sgpp::optimization::optimizer::UnconstrainedOptimizer Class Reference

Abstract class for optimizing objective functions. More...

#include <UnconstrainedOptimizer.hpp>

Inheritance diagram for sgpp::optimization::optimizer::UnconstrainedOptimizer:

Public Member Functions
virtual void	clone (std::unique_ptr< UnconstrainedOptimizer > &clone) const =0
	Pure virtual method for cloning the optimizer.
const base::DataMatrix &	getHistoryOfOptimalPoints () const
const base::DataVector &	getHistoryOfOptimalValues () const
size_t	getN () const
base::ScalarFunction &	getObjectiveFunction () const
base::ScalarFunctionGradient *	getObjectiveGradient () const
base::ScalarFunctionHessian *	getObjectiveHessian () const
const base::DataVector &	getOptimalPoint () const
double	getOptimalValue () const
const base::DataVector &	getStartingPoint () const
virtual void	optimize ()=0
	Pure virtual method for optimization of the objective function.
void	setN (size_t N)
virtual void	setObjectiveFunction (const base::ScalarFunction &f)
virtual void	setObjectiveGradient (const base::ScalarFunctionGradient *fGradient)
virtual void	setObjectiveHessian (const base::ScalarFunctionHessian *fHessian)
void	setStartingPoint (const base::DataVector &startingPoint)
	UnconstrainedOptimizer (const base::ScalarFunction &f, const base::ScalarFunctionGradient *fGradient, const base::ScalarFunctionHessian *fHessian, size_t N=DEFAULT_N)
	Constructor.
	UnconstrainedOptimizer (const UnconstrainedOptimizer &other)
	Copy constructor.
virtual	~UnconstrainedOptimizer ()
	Destructor.

Static Public Attributes
static const size_t	DEFAULT_N = 1000
	default maximal number of iterations or function evaluations

Protected Attributes
std::unique_ptr< base::ScalarFunction >	f
	objective function
std::unique_ptr< base::ScalarFunctionGradient >	fGradient
	objective function gradient
std::unique_ptr< base::ScalarFunctionHessian >	fHessian
	objective function Hessian
base::DataVector	fHist
	search history vector (optimal values)
double	fOpt
	result of optimization (optimal function value)
size_t	N
	maximal number of iterations or function evaluations
base::DataVector	x0
	starting point
base::DataMatrix	xHist
	search history matrix (optimal points)
base::DataVector	xOpt
	result of optimization (location of optimum)

Detailed Description

Abstract class for optimizing objective functions.

Constructor & Destructor Documentation

UnconstrainedOptimizer() [1/2]

sgpp::optimization::optimizer::UnconstrainedOptimizer::UnconstrainedOptimizer ( const base::ScalarFunction &  f, const base::ScalarFunctionGradient *  fGradient, const base::ScalarFunctionHessian *  fHessian, size_t  N = DEFAULT_N  )

inlineexplicit

Constructor.

The starting point is set to \((0.5, \dotsc, 0.5)^{\mathrm{T}}\).

Parameters

f	function to optimize
fGradient	gradient of f (nullptr to omit)
fHessian	Hessian of f (nullptr to omit)
N	maximal number of iterations or function evaluations (depending on the implementation)

References f, fGradient, and fHessian.

UnconstrainedOptimizer() [2/2]

sgpp::optimization::optimizer::UnconstrainedOptimizer::UnconstrainedOptimizer ( const UnconstrainedOptimizer &  other )

inline

Copy constructor.

Parameters

other

optimizer to be copied

References fHist, fOpt, x0, xHist, and xOpt.

~UnconstrainedOptimizer()

virtual sgpp::optimization::optimizer::UnconstrainedOptimizer::~UnconstrainedOptimizer ( )

inlinevirtual

Destructor.

Member Function Documentation

clone()

virtual void sgpp::optimization::optimizer::UnconstrainedOptimizer::clone ( std::unique_ptr< UnconstrainedOptimizer > &  clone ) const

pure virtual

Pure virtual method for cloning the optimizer.

It should generate a pointer to the cloned object and it's used for parallel computations.

