Base class of implementations of descent methods for optimization. More...

#include <num_collect/opt/descent_method_base.h>

Inheritance diagram for num_collect::opt::descent_method_base< Derived, LineSearcher >:

Collaboration diagram for num_collect::opt::descent_method_base< Derived, LineSearcher >:

Public Types
using	line_searcher_type = LineSearcher
	Type of class to perform line search.

using	objective_function_type
	Type of the objective function.

using	value_type = typename objective_function_type::value_type
	Type of function values.

using	variable_type = typename objective_function_type::variable_type
	Type of variables.

Public Member Functions
auto	calc_direction () -> variable_type
	Calculate search direction.

auto	evaluations () const noexcept -> index_type
	Get the number of function evaluations.

auto	gradient () const -> const variable_type &
	Get gradient for current optimal variable.

auto	gradient_norm () const -> value_type
	Calculate norm of gradient.

void	init (const variable_type &init_variable)
	Initialize.

auto	is_stop_criteria_satisfied () const -> bool
	Determine if stopping criteria of the algorithm are satisfied.

void	iterate ()
	Iterate the algorithm once.

auto	iterations () const noexcept -> index_type
	Get the number of iterations.

auto	line_searcher () -> line_searcher_type &
	Access object to perform line search.

auto	line_searcher () const -> const line_searcher_type &
	Access object to perform line search.

auto	opt_value () const -> const value_type &
	Get current optimal value.

auto	opt_variable () const -> const variable_type &
	Get current optimal variable.

auto	tol_gradient_norm (const value_type &value) -> Derived &
	Set tolerance of norm of gradient.

Public Member Functions inherited from num_collect::opt::optimizer_base< Derived >
	optimizer_base (logging::log_tag_view tag)
	Constructor.

auto	evaluations () const noexcept -> index_type
	Get the number of function evaluations.

auto	iterations () const noexcept -> index_type
	Get the number of iterations.

auto	opt_value () const
	Get current optimal value.

auto	opt_variable () const
	Get current optimal variable.

Public Member Functions inherited from num_collect::base::iterative_solver_base< Derived >
	iterative_solver_base (logging::log_tag_view tag)
	Constructor.

auto	is_stop_criteria_satisfied () const -> bool
	Determine if stopping criteria of the algorithm are satisfied.

void	iterate ()
	Iterate the algorithm once.

void	solve ()
	Solve the problem.

Public Member Functions inherited from num_collect::logging::logging_mixin
	logging_mixin (log_tag_view tag)
	Constructor.

template<typename Child >
void	configure_child_algorithm_logger_if_exists (Child &child)
	Configure a logger of a child algorithm if exists.

auto	logger () const noexcept -> const num_collect::logging::logger &
	Access to the logger.

auto	logger () noexcept -> num_collect::logging::logger &
	Access to the logger.

Public Member Functions inherited from num_collect::logging::iterations::iteration_logger_mixin< Derived >
	iteration_logger_mixin ()=default
	Constructor.

void	configure_iteration_logger (num_collect::logging::iterations::iteration_logger< Derived > &iteration_logger) const
	Configure an iteration logger.

auto	initialize_iteration_logger () -> num_collect::logging::iterations::iteration_logger< Derived > &
	Get the iteration logger.

Protected Member Functions
	descent_method_base (logging::log_tag_view tag, const objective_function_type &obj_fun=objective_function_type())
	Constructor.

Protected Member Functions inherited from num_collect::base::iterative_solver_base< Derived >
auto	derived () const noexcept -> const Derived &
	Access derived object.

auto	derived () noexcept -> Derived &
	Access derived object.

Private Attributes
index_type	iterations_ {0}
	Number of iterations.

line_searcher_type	line_searcher_
	Object to perform line search.

value_type	tol_grad_norm_ {default_tol_grad_norm}
	Tolerance of norm of gradient.

Static Private Attributes
static const auto	default_tol_grad_norm
	Default tolerance of norm of gradient.

Detailed Description

template<typename Derived, concepts::line_searcher LineSearcher>
class num_collect::opt::descent_method_base< Derived, LineSearcher >

Base class of implementations of descent methods for optimization.

Template Parameters

Derived	Type of derived class.
LineSearcher	Type of class to perform line search.

Definition at line 36 of file descent_method_base.h.

Member Typedef Documentation

◆ line_searcher_type

template<typename Derived , concepts::line_searcher LineSearcher>

using num_collect::opt::descent_method_base< Derived, LineSearcher >::line_searcher_type = LineSearcher

Type of class to perform line search.

Definition at line 39 of file descent_method_base.h.

◆ objective_function_type

template<typename Derived , concepts::line_searcher LineSearcher>

using num_collect::opt::descent_method_base< Derived, LineSearcher >::objective_function_type

Initial value:

typename line_searcher_type::objective_function_type

Type of the objective function.

Definition at line 42 of file descent_method_base.h.

◆ value_type

template<typename Derived , concepts::line_searcher LineSearcher>

using num_collect::opt::descent_method_base< Derived, LineSearcher >::value_type = typename objective_function_type::value_type

Type of function values.

