NonDGlobalReliability Class Reference

Class for global reliability methods within DAKOTA/UQ. More...

Inheritance diagram for NonDGlobalReliability:

Public Member Functions
	NonDGlobalReliability (ProblemDescDB &problem_db, Model &model)
	constructor
	~NonDGlobalReliability ()
	destructor
bool	resize ()
	reinitializes iterator based on new variable size
void	derived_init_communicators (ParLevLIter pl_iter)
	derived class contributions to initializing the communicators associated with this Iterator instance
void	derived_set_communicators (ParLevLIter pl_iter)
	derived class contributions to setting the communicators associated with this Iterator instance
void	derived_free_communicators (ParLevLIter pl_iter)
	derived class contributions to freeing the communicators associated with this Iterator instance
void	pre_run ()
	pre-run portion of run (optional); re-implemented by Iterators which can generate all Variables (parameter sets) a priori More...
void	core_run ()
	core portion of run; implemented by all derived classes and may include pre/post steps in lieu of separate pre/post More...
void	print_results (std::ostream &s, short results_state=FINAL_RESULTS)
	print the final iterator results More...

Private Member Functions
void	optimize_gaussian_process ()
	construct the GP using EGO/SKO
void	importance_sampling ()
	perform multimodal adaptive importance sampling on the GP
void	get_best_sample ()
	determine current best solution from among sample data for expected imporovement function in Performance Measure Approach (PMA)
Real	constraint_penalty (const Real &constraint, const RealVector &c_variables)
	calculate the penalty to be applied to the PMA constraint value
Real	expected_improvement (const RealVector &expected_values, const Variables &recast_vars)
	expected improvement function for the GP
Real	expected_feasibility (const RealVector &expected_values, const Variables &recast_vars)
	expected feasibility function for the GP
void	x_truth_evaluation (short mode)
	evaluate iteratedModel at current point to collect x-space truth data
void	x_truth_evaluation (const RealVector &c_vars_u, short mode)
	evaluate iteratedModel at specified point to collect x-space truth data
void	u_truth_evaluation (const RealVector &c_vars_u, short mode)
	evaluate uSpaceModel in BYPASS_SURROGATE mode to collect u-space truth data at specified point
void	u_evaluation (const RealVector &c_vars_u, short mode)
	evaluate uSpaceModel to collect u-space surrogate data at specified point

Static Private Member Functions
static void	EIF_objective_eval (const Variables &sub_model_vars, const Variables &recast_vars, const Response &sub_model_response, Response &recast_response)
	static function used as the objective function in the Expected Improvement (EIF) problem formulation for PMA
static void	EFF_objective_eval (const Variables &sub_model_vars, const Variables &recast_vars, const Response &sub_model_response, Response &recast_response)
	static function used as the objective function in the Expected Feasibility (EFF) problem formulation for RIA

Private Attributes
Real	fnStar
	minimum penalized response from among true function evaluations
short	meritFunctionType
	type of merit function used to penalize sample data
Real	lagrangeMult
	Lagrange multiplier for standard Lagrangian merit function.
Real	augLagrangeMult
	Lagrange multiplier for augmented Lagrangian merit function.
Real	penaltyParameter
	penalty parameter for augmented Lagrangian merit funciton
Real	lastConstraintViolation
	constraint violation at last iteration, used to determine if the current iterate should be accepted (must reduce violation)
bool	lastIterateAccepted
	flag to determine if last iterate was accepted this controls update of parameters for augmented Lagrangian merit fn
short	dataOrder
	order of the data used for surrogate construction, in ActiveSet request vector 3-bit format; user may override responses spec

Static Private Attributes
static NonDGlobalReliability *	nondGlobRelInstance
	pointer to the active object instance used within the static evaluator functions in order to avoid the need for static data

