A particle generation method following multivariate Gaussian distribution. More...

#include <MultiVariateGaussian.h>

Inheritance diagram for MultiVariateGaussian:

Collaboration diagram for MultiVariateGaussian:

Public Member Functions
	MultiVariateGaussian (std::shared_ptr< ParticleContainer_t > &pc, std::shared_ptr< FieldContainer_t > &fc, std::shared_ptr< Distribution_t > &opalDist)
	Constructor for MultiVariateGaussian.
	MultiVariateGaussian (std::shared_ptr< ParticleContainer_t > pc, const Vector_t< double, 3 > &meanR, const Vector_t< double, 3 > &meanP, const Vector_t< double, 3 > &sigmaR, const Vector_t< double, 3 > &sigmaP, const Vector_t< double, 3 > &cutoffR, const Vector_t< double, 3 > &cutoffP, bool fixMeanR=true, bool fixMeanP=true)
	Constructor for MultiVariateGaussian with specified parameters.
	MultiVariateGaussian (std::shared_ptr< ParticleContainer_t > pc, const Vector_t< double, 3 > &meanR, const Vector_t< double, 3 > &meanP, const Matrix_t &cov, const Vector_t< double, 3 > &cutoffR, const Vector_t< double, 3 > &cutoffP, bool fixMeanR=true, bool fixMeanP=true)
	Constructs the MultiVariateGaussian class.
void	ComputeCholeskyFactorization ()
	Computes the Cholesky factorization of the covariance matrix.
void	ComputeCenteredBounds ()
	Computes centered bounds for the particle distribution.
void	generateParticles (size_t &numberOfParticles, Vector_t< double, 3 > nr) override
	Generates particles based on the defined Gaussian distribution.
void	setMeanR (const Vector_t< double, 3 > &meanR)
void	setMeanP (const Vector_t< double, 3 > &meanP)
void	setCutoffR (const Vector_t< double, 3 > &cutoffR)
void	setCutoffP (const Vector_t< double, 3 > &cutoffP)
void	setFixMeanR (bool fixMeanR)
void	setFixMeanP (bool fixMeanP)
void	setSigmaR (const Vector_t< double, 3 > &sigmaR)
void	setSigmaP (const Vector_t< double, 3 > &sigmaP)
void	setCovarianceMatrix (const Matrix_t &cov)
void	setL (const Matrix_t &L)
virtual void	emitParticles (double t, double dt)
virtual void	testNumEmitParticles (size_t nsteps, double dt)
virtual void	testEmitParticles (size_t nsteps, double dt)
virtual void	initDomainDecomp (double BoxIncr)
virtual void	setWithDomainDecomp (bool withDomainDecomp)

Public Attributes
IpplTimings::TimerRef	samplerTimer_m
	Timer for performance profiling.

Protected Attributes
std::shared_ptr< ParticleContainer_t >	pc_m
std::shared_ptr< FieldContainer_t >	fc_m
std::shared_ptr< Distribution_t >	opalDist_m
std::string	samplingMethod_m

Private Member Functions
void	initRandomPool ()
	Initializes the random number generator pool.

Private Attributes
Vector_t< double, 3 >	meanR_m
Vector_t< double, 3 >	meanP_m
Matrix_t	cov_m
Matrix_t	L_m
Vector_t< double, 3 >	rmin_m
Vector_t< double, 3 >	rmax_m
Vector_t< double, 3 >	pmin_m
Vector_t< double, 3 >	pmax_m
Vector_t< double, 3 >	normRmin_m
Vector_t< double, 3 >	normRmax_m
Vector_t< double, 3 >	normPmin_m
Vector_t< double, 3 >	normPmax_m
Vector_t< double, 6 >	min_m
	Min and Max bounds for all 6 dimensions (R0,P0,R1,P1,R2,P2).
Vector_t< double, 6 >	max_m
Vector_t< double, 6 >	normMin_m
Vector_t< double, 6 >	normMax_m
Vector_t< double, 3 >	cutoffR_m
	Cutoff multipliers for position and momentum distributions.
Vector_t< double, 3 >	cutoffP_m
GeneratorPool	randPool_m
	Pool of random number generators for parallel sampling.
Vector_t< double, 3 >	sigmaR_m
	Standard deviations for position and momentum distributions.
Vector_t< double, 3 >	sigmaP_m
bool	fixMeanR_m
	Flag to exactly fix the mean R and P of particles after sampling.
bool	fixMeanP_m

Detailed Description

A particle generation method following multivariate Gaussian distribution.

