QMCObjectiveFunction Class Reference

Objective function optimized during a variational QMC (VMC) calculation to find the optimal wavefunction parameters. More...

#include <QMCObjectiveFunction.h>

List of all members.

Public Member Functions
void	initialize (QMCInput *input, int configsToSkip)
	Initializes this object.
QMCObjectiveFunctionResult	evaluate (Array1D< double > &params)
	Evaluates and returns the result of the objective function evaluated with a single set of parameters.
Array1D < QMCObjectiveFunctionResult >	evaluate (Array1D< Array1D< double > > &params)
	Evaluates and returns the result of the objective function evaluated with multiple single sets of parameters.
Array1D< double >	grad (Array1D< double > &params)
	Evaluates and returns the gradient of the objective function for one set of parameters.
Array1D< Array1D< double > >	grad (Array1D< Array1D< double > > &params)
	Evaluates and returns the gradient of the objective function for multiple sets of parameters.
Private Member Functions
void	numerical_grad (Array1D< double > &Params, Array1D< double > &GRAD)
Private Attributes
QMCInput *	Input
QMCReadAndEvaluateConfigs	RAEC

---

Detailed Description

Objective function optimized during a variational QMC (VMC) calculation to find the optimal wavefunction parameters.

As is standard in the field of numerical optimization, optimization means minimization. The particular form of the objective function is determined by parameters in the input file.

Definition at line 35 of file QMCObjectiveFunction.h.

---

Member Function Documentation

void QMCObjectiveFunction::initialize	(	QMCInput *	input,
		int	configsToSkip
	)

Initializes this object.

This must be called before any other functions in this object are called.

Parameters:

input

input data for the calculation

Definition at line 15 of file QMCObjectiveFunction.cpp.

References QMCReadAndEvaluateConfigs::initialize(), Input, and RAEC.

Referenced by QMCCorrelatedSamplingVMCOptimization::optimize().

QMCObjectiveFunctionResult QMCObjectiveFunction::evaluate

(

Array1D< double > &

params

)

Evaluates and returns the result of the objective function evaluated with a single set of parameters.

Parameters:

params

set of parameters to evaluate the objective function with.

Returns:

result of the objective function evaluation.

Definition at line 85 of file QMCObjectiveFunction.cpp.

Referenced by QMCMikesBracketingStepLengthSelector::bracket(), QMCWolfeStepLengthSelector::calculateLineSearchObjectiveFunction(), CKGeneticAlgorithm1::generateNewPopulation(), CKGeneticAlgorithm1::initializePopulation(), numerical_grad(), QMCLineSearch::optimize(), and QMCMikesBracketingStepLengthSelector::quadratic_rebracketer().

Array1D< QMCObjectiveFunctionResult > QMCObjectiveFunction::evaluate

(

Array1D< Array1D< double > > &

params

)

Evaluates and returns the result of the objective function evaluated with multiple single sets of parameters.

Parameters:

params

sets of parameters to evaluate the objective function with.

Returns:

results of the objective function evaluations. The index of the input parameters corresponds to the index of the returned values.

Definition at line 21 of file QMCObjectiveFunction.cpp.

References QMCFlags::energy_estimated, QMCInput::flags, QMCJastrowParameters::getPoles(), Input, QMCInput::JP, RAEC, QMCReadAndEvaluateConfigs::rootCalculateProperties(), and QMCInput::setAIParameters().

Array1D< double > QMCObjectiveFunction::grad

(

Array1D< double > &

params

)

Evaluates and returns the gradient of the objective function for one set of parameters.

Parameters:

params

sets of parameters to evaluate the gradient with.

Returns:

gradient of the objective function.

Definition at line 110 of file QMCObjectiveFunction.cpp.

References numerical_grad().

Array1D< Array1D< double > > QMCObjectiveFunction::grad

(

Array1D< Array1D< double > > &

params

)

Evaluates and returns the gradient of the objective function for multiple sets of parameters.

Parameters:

params

sets of parameters to evaluate the gradient with.

Returns:

gradients of the objective function. The index of the input parameters corresponds to the index of the returned values.

Definition at line 99 of file QMCObjectiveFunction.cpp.

void QMCObjectiveFunction::numerical_grad	(	Array1D< double > &	Params,
		Array1D< double > &	GRAD
	)			[private]

Definition at line 119 of file QMCObjectiveFunction.cpp.

References Array1D< T >::allocate(), Array1D< T >::dim1(), and evaluate().

Referenced by grad().

---

Member Data Documentation

QMCInput* QMCObjectiveFunction::Input [private]

Definition at line 90 of file QMCObjectiveFunction.h.

Referenced by evaluate(), and initialize().

QMCReadAndEvaluateConfigs QMCObjectiveFunction::RAEC [private]

Definition at line 92 of file QMCObjectiveFunction.h.

Referenced by evaluate(), and initialize().

---

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

• QMCObjectiveFunction.h
• QMCObjectiveFunction.cpp