Additional Inherited Members
Protected Member Functions inherited from NonDReliability
	NonDReliability (ProblemDescDB &problem_db, Model &model)
	constructor
	~NonDReliability ()
	destructor
void	nested_variable_mappings (const SizetArray &c_index1, const SizetArray &di_index1, const SizetArray &ds_index1, const SizetArray &dr_index1, const ShortArray &c_target2, const ShortArray &di_target2, const ShortArray &ds_target2, const ShortArray &dr_target2)
	set primaryA{CV,DIV,DRV}MapIndices, secondaryA{CV,DIV,DRV}MapTargets within derived Iterators; supports computation of higher-level sensitivities in nested contexts (e.g., derivatives of statistics w.r.t. inserted design variables)
bool	resize ()
	reinitializes iterator based on new variable size
void	post_run (std::ostream &s)
	post-run portion of run (optional); verbose to print results; re-implemented by Iterators that can read all Variables/Responses and perform final analysis phase in a standalone way More...
const Model &	algorithm_space_model () const
Static Protected Member Functions inherited from Iterator
static void	gnewton_set_recast (const Variables &recast_vars, const ActiveSet &recast_set, ActiveSet &sub_model_set)
	conversion of request vector values for the Gauss-Newton Hessian approximation More...
Protected Attributes inherited from NonDReliability
Model	uSpaceModel
	Model representing the limit state in u-space, after any recastings and data fits.
Model	mppModel
	RecastModel which formulates the optimization subproblem: RIA, PMA, EGO.
Iterator	mppOptimizer
	Iterator which optimizes the mppModel.
unsigned short	mppSearchType
	the MPP search type selection: Local: MV, x/u-space {AMV,AMV+,TANA,QMEA} or NO_APPROX Global x/u-space EGRA
Iterator	importanceSampler
	importance sampling instance used to compute/refine probabilities
unsigned short	integrationRefinement
	integration refinement type (NO_INT_REFINE, IS, AIS, or MMAIS) provided by refinement specification
size_t	numRelAnalyses
	number of invocations of core_run()
size_t	approxIters
	number of approximation cycles for the current respFnCount/levelCount
bool	approxConverged
	indicates convergence of approximation-based iterations
int	respFnCount
	counter for which response function is being analyzed
size_t	levelCount
	counter for which response/probability level is being analyzed
size_t	statCount
	counter for which final statistic is being computed
bool	pmaMaximizeG
	flag indicating maximization of G(u) within PMA formulation
Real	requestedTargetLevel
	the {response,reliability,generalized reliability} level target for the current response function
Static Protected Attributes inherited from NonD
static NonD *	nondInstance
	pointer to the active object instance used within static evaluator functions in order to avoid the need for static data

Detailed Description

Class for global reliability methods within DAKOTA/UQ.

The NonDGlobalReliability class implements EGO/SKO for global MPP search, which maximizes an expected improvement function derived from Gaussian process models. Once the limit state has been characterized, a multimodal importance sampling approach is used to compute probabilities.

Member Function Documentation

void pre_run ( )

virtual

pre-run portion of run (optional); re-implemented by Iterators which can generate all Variables (parameter sets) a priori

pre-run phase, which a derived iterator may optionally reimplement; when not present, pre-run is likely integrated into the derived run function. This is a virtual function; when re-implementing, a derived class must call its nearest parent's pre_run(), if implemented, typically before performing its own implementation steps.

Reimplemented from Analyzer.

References Model::initialize_mapping(), Iterator::methodPCIter, NonD::miPLIndex, NonDReliability::mppModel, Analyzer::pre_run(), and Model::update_from_subordinate_model().

void core_run ( )

virtual

core portion of run; implemented by all derived classes and may include pre/post steps in lieu of separate pre/post

Virtual run function for the iterator class hierarchy. All derived classes need to redefine it.

Reimplemented from Iterator.

References NonDGlobalReliability::importance_sampling(), NonDGlobalReliability::nondGlobRelInstance, and NonDGlobalReliability::optimize_gaussian_process().

void print_results ( std::ostream &  s, short  results_state = FINAL_RESULTS  )

virtual

print the final iterator results

This virtual function provides additional iterator-specific final results outputs beyond the function evaluation summary printed in finalize_run().

Reimplemented from Analyzer.

References NonD::cdfFlag, NonD::computedGenRelLevels, NonD::computedProbLevels, NonD::computedRespLevels, Iterator::iteratedModel, Analyzer::numFunctions, NonD::print_densities(), Model::response_labels(), and Dakota::write_precision.

---

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

• NonDGlobalReliability.hpp
• NonDGlobalReliability.cpp