Implementation of the probing strategy for computing the barrier parameter. More...

#include <IpQualityFunctionMuOracle.hpp>

Inheritance diagram for Ipopt::QualityFunctionMuOracle:

Public Types
Public enums.
Some of those are also used for the quality function.
enum	NormEnum { NM_NORM_1 = 0 , NM_NORM_2_SQUARED , NM_NORM_MAX , NM_NORM_2 }
	enum for norm type More...

enum	CentralityEnum { CEN_NONE = 0 , CEN_LOG , CEN_RECIPROCAL , CEN_CUBED_RECIPROCAL }
	enum for centrality type More...

enum	BalancingTermEnum { BT_NONE = 0 , BT_CUBIC }
	enum for the quality function balancing term type More...

Public Member Functions
virtual bool	InitializeImpl (const OptionsList &options, const std::string &prefix)
	Implementation of the initialization method that has to be overloaded by for each derived class.

virtual bool	CalculateMu (Number mu_min, Number mu_max, Number &new_mu)
	Method for computing the value of the barrier parameter that could be used in the current iteration (using the LOQO formula).

Constructors/Destructors
	QualityFunctionMuOracle (const SmartPtr< PDSystemSolver > &pd_solver)
	Constructor.

virtual	~QualityFunctionMuOracle ()
	Destructor.

Public Member Functions inherited from Ipopt::MuOracle
	MuOracle ()
	Default Constructor.

virtual	~MuOracle ()
	Destructor.

Public Member Functions inherited from Ipopt::AlgorithmStrategyObject
bool	Initialize (const Journalist &jnlst, IpoptNLP &ip_nlp, IpoptData &ip_data, IpoptCalculatedQuantities &ip_cq, const OptionsList &options, const std::string &prefix)
	This method is called every time the algorithm starts again - it is used to reset any internal state.

bool	ReducedInitialize (const Journalist &jnlst, const OptionsList &options, const std::string &prefix)
	Reduced version of the Initialize method, which does not require special Ipopt information.

	AlgorithmStrategyObject ()
	Default Constructor.

virtual	~AlgorithmStrategyObject ()
	Destructor.

Public Member Functions inherited from Ipopt::ReferencedObject
	ReferencedObject ()

virtual	~ReferencedObject ()

Index	ReferenceCount () const

void	AddRef (const Referencer *referencer) const

void	ReleaseRef (const Referencer *referencer) const

Static Public Member Functions
static void	RegisterOptions (SmartPtr< RegisteredOptions > roptions)

Private Member Functions
Number	CalculateQualityFunction (Number sigma, const Vector &step_aff_x_L, const Vector &step_aff_x_U, const Vector &step_aff_s_L, const Vector &step_aff_s_U, const Vector &step_aff_y_c, const Vector &step_aff_y_d, const Vector &step_aff_z_L, const Vector &step_aff_z_U, const Vector &step_aff_v_L, const Vector &step_aff_v_U, const Vector &step_cen_x_L, const Vector &step_cen_x_U, const Vector &step_cen_s_L, const Vector &step_cen_s_U, const Vector &step_cen_y_c, const Vector &step_cen_y_d, const Vector &step_cen_z_L, const Vector &step_cen_z_U, const Vector &step_cen_v_L, const Vector &step_cen_v_U)
	Auxiliary function for computing the average complementarity at a point, given step sizes and step.

Number	PerformGoldenSection (Number sigma_up, Number q_up, Number sigma_lo, Number q_lo, Number sigma_tol, Number qf_tol, const Vector &step_aff_x_L, const Vector &step_aff_x_U, const Vector &step_aff_s_L, const Vector &step_aff_s_U, const Vector &step_aff_y_c, const Vector &step_aff_y_d, const Vector &step_aff_z_L, const Vector &step_aff_z_U, const Vector &step_aff_v_L, const Vector &step_aff_v_U, const Vector &step_cen_x_L, const Vector &step_cen_x_U, const Vector &step_cen_s_L, const Vector &step_cen_s_U, const Vector &step_cen_y_c, const Vector &step_cen_y_d, const Vector &step_cen_z_L, const Vector &step_cen_z_U, const Vector &step_cen_v_L, const Vector &step_cen_v_U)
	Auxiliary function performing the golden section.

Default Compiler Generated Methods
(Hidden to avoid implicit creation/calling). These methods are not implemented and we do not want the compiler to implement them for us, so we declare them private and do not define them. This ensures that they will not be implicitly created/called.
	QualityFunctionMuOracle ()
	Default Constructor.

