MADNESS: Fred Class Reference

MADNESS 0.10.1

Wrapper around vector demonstrating interface necessary. More...

Inheritance diagram for Fred:

Inheritance graph

[legend]

Collaboration diagram for Fred:

Collaboration graph

[legend]

Public Member Functions
	Fred (const Fred &f)

	Fred (const Tensor< double > f, const Tensor< double > x, const Tensor< double > w, const double mu, const double p)

	Fred (double a, double b, double c)

string	a (const string &input) const

string	b (const string &input) const

double	get (int i) const

Tensor< double >	gradient (const Tensor< double > &expnt)
	Should return the derivative of the function.

Fred	operator* (const double &d) const

Fred &	operator+= (const Fred &f)

Fred &	operator= (const Fred &f)

bool	provides_gradient () const
	Override this to return true if the derivative is implemented.

void	set (int i, double a)

double	value (const Tensor< double > &expnt)
	Should return the value of the objective function.

void	value_and_gradient (const Tensor< double > &expnt, double &value, Tensor< double > &gradient)
	Reimplement if more efficient to evaluate both value and gradient in one call.

Public Member Functions inherited from madness::OptimizationTargetInterface
virtual	~OptimizationTargetInterface ()

double	test_gradient (Tensor< double > &x, double value_precision, bool doprint=true)
	Numerical test of the derivative ... optionally prints to stdout, returns max abs error.

Private Member Functions
	Fred ()

double	dpenalty (double expnt) const

Tensor< double >	make_g (const Tensor< double > &expnt)
	Makes the matrix of Gaussians g(i,j) = myexp(-expnt[j]*x[i]x[i])

double	penalty (double expnt) const

Private Attributes
const double	alpha

const Tensor< double >	f

const double	mu

const long	nx

const double	p

std::vector< double >	v

const Tensor< double >	w

const Tensor< double >	x

Detailed Description

Wrapper around vector demonstrating interface necessary.

Constructor & Destructor Documentation

◆ Fred() [1/4]

Fred::Fred ( )

private

◆ Fred() [2/4]

Fred::Fred	(	double	a,
		double	b,
		double	c
	)

inline

References a, b, c, and v.

◆ Fred() [3/4]

Fred::Fred ( const Fred & f )

inline

◆ Fred() [4/4]

Fred::Fred	(	const Tensor< double >	f,
		const Tensor< double >	x,
		const Tensor< double >	w,
		const double	mu,
		const double	p
	)

inline

Member Function Documentation

◆ a()

string Fred::a ( const string & input ) const

inline

Referenced by main().

◆ b()

string Fred::b ( const string & input ) const

inline

Referenced by main().

◆ dpenalty()

double Fred::dpenalty ( double expnt ) const

inlineprivate

References alpha.

Referenced by value_and_gradient().

◆ get()

double Fred::get ( int i ) const

inline

References v.

Referenced by test1().

◆ gradient()

Tensor< double > Fred::gradient ( const Tensor< double > & x )

inlinevirtual

Should return the derivative of the function.

Reimplemented from madness::OptimizationTargetInterface.

References errsq(), madness::g, and value_and_gradient().

Referenced by value(), and value_and_gradient().

◆ make_g()

Tensor< double > Fred::make_g ( const Tensor< double > & expnt )

inlineprivate

Makes the matrix of Gaussians g(i,j) = myexp(-expnt[j]*x[i]x[i])

References madness::g, myexp(), nx, madness::BaseTensor::size(), and x.

Referenced by value_and_gradient().

◆ operator*()

Fred Fred::operator* ( const double & d ) const

inline

References d, and f.

◆ operator+=()

Fred & Fred::operator+= ( const Fred & f )

inline

References f, and v.

◆ operator=()

Fred & Fred::operator= ( const Fred & f )

inline

References f, and v.

◆ penalty()

double Fred::penalty ( double expnt ) const

inlineprivate

References alpha.

Referenced by value_and_gradient().

◆ provides_gradient()

bool Fred::provides_gradient ( ) const

inlinevirtual

Override this to return true if the derivative is implemented.

Reimplemented from madness::OptimizationTargetInterface.

◆ set()

void Fred::set	(	int	i,
		double	a
	)

inline

References a, and v.

Referenced by expL(), and N().

◆ value()

double Fred::value ( const Tensor< double > & x )

inlinevirtual

Should return the value of the objective function.

Implements madness::OptimizationTargetInterface.

References errsq(), gradient(), and value_and_gradient().

Referenced by value_and_gradient().

◆ value_and_gradient()

void Fred::value_and_gradient	(	const Tensor< double > &	x,
		double &	value,
		Tensor< double > &	gradient
	)

inlinevirtual

Reimplement if more efficient to evaluate both value and gradient in one call.

Reimplemented from madness::OptimizationTargetInterface.

References c, dpenalty(), e(), errsq(), f, madness::g, gradient(), make_g(), mu, nx, optimize_coeffs(), p, penalty(), pow(), madness::BaseTensor::size(), v, value(), w, and x.

Referenced by gradient(), and value().

Member Data Documentation

◆ alpha

const double Fred::alpha

private

Referenced by dpenalty(), and penalty().

◆ f

const Tensor<double> Fred::f

private

Referenced by operator*(), operator+=(), operator=(), and value_and_gradient().

◆ mu

const double Fred::mu

private

Referenced by value_and_gradient().

◆ nx

const long Fred::nx

private

Referenced by make_g(), and value_and_gradient().

◆ p

const double Fred::p

private

Referenced by value_and_gradient().

◆ v

std::vector<double> Fred::v

private

Referenced by Fred(), get(), operator+=(), operator=(), set(), and value_and_gradient().

◆ w

const Tensor<double> Fred::w

private

Referenced by value_and_gradient().

◆ x

const Tensor<double> Fred::x

private

Referenced by make_g(), and value_and_gradient().

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