Parameters

[out]

clone

pointer to cloned object

Implemented in sgpp::optimization::optimizer::AugmentedLagrangian, sgpp::optimization::optimizer::LogBarrier, sgpp::optimization::optimizer::SquaredPenalty, sgpp::optimization::optimizer::AdaptiveGradientDescent, sgpp::optimization::optimizer::AdaptiveNewton, sgpp::optimization::optimizer::BFGS, sgpp::optimization::optimizer::CMAES, sgpp::optimization::optimizer::DifferentialEvolution, sgpp::optimization::optimizer::GradientDescent, sgpp::optimization::optimizer::MultiStart, sgpp::optimization::optimizer::NelderMead, sgpp::optimization::optimizer::Newton, sgpp::optimization::optimizer::NLCG, and sgpp::optimization::optimizer::Rprop.

Referenced by sgpp::optimization::optimizer::ConstrainedOptimizer::ConstrainedOptimizer(), sgpp::optimization::optimizer::ConstrainedOptimizer::ConstrainedOptimizer(), sgpp::optimization::optimizer::ConstrainedOptimizer::ConstrainedOptimizer(), sgpp::optimization::optimizer::MultiStart::MultiStart(), and sgpp::optimization::optimizer::MultiStart::optimize().

getHistoryOfOptimalPoints()

const base::DataMatrix & sgpp::optimization::optimizer::UnconstrainedOptimizer::getHistoryOfOptimalPoints ( ) const

inline

Returns

tall matrix (d columns) in which the k-th row indicates the best point after k iterations of the optimization, empty matrix on error or if not supported

References xHist.

Referenced by sgpp::optimization::optimizer::MultiStart::optimize().

getHistoryOfOptimalValues()

const base::DataVector & sgpp::optimization::optimizer::UnconstrainedOptimizer::getHistoryOfOptimalValues ( ) const

inline

Returns

vector in which the k-th entry indicates the best function value after k iterations of the optimization, empty vector on error or if not supported

References fHist.

getN()

size_t sgpp::optimization::optimizer::UnconstrainedOptimizer::getN ( ) const

inline

Returns

maximal number of iterations or function evaluations

References N.

getObjectiveFunction()

base::ScalarFunction & sgpp::optimization::optimizer::UnconstrainedOptimizer::getObjectiveFunction ( ) const

inline

Returns

objective function

References f.

getObjectiveGradient()

base::ScalarFunctionGradient * sgpp::optimization::optimizer::UnconstrainedOptimizer::getObjectiveGradient ( ) const

inline

Returns

objective gradient

References fGradient.

getObjectiveHessian()

base::ScalarFunctionHessian * sgpp::optimization::optimizer::UnconstrainedOptimizer::getObjectiveHessian ( ) const

inline

Returns

objective Hessian

References fHessian.

getOptimalPoint()

const base::DataVector & sgpp::optimization::optimizer::UnconstrainedOptimizer::getOptimalPoint ( ) const

inline

Returns

result of optimization (location of optimum), empty vector on error

References xOpt.

Referenced by sgpp::datadriven::OperationInverseRosenblattTransformation1DBspline::doTransformation1D(), sgpp::datadriven::OperationInverseRosenblattTransformation1DBsplineBoundary::doTransformation1D(), sgpp::datadriven::OperationInverseRosenblattTransformation1DBsplineClenshawCurtis::doTransformation1D(), sgpp::datadriven::OperationInverseRosenblattTransformation1DModBspline::doTransformation1D(), sgpp::datadriven::OperationInverseRosenblattTransformation1DModBsplineClenshawCurtis::doTransformation1D(), sgpp::datadriven::OperationInverseRosenblattTransformation1DModPoly::doTransformation1D(), sgpp::datadriven::OperationInverseRosenblattTransformation1DModPolyClenshawCurtis::doTransformation1D(), sgpp::datadriven::OperationInverseRosenblattTransformation1DPoly::doTransformation1D(), sgpp::datadriven::OperationInverseRosenblattTransformation1DPolyBoundary::doTransformation1D(), sgpp::datadriven::OperationInverseRosenblattTransformation1DPolyClenshawCurtis::doTransformation1D(), sgpp::datadriven::OperationInverseRosenblattTransformation1DPolyClenshawCurtisBoundary::doTransformation1D(), sgpp::datadriven::BayesianOptimization::fitScales(), sgpp::datadriven::BayesianOptimization::main(), sgpp::optimization::SplineResponseSurface::optimize(), sgpp::optimization::optimizer::MultiStart::optimize(), and sgpp::datadriven::MaximumLikelihoodCrossValidation::optimizeBandwidths().

getOptimalValue()

double sgpp::optimization::optimizer::UnconstrainedOptimizer::getOptimalValue ( ) const

inline

Returns

result of optimization (optimal function value), NaN on error

References fOpt.