This class generates particles following a multivariate Gaussian distribution using Cholesky factorization and inverse transformation sampling.

Given covariance matrix cov_m = [ Cov(R0,R0), Cov(R0,P0), Cov(R0,R1), Cov(R0,P1), ...] whose values are read from opalDist_m->correlationMatrix_m. First, the Cholesky factorization is computed cov_m = L_m * L_m^T Then, normally distribution particles R=P~N(0,I) are transformed to multivariate using L_m.

Definition at line 32 of file MultiVariateGaussian.h.

Constructor & Destructor Documentation

◆ MultiVariateGaussian() [1/3]

MultiVariateGaussian::MultiVariateGaussian	(	std::shared_ptr< ParticleContainer_t > &	pc,
		std::shared_ptr< FieldContainer_t > &	fc,
		std::shared_ptr< Distribution_t > &	opalDist )

Constructor for MultiVariateGaussian.

Constructs the MultiVariateGaussian class.

Parameters

pc	Shared pointer to the particle container.
fc	Shared pointer to the field container.
opalDist	Shared pointer to the distribution.

Definition at line 13 of file MultiVariateGaussian.cpp.

References cov_m, initRandomPool(), meanP_m, meanR_m, SamplingBase::opalDist_m, samplerTimer_m, SamplingBase::SamplingBase(), setCutoffP(), setCutoffR(), setSigmaP(), and setSigmaR().

Here is the call graph for this function:

◆ MultiVariateGaussian() [2/3]

MultiVariateGaussian::MultiVariateGaussian	(	std::shared_ptr< ParticleContainer_t >	pc,
		const Vector_t< double, 3 > &	meanR,
		const Vector_t< double, 3 > &	meanP,
		const Vector_t< double, 3 > &	sigmaR,
		const Vector_t< double, 3 > &	sigmaP,
		const Vector_t< double, 3 > &	cutoffR,
		const Vector_t< double, 3 > &	cutoffP,
		bool	fixMeanR = true,
		bool	fixMeanP = true )

Constructor for MultiVariateGaussian with specified parameters.

Parameters

meanR	Mean position vector.
meanP	Mean momentum vector.
sigmaR	Standard deviation for position distribution.
sigmaP	Standard deviation for momentum distribution.
cutoffR	Cutoff position vector.
cutoffP	Cutoff momentum vector.
fixMeanR	Boolean flag to fix mean position.
fixMeanP	Boolean flag to fix mean momentum.

Definition at line 38 of file MultiVariateGaussian.cpp.

References cov_m, initRandomPool(), samplerTimer_m, SamplingBase::SamplingBase(), setCutoffP(), setCutoffR(), setFixMeanP(), setFixMeanR(), setMeanP(), setMeanR(), setSigmaP(), and setSigmaR().

Here is the call graph for this function:

◆ MultiVariateGaussian() [3/3]

MultiVariateGaussian::MultiVariateGaussian	(	std::shared_ptr< ParticleContainer_t >	pc,
		const Vector_t< double, 3 > &	meanR,
		const Vector_t< double, 3 > &	meanP,
		const Matrix_t &	cov,
		const Vector_t< double, 3 > &	cutoffR,
		const Vector_t< double, 3 > &	cutoffP,
		bool	fixMeanR = true,
		bool	fixMeanP = true )

Constructs the MultiVariateGaussian class.

