_cogs"

string specifying the indicator for the training set nodes.

Possible values

• nodes_cogs
• nodes
• nodes_skippedrefs
• nodes_cogs_skippedrefs

Default: "nodes_cogs"

Definition at line 104 of file TrainingSetAdaptation.m.

Greedy.TrainingSetAdaptation.element_indicator_s_max = 100

specifies maximum number of skipped refinements in case of *_skippedrefs indicator modes.

Default: 100

Definition at line 119 of file TrainingSetAdaptation.m.

Greedy.TrainingSetAdaptation.force_stop_at_max_refinement_level = true

boolean flag indicating whether at the maximum refinement level the algorithm shall stop.

If this is set to false, no further refinement is done at the last level, but the reduced basis spaces are generated, anyways.

Default: true

Definition at line 129 of file TrainingSetAdaptation.m.

Greedy.TrainingSetAdaptation.info_fields

static

Initial value:

= {" 'max_refinement_steps', 'max_refinement_level',    \
                    'refinement_mode', 'refinement_theta', 'element_indicator_mode',    \
                    'element_indicator_s_max', 'force_stop_at_max_refinement_level' "}

cell array of field names to be copied to the detailed data node instance during init_basis() call

Default: {" 'max_refinement_steps', 'max_refinement_level', \ 'refinement_mode', 'refinement_theta', 'element_indicator_mode', \ 'element_indicator_s_max', 'force_stop_at_max_refinement_level' "}

Definition at line 33 of file TrainingSetAdaptation.m.

Greedy.TrainingSetAdaptation.max_refinement_level = 5

maximum number of refinement levels in the refineable trainings parameter set

If this level is reached, the algorithm terminates and returns the generated reduced basis spaces.

Default: 5

Definition at line 62 of file TrainingSetAdaptation.m.

Greedy.TrainingSetAdaptation.max_refinement_steps = inf

maximum number of refinement steps before the algorithm stops.

If this level is reached, the algorithm terminates and returns the generated reduced basis spaces.

Default: inf

Definition at line 50 of file TrainingSetAdaptation.m.

Greedy.TrainingSetAdaptation.refinement_mode = "uniform"

string specifying the method type for the adaptation of the parameter sampling. uniform or adaptive

Possible values are

• uniform: the parameter sampling is refined uniformly
• adaptive: the parameter sampling is refined adaptively, such that only regions of the parameter space with high estimated errors are refined.

Default: "uniform"

Definition at line 75 of file TrainingSetAdaptation.m.

Greedy.TrainingSetAdaptation.refinement_theta = 0.5

ratio of elements which are refined during an adaptive refinement step

This value \(\theta\) is multiplied by the number of leaf elements in the current parameter grid in order to compute the number grid cells to be refined.

Default: 0.5

Definition at line 91 of file TrainingSetAdaptation.m.

---

The documentation for this class was generated from the following file:

• rbasis/basisgen/oop/+Greedy/TrainingSetAdaptation.m