Referenced by sgpp::datadriven::BayesianOptimization::main(), and sgpp::optimization::optimizer::MultiStart::optimize().

getStartingPoint()

const base::DataVector & sgpp::optimization::optimizer::UnconstrainedOptimizer::getStartingPoint ( ) const

inline

Returns

starting point

References x0.

optimize()

virtual void sgpp::optimization::optimizer::UnconstrainedOptimizer::optimize ( )

pure virtual

Pure virtual method for optimization of the objective function.

The result of the optimization process can be obtained by member functions, e.g., getOptimalPoint() and getOptimalValue().

Implemented in sgpp::optimization::optimizer::AugmentedLagrangian, sgpp::optimization::optimizer::LogBarrier, sgpp::optimization::optimizer::SquaredPenalty, sgpp::optimization::optimizer::AdaptiveGradientDescent, sgpp::optimization::optimizer::AdaptiveNewton, sgpp::optimization::optimizer::BFGS, sgpp::optimization::optimizer::CMAES, sgpp::optimization::optimizer::DifferentialEvolution, sgpp::optimization::optimizer::GradientDescent, sgpp::optimization::optimizer::MultiStart, sgpp::optimization::optimizer::NelderMead, sgpp::optimization::optimizer::Newton, sgpp::optimization::optimizer::NLCG, and sgpp::optimization::optimizer::Rprop.

Referenced by sgpp::optimization::optimizer::MultiStart::optimize().

setN()

void sgpp::optimization::optimizer::UnconstrainedOptimizer::setN ( size_t  N )

inline

Parameters

N	maximal number of iterations or function evaluations

References N.

Referenced by sgpp::optimization::optimizer::MultiStart::optimize().

setObjectiveFunction()

virtual void sgpp::optimization::optimizer::UnconstrainedOptimizer::setObjectiveFunction ( const base::ScalarFunction &  f )

inlinevirtual

Parameters

f	objective function

Reimplemented in sgpp::optimization::optimizer::MultiStart.

References f.

Referenced by sgpp::optimization::optimizer::MultiStart::setObjectiveFunction().

setObjectiveGradient()

virtual void sgpp::optimization::optimizer::UnconstrainedOptimizer::setObjectiveGradient ( const base::ScalarFunctionGradient *  fGradient )

inlinevirtual

Parameters

fGradient

objective gradient

Reimplemented in sgpp::optimization::optimizer::MultiStart.

References fGradient.

Referenced by sgpp::optimization::optimizer::MultiStart::setObjectiveGradient().

setObjectiveHessian()

virtual void sgpp::optimization::optimizer::UnconstrainedOptimizer::setObjectiveHessian ( const base::ScalarFunctionHessian *  fHessian )

inlinevirtual

Parameters

fHessian

objective Hessian

Reimplemented in sgpp::optimization::optimizer::MultiStart.

References fHessian.

Referenced by sgpp::optimization::optimizer::MultiStart::setObjectiveHessian().

setStartingPoint()

void sgpp::optimization::optimizer::UnconstrainedOptimizer::setStartingPoint ( const base::DataVector &  startingPoint )

inline

Parameters

startingPoint

starting point

References x0.

Referenced by sgpp::optimization::SplineResponseSurface::optimize(), and sgpp::optimization::optimizer::MultiStart::optimize().

Member Data Documentation

DEFAULT_N

const size_t sgpp::optimization::optimizer::UnconstrainedOptimizer::DEFAULT_N = 1000

static

default maximal number of iterations or function evaluations

f

std::unique_ptr<base::ScalarFunction> sgpp::optimization::optimizer::UnconstrainedOptimizer::f

protected

objective function

Referenced by sgpp::optimization::optimizer::AugmentedLagrangian::findFeasiblePoint(), getObjectiveFunction(), sgpp::optimization::optimizer::MultiStart::initialize(), sgpp::optimization::optimizer::AugmentedLagrangian::optimize(), sgpp::optimization::optimizer::LogBarrier::optimize(), sgpp::optimization::optimizer::SquaredPenalty::optimize(), sgpp::optimization::optimizer::AdaptiveGradientDescent::optimize(), sgpp::optimization::optimizer::AdaptiveNewton::optimize(), sgpp::optimization::optimizer::BFGS::optimize(), sgpp::optimization::optimizer::CMAES::optimize(), sgpp::optimization::optimizer::DifferentialEvolution::optimize(), sgpp::optimization::optimizer::GradientDescent::optimize(), sgpp::optimization::optimizer::MultiStart::optimize(), sgpp::optimization::optimizer::NelderMead::optimize(), sgpp::optimization::optimizer::Newton::optimize(), sgpp::optimization::optimizer::NLCG::optimize(), sgpp::optimization::optimizer::Rprop::optimize(), python.uq.refinement.RefinementStrategy.PredictiveRanking::rank(), setObjectiveFunction(), sgpp::optimization::optimizer::MultiStart::setObjectiveFunction(), and UnconstrainedOptimizer().