	QualityFunctionMuOracle (const QualityFunctionMuOracle &)
	Copy Constructor.

void	operator= (const QualityFunctionMuOracle &)
	Default Assignment Operator.


Number	ScaleSigma (Number sigma)
	Auxiliary functions for scaling the sigma axis in the golden section procedure.

Number	UnscaleSigma (Number scaled_sigma)

Private Attributes
SmartPtr< PDSystemSolver >	pd_solver_
	Pointer to the object that should be used to solve the primal-dual system.

Index	count_qf_evals_

Algorithmic parameters
Auxiliary function performing the golden section in the logarithmic scale
Number	sigma_max_
	Upper bound on centering parameter sigma.

Number	sigma_min_
	Lower bound on centering parameter sigma.

NormEnum	quality_function_norm_
	Norm to be used for the quality function.

CentralityEnum	quality_function_centrality_
	Flag indicating how centrality should be involved in the quality function.

BalancingTermEnum	quality_function_balancing_term_
	Flag indicating whether we use a balancing term in the quality function.

Number	quality_function_section_sigma_tol_
	Relative tolerance for golden bi-section algorithm in sigma space.

Number	quality_function_section_qf_tol_
	Relative tolerance for golden bi-section algorithm in function value space.

Index	quality_function_max_section_steps_
	Maximal number of bi-section steps in the golden section search for sigma.

Temporary work space vectors.
We use those to avoid repeated reallocation in CalculateQualityFunction.
SmartPtr< Vector >	tmp_step_x_L_

SmartPtr< Vector >	tmp_step_x_U_

SmartPtr< Vector >	tmp_step_s_L_

SmartPtr< Vector >	tmp_step_s_U_

SmartPtr< Vector >	tmp_step_z_L_

SmartPtr< Vector >	tmp_step_z_U_

SmartPtr< Vector >	tmp_step_v_L_

SmartPtr< Vector >	tmp_step_v_U_

SmartPtr< Vector >	tmp_slack_x_L_

SmartPtr< Vector >	tmp_slack_x_U_

SmartPtr< Vector >	tmp_slack_s_L_

SmartPtr< Vector >	tmp_slack_s_U_

SmartPtr< Vector >	tmp_z_L_

SmartPtr< Vector >	tmp_z_U_

SmartPtr< Vector >	tmp_v_L_

SmartPtr< Vector >	tmp_v_U_

Quantities used many times in CalculateQualityFunction,
which we store here instead of retrieving them from cache every time. I (AW) don't know if that really makes a difference, but some of those things showed up in gprof.
bool	initialized_

Index	n_dual_

Index	n_pri_

Index	n_comp_

SmartPtr< const Vector >	curr_slack_x_L_

SmartPtr< const Vector >	curr_slack_x_U_

SmartPtr< const Vector >	curr_slack_s_L_

SmartPtr< const Vector >	curr_slack_s_U_

SmartPtr< const Vector >	curr_z_L_

SmartPtr< const Vector >	curr_z_U_

SmartPtr< const Vector >	curr_v_L_

SmartPtr< const Vector >	curr_v_U_

Number	curr_grad_lag_x_asum_

Number	curr_grad_lag_s_asum_

Number	curr_c_asum_

Number	curr_d_minus_s_asum_

Number	curr_grad_lag_x_nrm2_

Number	curr_grad_lag_s_nrm2_

Number	curr_c_nrm2_

Number	curr_d_minus_s_nrm2_

Number	curr_grad_lag_x_amax_

Number	curr_grad_lag_s_amax_

Number	curr_c_amax_

Number	curr_d_minus_s_amax_

Additional Inherited Members
Protected Member Functions inherited from Ipopt::AlgorithmStrategyObject
const Journalist &	Jnlst () const

IpoptNLP &	IpNLP () const

IpoptData &	IpData () const

IpoptCalculatedQuantities &	IpCq () const

bool	HaveIpData () const

Detailed Description

Implementation of the probing strategy for computing the barrier parameter.

Definition at line 20 of file IpQualityFunctionMuOracle.hpp.

Member Enumeration Documentation

◆ NormEnum

enum Ipopt::QualityFunctionMuOracle::NormEnum

enum for norm type

Enumerator
NM_NORM_1
NM_NORM_2_SQUARED
NM_NORM_MAX
NM_NORM_2

Definition at line 58 of file IpQualityFunctionMuOracle.hpp.

