NonDSampling Class Reference

NonDIncremLHSSampling, and NonDAdaptImpSampling. More...

Inheritance diagram for NonDSampling:

List of all members.

Public Member Functions
void	compute_distribution_mappings (const ResponseArray &samples)
	z to p/beta and of p/beta to z
void	compute_correlations (const VariablesArray &vars_samples, const ResponseArray &resp_samples)
	simple, partial, simple rank, and partial rank
void	update_final_statistics ()
	and computedProbLevels/computedRelLevels/computedRespLevels
void	print_distribution_mappings (ostream &s) const
	prints the p/beta/z mappings computed in compute_distribution_mappings()
void	print_correlations (ostream &s) const
	prints the correlations computed in compute_correlations()
Protected Member Functions
	NonDSampling (Model &model)
	constructor
	NonDSampling (NoDBBaseConstructor, Model &model, int samples, int seed)
	alternate constructor for sample generation and evaluation "on the fly"
	NonDSampling (NoDBBaseConstructor, int samples, int seed, const RealVector &lower_bnds, const RealVector &upper_bnds)
	alternate constructor for sample generation "on the fly"
	~NonDSampling ()
	destructor
void	sampling_reset (int min_samples, int rec_samples, bool all_data_flag, bool stats_flag)
	resets number of samples and sampling flags
const String &	sampling_scheme () const
	return sampleType: "lhs" or "random"
void	vary_pattern (bool pattern_flag)
	set varyPattern
void	get_parameter_sets (const Model &model)
	distributions/bounds defined in the incoming model.
void	get_parameter_sets (const RealVector &lower_bnds, const RealVector &upper_bnds)
	lower_bnds/upper_bnds.
void	initialize_lhs (bool write_message)
	increments numLHSRuns, sets randomSeed, and initialized lhsDriver
void	finalize_lhs (RealDenseMatrix &samples_array)
	converts samples_array into allVariables
void	compute_statistics (const VariablesArray &vars_samples, const ResponseArray &resp_samples)
	or intervals (epsitemic or mixed uncertainties)
void	compute_intervals (const ResponseArray &samples)
	called by compute_statistics() to calculate min/max intervals
void	compute_moments (const ResponseArray &samples)
	deviations, and confidence intervals
void	print_statistics (ostream &s) const
	prints the statistics computed in compute_statistics()
void	print_intervals (ostream &s) const
	prints the intervals computed in compute_intervals()
void	print_moments (ostream &s) const
	prints the moments computed in compute_moments()
void	simple_corr (RealDenseMatrix &total_data, bool rank_on, const int &num_in)
	computes simple correlations
void	partial_corr (RealDenseMatrix &total_data, bool rank_on, const int &num_in)
	computes partial correlations
Static Protected Member Functions
static bool	rank_sort (const int &x, const int &y)
	sort algorithm to compute ranks for rank correlations
Protected Attributes
const int	originalSeed
	the user seed specification (default is 0)
int	randomSeed
	the current random number seed
const int	samplesSpec
	initial specification of number of samples
int	numSamples
	the current number of samples to evaluate
String	sampleType
	the sample type: random, lhs, or incremental_lhs
Pecos::LHSDriver	lhsDriver
	the C++ wrapper for the F90 LHS library
bool	statsFlag
	flags computation/output of statistics
bool	allDataFlag
	flags update of allVariables/allResponses
short	samplingVarsMode
	the sampling mode: ACTIVE, ACTIVE_UNIFORM, ALL, or ALL_UNIFORM
short	sampleRanksMode
	SET_RANKS, or SET_GET_RANKS.
bool	varyPattern
	repeatable
RealDenseMatrix	sampleRanks
	data structure to hold the sample ranks
RealVector	mean95CIDeltas
	intervals (calculated in compute_moments())
RealVector	stdDev95CILowerBnds
	(calculated in compute_moments())
RealVector	stdDev95CIUpperBnds
	(calculated in compute_moments())
Private Attributes
size_t	numLHSRuns
	counter for number of executions of get_parameter_sets() for this object
RealVector	minValues
	(calculated in compute_intervals())
RealVector	maxValues
	(calculated in compute_intervals())
RealDenseMatrix	simpleCorr
	matrix to hold simple raw correlations
RealDenseMatrix	simpleRankCorr
	matrix to hold simple rank correlations
RealDenseMatrix	partialCorr
	matrix to hold partial raw correlations
RealDenseMatrix	partialRankCorr
	matrix to hold partial rank correlations
Static Private Attributes
static RealArray	rawData
	vector to hold raw data before rank sort

---

Detailed Description

NonDIncremLHSSampling, and NonDAdaptImpSampling.

This base class provides common code for sampling methods which employ the Latin Hypercube Sampling (LHS) package from Sandia Albuquerque's Risk and Reliability organization. NonDSampling now exclusively utilizes the 1998 Fortran 90 LHS version as documented in SAND98-0210, which was converted to a UNIX link library in 2001. The 1970's vintage LHS (that had been f2c'd and converted to incomplete classes) has been removed.

---

Constructor & Destructor Documentation

NonDSampling

(

Model &

model

)

[protected]

constructor

This constructor is called for a standard letter-envelope iterator instantiation. In this case, set_db_list_nodes has been called and probDescDB can be queried for settings from the method specification.

NonDSampling	(	NoDBBaseConstructor	,
		Model &	model,
		int	samples,
		int	seed
	)			[protected]

alternate constructor for sample generation and evaluation "on the fly"

This alternate constructor is used for generation and evaluation of on-the-fly sample sets.

NonDSampling	(	NoDBBaseConstructor	,
		int	samples,
		int	seed,
		const RealVector &	lower_bnds,
		const RealVector &	upper_bnds
	)			[protected]

alternate constructor for sample generation "on the fly"

This alternate constructor is used by ConcurrentStrategy for generation of uniform, uncorrelated sample sets.

---

Member Function Documentation

void sampling_reset	(	int	min_samples,
		int	rec_samples,
		bool	all_data_flag,
		bool	stats_flag
	)			[inline, protected, virtual]

resets number of samples and sampling flags

used by DataFitSurrModel::build_global() to publish the minimum number of samples needed from the sampling routine (to build a particular global approximation) and to set allDataFlag and statsFlag. In this case, allDataFlag is set to true (vectors of variable and response sets must be returned to build the global approximation) and statsFlag is set to false (statistics computations are not needed).

Reimplemented from Iterator.

void get_parameter_sets

(

const Model &

model

)

[protected, virtual]

distributions/bounds defined in the incoming model.

This version of get_parameter_sets() extracts data from the user-defined model in any of the four sampling modes.

Reimplemented from Analyzer.

void get_parameter_sets	(	const RealVector &	lower_bnds,
		const RealVector &	upper_bnds
	)			[protected]

lower_bnds/upper_bnds.

This version of get_parameter_sets() does not extract data from the user-defined model, but instead relies on the incoming bounded region definition. It only support a UNIFORM sampling mode, where the distinction of ACTIVE_UNIFORM vs. ALL_UNIFORM is handled elsewhere.

---

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

• NonDSampling.H
• NonDSampling.C

---