Parameters

pc	Shared pointer to the particle container.
meanR	Mean position vector.
meanP	Mean momentum vector.
cov	Covariance matrix.
cutoffR	Cutoff position vector.
cutoffP	Cutoff momentum vector.
fixMeanR	Boolean flag to fix mean position.
fixMeanP	Boolean flag to fix mean momentum.

Definition at line 74 of file MultiVariateGaussian.cpp.

References cov_m, initRandomPool(), samplerTimer_m, SamplingBase::SamplingBase(), setCutoffP(), setCutoffR(), setFixMeanR(), setMeanP(), setMeanR(), setSigmaP(), and setSigmaR().

Here is the call graph for this function:

Member Function Documentation

◆ ComputeCenteredBounds()

void MultiVariateGaussian::ComputeCenteredBounds ( )

Computes centered bounds for the particle distribution.

Computes normalized boundaries for the multivariate Gaussian sampling.

Definition at line 150 of file MultiVariateGaussian.cpp.

References cutoffP_m, cutoffR_m, L_m, max_m, min_m, normMax_m, normMin_m, normPmax_m, normPmin_m, normRmax_m, normRmin_m, pmax_m, pmin_m, rmax_m, rmin_m, sigmaP_m, and sigmaR_m.

Referenced by generateParticles().

Here is the caller graph for this function:

◆ ComputeCholeskyFactorization()

void MultiVariateGaussian::ComputeCholeskyFactorization ( )

Computes the Cholesky factorization of the covariance matrix.

Computes the Cholesky decomposition of the covariance matrix.

Definition at line 125 of file MultiVariateGaussian.cpp.

References cov_m, and L_m.

Referenced by generateParticles().

Here is the caller graph for this function:

◆ emitParticles()

virtual void SamplingBase::emitParticles	(	double	t,
		double	dt )

inlinevirtualinherited

Reimplemented in FlatTop.

Definition at line 31 of file SamplingBase.hpp.

◆ generateParticles()

void MultiVariateGaussian::generateParticles	(	size_t &	numberOfParticles,
		Vector_t< double, 3 >	nr )

overridevirtual

Generates particles based on the defined Gaussian distribution.

Generates particles following a multivariate Gaussian distribution.

Parameters

numberOfParticles	Number of particles to generate.
nr	Vector specifying additional sampling parameters.

Reimplemented from SamplingBase.

Definition at line 198 of file MultiVariateGaussian.cpp.

References ComputeCenteredBounds(), ComputeCholeskyFactorization(), fixMeanP_m, fixMeanR_m, L_m, meanP_m, meanR_m, normPmax_m, normPmin_m, normRmax_m, normRmin_m, nr, SamplingBase::pc_m, randPool_m, and samplerTimer_m.

Here is the call graph for this function:

◆ initDomainDecomp()

virtual void SamplingBase::initDomainDecomp ( double BoxIncr )

inlinevirtualinherited

Reimplemented in FlatTop.

Definition at line 39 of file SamplingBase.hpp.

◆ initRandomPool()

void MultiVariateGaussian::initRandomPool ( )

private

Initializes the random number generator pool.

Definition at line 106 of file MultiVariateGaussian.cpp.

References gmsg, randPool_m, and Options::seed.

Referenced by MultiVariateGaussian(), MultiVariateGaussian(), and MultiVariateGaussian().

Here is the caller graph for this function:

◆ setCovarianceMatrix()

void MultiVariateGaussian::setCovarianceMatrix ( const Matrix_t & cov )

inline

Definition at line 139 of file MultiVariateGaussian.h.

References cov_m.

◆ setCutoffP()

void MultiVariateGaussian::setCutoffP ( const Vector_t< double, 3 > & cutoffP )

inline

Definition at line 120 of file MultiVariateGaussian.h.

References cutoffP_m.

Referenced by MultiVariateGaussian(), MultiVariateGaussian(), and MultiVariateGaussian().