◆ CentralityEnum

enum Ipopt::QualityFunctionMuOracle::CentralityEnum

enum for centrality type

Enumerator
CEN_NONE
CEN_LOG
CEN_RECIPROCAL
CEN_CUBED_RECIPROCAL

Definition at line 67 of file IpQualityFunctionMuOracle.hpp.

◆ BalancingTermEnum

enum Ipopt::QualityFunctionMuOracle::BalancingTermEnum

enum for the quality function balancing term type

Enumerator
BT_NONE
BT_CUBIC

Definition at line 76 of file IpQualityFunctionMuOracle.hpp.

Constructor & Destructor Documentation

◆ QualityFunctionMuOracle() [1/3]

Ipopt::QualityFunctionMuOracle::QualityFunctionMuOracle ( const SmartPtr< PDSystemSolver > & pd_solver )

Constructor.

◆ ~QualityFunctionMuOracle()

virtual Ipopt::QualityFunctionMuOracle::~QualityFunctionMuOracle ( )

virtual

Destructor.

◆ QualityFunctionMuOracle() [2/3]

Ipopt::QualityFunctionMuOracle::QualityFunctionMuOracle ( )

private

Default Constructor.

◆ QualityFunctionMuOracle() [3/3]

Ipopt::QualityFunctionMuOracle::QualityFunctionMuOracle ( const QualityFunctionMuOracle & )

private

Copy Constructor.

Member Function Documentation

◆ InitializeImpl()

virtual bool Ipopt::QualityFunctionMuOracle::InitializeImpl	(	const OptionsList &	options,
		const std::string &	prefix
	)

virtual

Implementation of the initialization method that has to be overloaded by for each derived class.

Implements Ipopt::MuOracle.

◆ CalculateMu()

virtual bool Ipopt::QualityFunctionMuOracle::CalculateMu	(	Number	mu_min,
		Number	mu_max,
		Number &	new_mu
	)

virtual

Method for computing the value of the barrier parameter that could be used in the current iteration (using the LOQO formula).

Implements Ipopt::MuOracle.

◆ RegisterOptions()

static void Ipopt::QualityFunctionMuOracle::RegisterOptions ( SmartPtr< RegisteredOptions > roptions )

static

◆ operator=()

void Ipopt::QualityFunctionMuOracle::operator= ( const QualityFunctionMuOracle & )

private

Default Assignment Operator.

◆ CalculateQualityFunction()

Number Ipopt::QualityFunctionMuOracle::CalculateQualityFunction	(	Number	sigma,
		const Vector &	step_aff_x_L,
		const Vector &	step_aff_x_U,
		const Vector &	step_aff_s_L,
		const Vector &	step_aff_s_U,
		const Vector &	step_aff_y_c,
		const Vector &	step_aff_y_d,
		const Vector &	step_aff_z_L,
		const Vector &	step_aff_z_U,
		const Vector &	step_aff_v_L,
		const Vector &	step_aff_v_U,
		const Vector &	step_cen_x_L,
		const Vector &	step_cen_x_U,
		const Vector &	step_cen_s_L,
		const Vector &	step_cen_s_U,
		const Vector &	step_cen_y_c,
		const Vector &	step_cen_y_d,
		const Vector &	step_cen_z_L,
		const Vector &	step_cen_z_U,
		const Vector &	step_cen_v_L,
		const Vector &	step_cen_v_U
	)

private

Auxiliary function for computing the average complementarity at a point, given step sizes and step.

◆ PerformGoldenSection()

Number Ipopt::QualityFunctionMuOracle::PerformGoldenSection	(	Number	sigma_up,
		Number	q_up,
		Number	sigma_lo,
		Number	q_lo,
		Number	sigma_tol,
		Number	qf_tol,
		const Vector &	step_aff_x_L,
		const Vector &	step_aff_x_U,
		const Vector &	step_aff_s_L,
		const Vector &	step_aff_s_U,
		const Vector &	step_aff_y_c,
		const Vector &	step_aff_y_d,
		const Vector &	step_aff_z_L,
		const Vector &	step_aff_z_U,
		const Vector &	step_aff_v_L,
		const Vector &	step_aff_v_U,
		const Vector &	step_cen_x_L,
		const Vector &	step_cen_x_U,
		const Vector &	step_cen_s_L,
		const Vector &	step_cen_s_U,
		const Vector &	step_cen_y_c,
		const Vector &	step_cen_y_d,
		const Vector &	step_cen_z_L,
		const Vector &	step_cen_z_U,
		const Vector &	step_cen_v_L,
		const Vector &	step_cen_v_U
	)