Provides a simplified interface for solving a wide range of optimization problems. More...

#include <ROL_OptimizationSolver.hpp>

Public Member Functions
	OptimizationSolver (OptimizationProblem< Real > &opt, ROL::ParameterList &parlist)
	Constructor.
std::vector< std::string >	getOutput (void) const
	Returns iteration history as a vector of strings.
int	solve (const ROL::Ptr< StatusTest< Real > > &status=ROL::nullPtr, const bool combineStatus=true)
	Solve optimization problem with no iteration output.
int	solve (std::ostream &outStream, const ROL::Ptr< StatusTest< Real > > &status=ROL::nullPtr, const bool combineStatus=true)
	Solve optimization problem.
ROL::Ptr< const AlgorithmState< Real > >	getAlgorithmState (void) const
	Return the AlgorithmState.
void	resetAlgorithmState (void)
	Reset the AlgorithmState.
void	reset (const bool resetAlgo=true)
	Reset both Algorithm and Step.
std::string	getStepName (void) const
	Grab step name (after check for consistency).

Private Attributes
ROL::Ptr< Algorithm< Real > >	algo_
ROL::Ptr< Step< Real > >	step_
ROL::Ptr< StatusTest< Real > >	status0_
ROL::Ptr< CombinedStatusTest< Real > >	status_
ROL::Ptr< AlgorithmState< Real > >	state_
ROL::Ptr< Vector< Real > >	x_
ROL::Ptr< Vector< Real > >	g_
ROL::Ptr< Vector< Real > >	l_
ROL::Ptr< Vector< Real > >	c_
ROL::Ptr< Objective< Real > >	obj_
ROL::Ptr< BoundConstraint< Real > >	bnd_
ROL::Ptr< Constraint< Real > >	con_
std::vector< std::string >	output_
EProblem	problemType_
EStep	stepType_
std::string	stepname_
Real	pen_

Detailed Description

template<class Real>
class ROL::OptimizationSolver< Real >

Provides a simplified interface for solving a wide range of optimization problems.

Definition at line 29 of file ROL_OptimizationSolver.hpp.

Constructor & Destructor Documentation

◆ OptimizationSolver()

template<class Real>

ROL::OptimizationSolver< Real >::OptimizationSolver	(	OptimizationProblem< Real > &	opt,
		ROL::ParameterList &	parlist )

inline

Constructor.

Parameters

[in]	opt	the OptimizationProblem to be solved
[in]	parlist	algorithm and step input parameters

Definition at line 64 of file ROL_OptimizationSolver.hpp.

References bnd_, c_, con_, ROL::EStepToString(), g_, ROL::OptimizationProblem< Real >::getBoundConstraint(), ROL::OptimizationProblem< Real >::getConstraint(), ROL::OptimizationProblem< Real >::getMultiplierVector(), ROL::OptimizationProblem< Real >::getObjective(), ROL::OptimizationProblem< Real >::getProblemType(), ROL::OptimizationProblem< Real >::getSolutionVector(), ROL::StatusTestFactory< Real >::getStatusTest(), ROL::isCompatibleStep(), l_, obj_, pen_, problemType_, state_, status0_, status_, step_, ROL::STEP_AUGMENTEDLAGRANGIAN, ROL::STEP_BUNDLE, ROL::STEP_COMPOSITESTEP, ROL::STEP_FLETCHER, ROL::STEP_INTERIORPOINT, ROL::STEP_MOREAUYOSIDAPENALTY, ROL::STEP_TRUSTREGION, ROL::stepFactory(), stepname_, stepType_, ROL::StringToEStep(), ROL::TYPE_B, ROL::TYPE_E, ROL::TYPE_EB, ROL::TYPE_LAST, ROL::TYPE_U, and x_.

Member Function Documentation

◆ getOutput()

template<class Real>

std::vector< std::string > ROL::OptimizationSolver< Real >::getOutput ( void ) const

inline

Returns iteration history as a vector of strings.

Definition at line 169 of file ROL_OptimizationSolver.hpp.

References output_.

◆ solve() [1/2]

template<class Real>

int ROL::OptimizationSolver< Real >::solve	(	const ROL::Ptr< StatusTest< Real > > &	status = ROL::nullPtr,
		const bool	combineStatus = true )

inline

Solve optimization problem with no iteration output.

Parameters

[in]	status	is a user-defined StatusTest
[in]	combineStatus	if true, the user-defined StatusTest will be combined with the default StatusTest

Definition at line 180 of file ROL_OptimizationSolver.hpp.

References solve().

Referenced by main(), solve(), and ROL::SROMGenerator< Real >::SROMGenerator().

◆ solve() [2/2]

template<class Real>

int ROL::OptimizationSolver< Real >::solve	(	std::ostream &	outStream,
		const ROL::Ptr< StatusTest< Real > > &	status = ROL::nullPtr,
		const bool	combineStatus = true )

inline

Solve optimization problem.

Parameters

[in]	outStream	is the output stream to collect iteration history
[in]	status	is a user-defined StatusTest
[in]	combineStatus	if true, the user-defined StatusTest will be combined with the default StatusTest

Definition at line 194 of file ROL_OptimizationSolver.hpp.

References algo_, bnd_, c_, con_, g_, l_, obj_, output_, problemType_, state_, status0_, status_, step_, ROL::TYPE_B, ROL::TYPE_E, ROL::TYPE_EB, ROL::TYPE_LAST, ROL::TYPE_U, and x_.

◆ getAlgorithmState()

template<class Real>

ROL::Ptr< const AlgorithmState< Real > > ROL::OptimizationSolver< Real >::getAlgorithmState ( void ) const

inline

Return the AlgorithmState.

Definition at line 238 of file ROL_OptimizationSolver.hpp.

References state_.

Referenced by main().

◆ resetAlgorithmState()

template<class Real>

void ROL::OptimizationSolver< Real >::resetAlgorithmState ( void )

inline

Reset the AlgorithmState.

This function does not reset the Step or the StepState.

Definition at line 248 of file ROL_OptimizationSolver.hpp.

References state_.

Referenced by reset().

◆ reset()

template<class Real>

void ROL::OptimizationSolver< Real >::reset ( const bool resetAlgo = true )

inline

Reset both Algorithm and Step.

Parameters

[in] resetAlgo if true, then AlgorithmState will be reset

This function will reset the AlgorithmState and reinitialize the Step. This function does not permit changing the Step specified upon construction. To change the Step, reinitialize the OptimizationSolver.

Definition at line 263 of file ROL_OptimizationSolver.hpp.

References l_, obj_, pen_, reset(), resetAlgorithmState(), step_, ROL::STEP_AUGMENTEDLAGRANGIAN, ROL::STEP_INTERIORPOINT, ROL::STEP_MOREAUYOSIDAPENALTY, and stepType_.

Referenced by reset().