Here is the caller graph for this function:

◆ setCutoffR()

void MultiVariateGaussian::setCutoffR ( const Vector_t< double, 3 > & cutoffR )

inline

Definition at line 116 of file MultiVariateGaussian.h.

References cutoffR_m.

Referenced by MultiVariateGaussian(), MultiVariateGaussian(), and MultiVariateGaussian().

Here is the caller graph for this function:

◆ setFixMeanP()

void MultiVariateGaussian::setFixMeanP ( bool fixMeanP )

inline

Definition at line 128 of file MultiVariateGaussian.h.

References fixMeanP_m.

Referenced by MultiVariateGaussian().

Here is the caller graph for this function:

◆ setFixMeanR()

void MultiVariateGaussian::setFixMeanR ( bool fixMeanR )

inline

Definition at line 124 of file MultiVariateGaussian.h.

References fixMeanR_m.

Referenced by MultiVariateGaussian(), and MultiVariateGaussian().

Here is the caller graph for this function:

◆ setL()

void MultiVariateGaussian::setL ( const Matrix_t & L )

inline

Definition at line 143 of file MultiVariateGaussian.h.

References L_m.

◆ setMeanP()

void MultiVariateGaussian::setMeanP ( const Vector_t< double, 3 > & meanP )

inline

Definition at line 112 of file MultiVariateGaussian.h.

References meanP_m.

Referenced by MultiVariateGaussian(), and MultiVariateGaussian().

Here is the caller graph for this function:

◆ setMeanR()

void MultiVariateGaussian::setMeanR ( const Vector_t< double, 3 > & meanR )

inline

Definition at line 108 of file MultiVariateGaussian.h.

References meanR_m.

Referenced by MultiVariateGaussian(), and MultiVariateGaussian().

Here is the caller graph for this function:

◆ setSigmaP()

void MultiVariateGaussian::setSigmaP ( const Vector_t< double, 3 > & sigmaP )

inline

Definition at line 135 of file MultiVariateGaussian.h.

References sigmaP_m.

Referenced by MultiVariateGaussian(), MultiVariateGaussian(), and MultiVariateGaussian().

Here is the caller graph for this function:

◆ setSigmaR()

void MultiVariateGaussian::setSigmaR ( const Vector_t< double, 3 > & sigmaR )

inline

Definition at line 132 of file MultiVariateGaussian.h.

References sigmaR_m.

Referenced by MultiVariateGaussian(), MultiVariateGaussian(), and MultiVariateGaussian().

Here is the caller graph for this function:

◆ setWithDomainDecomp()

virtual void SamplingBase::setWithDomainDecomp ( bool withDomainDecomp )

inlinevirtualinherited

Reimplemented in FlatTop.

Definition at line 41 of file SamplingBase.hpp.

◆ testEmitParticles()

virtual void SamplingBase::testEmitParticles	(	size_t	nsteps,
		double	dt )

inlinevirtualinherited

Definition at line 37 of file SamplingBase.hpp.

◆ testNumEmitParticles()

virtual void SamplingBase::testNumEmitParticles	(	size_t	nsteps,
		double	dt )

inlinevirtualinherited

Definition at line 34 of file SamplingBase.hpp.

Member Data Documentation

◆ cov_m

Matrix_t MultiVariateGaussian::cov_m

private

Definition at line 156 of file MultiVariateGaussian.h.

Referenced by ComputeCholeskyFactorization(), MultiVariateGaussian(), MultiVariateGaussian(), MultiVariateGaussian(), and setCovarianceMatrix().

◆ cutoffP_m

Vector_t<double, 3> MultiVariateGaussian::cutoffP_m

private

Definition at line 182 of file MultiVariateGaussian.h.

Referenced by ComputeCenteredBounds(), and setCutoffP().

◆ cutoffR_m

Vector_t<double, 3> MultiVariateGaussian::cutoffR_m

private

Cutoff multipliers for position and momentum distributions.

Definition at line 181 of file MultiVariateGaussian.h.

Referenced by ComputeCenteredBounds(), and setCutoffR().