fGradient

std::unique_ptr<base::ScalarFunctionGradient> sgpp::optimization::optimizer::UnconstrainedOptimizer::fGradient

protected

objective function gradient

Referenced by getObjectiveGradient(), sgpp::optimization::optimizer::AugmentedLagrangian::optimize(), sgpp::optimization::optimizer::LogBarrier::optimize(), sgpp::optimization::optimizer::SquaredPenalty::optimize(), sgpp::optimization::optimizer::AdaptiveGradientDescent::optimize(), sgpp::optimization::optimizer::BFGS::optimize(), sgpp::optimization::optimizer::GradientDescent::optimize(), sgpp::optimization::optimizer::NLCG::optimize(), sgpp::optimization::optimizer::Rprop::optimize(), setObjectiveGradient(), sgpp::optimization::optimizer::MultiStart::setObjectiveGradient(), and UnconstrainedOptimizer().

fHessian

std::unique_ptr<base::ScalarFunctionHessian> sgpp::optimization::optimizer::UnconstrainedOptimizer::fHessian

protected

objective function Hessian

Referenced by getObjectiveHessian(), sgpp::optimization::optimizer::AdaptiveNewton::optimize(), sgpp::optimization::optimizer::Newton::optimize(), setObjectiveHessian(), sgpp::optimization::optimizer::MultiStart::setObjectiveHessian(), and UnconstrainedOptimizer().

fHist

base::DataVector sgpp::optimization::optimizer::UnconstrainedOptimizer::fHist

protected

search history vector (optimal values)

Referenced by getHistoryOfOptimalValues(), sgpp::optimization::optimizer::AugmentedLagrangian::optimize(), sgpp::optimization::optimizer::LogBarrier::optimize(), sgpp::optimization::optimizer::SquaredPenalty::optimize(), sgpp::optimization::optimizer::AdaptiveGradientDescent::optimize(), sgpp::optimization::optimizer::AdaptiveNewton::optimize(), sgpp::optimization::optimizer::BFGS::optimize(), sgpp::optimization::optimizer::CMAES::optimize(), sgpp::optimization::optimizer::DifferentialEvolution::optimize(), sgpp::optimization::optimizer::GradientDescent::optimize(), sgpp::optimization::optimizer::MultiStart::optimize(), sgpp::optimization::optimizer::NelderMead::optimize(), sgpp::optimization::optimizer::Newton::optimize(), sgpp::optimization::optimizer::NLCG::optimize(), sgpp::optimization::optimizer::Rprop::optimize(), and UnconstrainedOptimizer().

fOpt

double sgpp::optimization::optimizer::UnconstrainedOptimizer::fOpt

protected

result of optimization (optimal function value)

Referenced by getOptimalValue(), sgpp::optimization::optimizer::AugmentedLagrangian::optimize(), sgpp::optimization::optimizer::LogBarrier::optimize(), sgpp::optimization::optimizer::SquaredPenalty::optimize(), sgpp::optimization::optimizer::AdaptiveGradientDescent::optimize(), sgpp::optimization::optimizer::AdaptiveNewton::optimize(), sgpp::optimization::optimizer::BFGS::optimize(), sgpp::optimization::optimizer::CMAES::optimize(), sgpp::optimization::optimizer::DifferentialEvolution::optimize(), sgpp::optimization::optimizer::GradientDescent::optimize(), sgpp::optimization::optimizer::MultiStart::optimize(), sgpp::optimization::optimizer::NelderMead::optimize(), sgpp::optimization::optimizer::Newton::optimize(), sgpp::optimization::optimizer::NLCG::optimize(), sgpp::optimization::optimizer::Rprop::optimize(), and UnconstrainedOptimizer().