◆ fc_m

std::shared_ptr<FieldContainer_t> SamplingBase::fc_m

protectedinherited

Definition at line 14 of file SamplingBase.hpp.

Referenced by FlatTop::initDomainDecomp(), SamplingBase(), and FlatTop::setParameters().

◆ fixMeanP_m

bool MultiVariateGaussian::fixMeanP_m

private

Definition at line 204 of file MultiVariateGaussian.h.

Referenced by generateParticles(), and setFixMeanP().

◆ fixMeanR_m

bool MultiVariateGaussian::fixMeanR_m

private

Flag to exactly fix the mean R and P of particles after sampling.

Definition at line 203 of file MultiVariateGaussian.h.

Referenced by generateParticles(), and setFixMeanR().

◆ L_m

Matrix_t MultiVariateGaussian::L_m

private

Definition at line 161 of file MultiVariateGaussian.h.

Referenced by ComputeCenteredBounds(), ComputeCholeskyFactorization(), generateParticles(), and setL().

◆ max_m

Vector_t<double, 6> MultiVariateGaussian::max_m

private

Definition at line 176 of file MultiVariateGaussian.h.

Referenced by ComputeCenteredBounds().

◆ meanP_m

Vector_t<double, 3> MultiVariateGaussian::meanP_m

private

Definition at line 151 of file MultiVariateGaussian.h.

Referenced by generateParticles(), MultiVariateGaussian(), and setMeanP().

◆ meanR_m

Vector_t<double, 3> MultiVariateGaussian::meanR_m

private

Definition at line 151 of file MultiVariateGaussian.h.

Referenced by generateParticles(), MultiVariateGaussian(), and setMeanR().

◆ min_m

Vector_t<double, 6> MultiVariateGaussian::min_m

private

Min and Max bounds for all 6 dimensions (R0,P0,R1,P1,R2,P2).

Definition at line 176 of file MultiVariateGaussian.h.

Referenced by ComputeCenteredBounds().

◆ normMax_m

Vector_t<double, 6> MultiVariateGaussian::normMax_m

private

Definition at line 176 of file MultiVariateGaussian.h.

Referenced by ComputeCenteredBounds().

◆ normMin_m

Vector_t<double, 6> MultiVariateGaussian::normMin_m

private

Definition at line 176 of file MultiVariateGaussian.h.

Referenced by ComputeCenteredBounds().

◆ normPmax_m

Vector_t<double, 3> MultiVariateGaussian::normPmax_m

private

Definition at line 171 of file MultiVariateGaussian.h.

Referenced by ComputeCenteredBounds(), and generateParticles().

◆ normPmin_m

Vector_t<double, 3> MultiVariateGaussian::normPmin_m

private

Definition at line 171 of file MultiVariateGaussian.h.

Referenced by ComputeCenteredBounds(), and generateParticles().

◆ normRmax_m

Vector_t<double, 3> MultiVariateGaussian::normRmax_m

private

Definition at line 171 of file MultiVariateGaussian.h.

Referenced by ComputeCenteredBounds(), and generateParticles().

◆ normRmin_m

Vector_t<double, 3> MultiVariateGaussian::normRmin_m

private

Definition at line 171 of file MultiVariateGaussian.h.

Referenced by ComputeCenteredBounds(), and generateParticles().

◆ opalDist_m

std::shared_ptr<Distribution_t> SamplingBase::opalDist_m

protectedinherited

Definition at line 15 of file SamplingBase.hpp.

Referenced by FlatTop::countEnteringParticlesPerRank(), FlatTop::generateUniformDisk(), FlatTop::initDomainDecomp(), MultiVariateGaussian::MultiVariateGaussian(), SamplingBase(), and FlatTop::setParameters().

◆ pc_m

std::shared_ptr<ParticleContainer_t> SamplingBase::pc_m

protectedinherited

Definition at line 13 of file SamplingBase.hpp.

Referenced by FlatTop::allocateParticles(), FlatTop::countEnteringParticlesPerRank(), FlatTop::emitParticles(), FlatTop::generateParticles(), Gaussian::generateParticles(), MultiVariateGaussian::generateParticles(), FlatTop::generateUniformDisk(), FlatTop::initDomainDecomp(), SamplingBase(), SamplingBase(), and FlatTop::testNumEmitParticles().

◆ pmax_m

Vector_t<double, 3> MultiVariateGaussian::pmax_m

private

Definition at line 166 of file MultiVariateGaussian.h.

Referenced by ComputeCenteredBounds().

◆ pmin_m

Vector_t<double, 3> MultiVariateGaussian::pmin_m

private

Definition at line 166 of file MultiVariateGaussian.h.

Referenced by ComputeCenteredBounds().

◆ randPool_m

GeneratorPool MultiVariateGaussian::randPool_m

private

Pool of random number generators for parallel sampling.

Definition at line 192 of file MultiVariateGaussian.h.

Referenced by generateParticles(), and initRandomPool().

◆ rmax_m

Vector_t<double, 3> MultiVariateGaussian::rmax_m

private

Definition at line 166 of file MultiVariateGaussian.h.

Referenced by ComputeCenteredBounds().

◆ rmin_m

Vector_t<double, 3> MultiVariateGaussian::rmin_m

private

Definition at line 166 of file MultiVariateGaussian.h.

Referenced by ComputeCenteredBounds().

◆ samplerTimer_m

IpplTimings::TimerRef MultiVariateGaussian::samplerTimer_m

Timer for performance profiling.

Definition at line 106 of file MultiVariateGaussian.h.

Referenced by generateParticles(), MultiVariateGaussian(), MultiVariateGaussian(), and MultiVariateGaussian().

◆ samplingMethod_m

std::string SamplingBase::samplingMethod_m

protectedinherited

Definition at line 16 of file SamplingBase.hpp.

◆ sigmaP_m

Vector_t<double, 3> MultiVariateGaussian::sigmaP_m

private

Definition at line 198 of file MultiVariateGaussian.h.

Referenced by ComputeCenteredBounds(), and setSigmaP().

◆ sigmaR_m

Vector_t<double, 3> MultiVariateGaussian::sigmaR_m

private

Standard deviations for position and momentum distributions.

Definition at line 197 of file MultiVariateGaussian.h.

Referenced by ComputeCenteredBounds(), and setSigmaR().

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

Public Member Functions

Public Attributes

Protected Attributes

Private Member Functions

Private Attributes

Detailed Description

Constructor & Destructor Documentation

◆ MultiVariateGaussian() [1/3]

◆ MultiVariateGaussian() [2/3]

◆ MultiVariateGaussian() [3/3]

Member Function Documentation

◆ ComputeCenteredBounds()

◆ ComputeCholeskyFactorization()

◆ emitParticles()

◆ generateParticles()

◆ initDomainDecomp()

◆ initRandomPool()

◆ setCovarianceMatrix()

◆ setCutoffP()

◆ setCutoffR()

◆ setFixMeanP()

◆ setFixMeanR()

◆ setL()

◆ setMeanP()

◆ setMeanR()

◆ setSigmaP()

◆ setSigmaR()

◆ setWithDomainDecomp()

◆ testEmitParticles()

◆ testNumEmitParticles()

Member Data Documentation

◆ cov_m

◆ cutoffP_m

◆ cutoffR_m

◆ fc_m

◆ fixMeanP_m

◆ fixMeanR_m

◆ L_m

◆ max_m

◆ meanP_m

◆ meanR_m

◆ min_m

◆ normMax_m

◆ normMin_m

◆ normPmax_m

◆ normPmin_m

◆ normRmax_m

◆ normRmin_m

◆ opalDist_m

◆ pc_m

◆ pmax_m

◆ pmin_m

◆ randPool_m

◆ rmax_m

◆ rmin_m

◆ samplerTimer_m

◆ samplingMethod_m

◆ sigmaP_m

◆ sigmaR_m