N

size_t sgpp::optimization::optimizer::UnconstrainedOptimizer::N

protected

maximal number of iterations or function evaluations

Referenced by getN(), sgpp::optimization::optimizer::AugmentedLagrangian::optimize(), sgpp::optimization::optimizer::LogBarrier::optimize(), sgpp::optimization::optimizer::SquaredPenalty::optimize(), sgpp::optimization::optimizer::AdaptiveGradientDescent::optimize(), sgpp::optimization::optimizer::AdaptiveNewton::optimize(), sgpp::optimization::optimizer::BFGS::optimize(), sgpp::optimization::optimizer::CMAES::optimize(), sgpp::optimization::optimizer::DifferentialEvolution::optimize(), sgpp::optimization::optimizer::GradientDescent::optimize(), sgpp::optimization::optimizer::MultiStart::optimize(), sgpp::optimization::optimizer::NelderMead::optimize(), sgpp::optimization::optimizer::Newton::optimize(), sgpp::optimization::optimizer::NLCG::optimize(), sgpp::optimization::optimizer::Rprop::optimize(), and setN().

x0

base::DataVector sgpp::optimization::optimizer::UnconstrainedOptimizer::x0

protected

starting point

Referenced by getStartingPoint(), sgpp::optimization::optimizer::AugmentedLagrangian::optimize(), sgpp::optimization::optimizer::LogBarrier::optimize(), sgpp::optimization::optimizer::SquaredPenalty::optimize(), sgpp::optimization::optimizer::AdaptiveGradientDescent::optimize(), sgpp::optimization::optimizer::AdaptiveNewton::optimize(), sgpp::optimization::optimizer::BFGS::optimize(), sgpp::optimization::optimizer::CMAES::optimize(), sgpp::optimization::optimizer::GradientDescent::optimize(), sgpp::optimization::optimizer::MultiStart::optimize(), sgpp::optimization::optimizer::NelderMead::optimize(), sgpp::optimization::optimizer::Newton::optimize(), sgpp::optimization::optimizer::NLCG::optimize(), sgpp::optimization::optimizer::Rprop::optimize(), setStartingPoint(), and UnconstrainedOptimizer().

xHist

base::DataMatrix sgpp::optimization::optimizer::UnconstrainedOptimizer::xHist

protected

search history matrix (optimal points)

Referenced by getHistoryOfOptimalPoints(), sgpp::optimization::optimizer::AugmentedLagrangian::optimize(), sgpp::optimization::optimizer::LogBarrier::optimize(), sgpp::optimization::optimizer::SquaredPenalty::optimize(), sgpp::optimization::optimizer::AdaptiveGradientDescent::optimize(), sgpp::optimization::optimizer::AdaptiveNewton::optimize(), sgpp::optimization::optimizer::BFGS::optimize(), sgpp::optimization::optimizer::CMAES::optimize(), sgpp::optimization::optimizer::DifferentialEvolution::optimize(), sgpp::optimization::optimizer::GradientDescent::optimize(), sgpp::optimization::optimizer::MultiStart::optimize(), sgpp::optimization::optimizer::NelderMead::optimize(), sgpp::optimization::optimizer::Newton::optimize(), sgpp::optimization::optimizer::NLCG::optimize(), sgpp::optimization::optimizer::Rprop::optimize(), and UnconstrainedOptimizer().

xOpt

base::DataVector sgpp::optimization::optimizer::UnconstrainedOptimizer::xOpt

protected

result of optimization (location of optimum)

Referenced by getOptimalPoint(), sgpp::optimization::optimizer::AugmentedLagrangian::optimize(), sgpp::optimization::optimizer::LogBarrier::optimize(), sgpp::optimization::optimizer::SquaredPenalty::optimize(), sgpp::optimization::optimizer::AdaptiveGradientDescent::optimize(), sgpp::optimization::optimizer::AdaptiveNewton::optimize(), sgpp::optimization::optimizer::BFGS::optimize(), sgpp::optimization::optimizer::CMAES::optimize(), sgpp::optimization::optimizer::DifferentialEvolution::optimize(), sgpp::optimization::optimizer::GradientDescent::optimize(), sgpp::optimization::optimizer::MultiStart::optimize(), sgpp::optimization::optimizer::NelderMead::optimize(), sgpp::optimization::optimizer::Newton::optimize(), sgpp::optimization::optimizer::NLCG::optimize(), sgpp::optimization::optimizer::Rprop::optimize(), and UnconstrainedOptimizer().

---

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

• optimization/src/sgpp/optimization/optimizer/unconstrained/UnconstrainedOptimizer.hpp