Definition at line 49 of file descent_method_base.h.

◆ variable_type

template<typename Derived , concepts::line_searcher LineSearcher>

using num_collect::opt::descent_method_base< Derived, LineSearcher >::variable_type = typename objective_function_type::variable_type

Type of variables.

Definition at line 46 of file descent_method_base.h.

Constructor & Destructor Documentation

◆ descent_method_base()

template<typename Derived , concepts::line_searcher LineSearcher>

num_collect::opt::descent_method_base< Derived, LineSearcher >::descent_method_base	(	logging::log_tag_view	tag,
		const objective_function_type &	obj_fun = objective_function_type() )

inlineexplicitprotected

Constructor.

Parameters

[in]	tag	Log tag.
[in]	obj_fun	Objective function.

Definition at line 170 of file descent_method_base.h.

Member Function Documentation

◆ calc_direction()

template<typename Derived , concepts::line_searcher LineSearcher>

auto num_collect::opt::descent_method_base< Derived, LineSearcher >::calc_direction ( ) -> variable_type

inlinenodiscard

Calculate search direction.

Returns: Search direction.

Definition at line 66 of file descent_method_base.h.

◆ evaluations()

template<typename Derived , concepts::line_searcher LineSearcher>

auto num_collect::opt::descent_method_base< Derived, LineSearcher >::evaluations ( ) const -> index_type

inlinenodiscardnoexcept

Get the number of function evaluations.

Returns: Number of function evaluations.

Definition at line 127 of file descent_method_base.h.

◆ gradient()

template<typename Derived , concepts::line_searcher LineSearcher>

auto num_collect::opt::descent_method_base< Derived, LineSearcher >::gradient ( ) const -> const variable_type&

inlinenodiscard

Get gradient for current optimal variable.

Returns: Gradient for current optimal variable.

Definition at line 136 of file descent_method_base.h.

◆ gradient_norm()

template<typename Derived , concepts::line_searcher LineSearcher>

auto num_collect::opt::descent_method_base< Derived, LineSearcher >::gradient_norm ( ) const -> value_type

inlinenodiscard

Calculate norm of gradient.

Returns: Norm of gradient.

Definition at line 145 of file descent_method_base.h.

◆ init()

template<typename Derived , concepts::line_searcher LineSearcher>

void num_collect::opt::descent_method_base< Derived, LineSearcher >::init ( const variable_type & init_variable )

inline

Initialize.

Parameters

[in] init_variable Initial variable.

Definition at line 56 of file descent_method_base.h.

◆ is_stop_criteria_satisfied()

template<typename Derived , concepts::line_searcher LineSearcher>

auto num_collect::opt::descent_method_base< Derived, LineSearcher >::is_stop_criteria_satisfied ( ) const -> bool

inlinenodiscard

Determine if stopping criteria of the algorithm are satisfied.

Returns: If stopping criteria of the algorithm are satisfied.

Definition at line 81 of file descent_method_base.h.

◆ iterate()

template<typename Derived , concepts::line_searcher LineSearcher>

void num_collect::opt::descent_method_base< Derived, LineSearcher >::iterate ( )

inline

Iterate the algorithm once.

Warning: Any required initializations (with init functions) are assumed to have been done.

Definition at line 73 of file descent_method_base.h.

◆ iterations()

template<typename Derived , concepts::line_searcher LineSearcher>

auto num_collect::opt::descent_method_base< Derived, LineSearcher >::iterations ( ) const -> index_type

inlinenodiscardnoexcept

Get the number of iterations.

Returns: Number of iterations.

Definition at line 120 of file descent_method_base.h.

◆ line_searcher() [1/2]

template<typename Derived , concepts::line_searcher LineSearcher>

auto num_collect::opt::descent_method_base< Derived, LineSearcher >::line_searcher ( ) -> line_searcher_type&

inlinenodiscard

Access object to perform line search.

Returns: Object to perform line search.

Definition at line 90 of file descent_method_base.h.

◆ line_searcher() [2/2]

template<typename Derived , concepts::line_searcher LineSearcher>

auto num_collect::opt::descent_method_base< Derived, LineSearcher >::line_searcher ( ) const -> const line_searcher_type&

inlinenodiscard

Access object to perform line search.

Returns: Object to perform line search.

Definition at line 99 of file descent_method_base.h.

◆ opt_value()

template<typename Derived , concepts::line_searcher LineSearcher>

auto num_collect::opt::descent_method_base< Derived, LineSearcher >::opt_value ( ) const -> const value_type&

inlinenodiscard

Get current optimal value.

Returns: Current optimal value.

Definition at line 113 of file descent_method_base.h.

◆ opt_variable()

template<typename Derived , concepts::line_searcher LineSearcher>

auto num_collect::opt::descent_method_base< Derived, LineSearcher >::opt_variable ( ) const -> const variable_type&

inlinenodiscard

Get current optimal variable.

Returns: Current optimal variable.

Definition at line 106 of file descent_method_base.h.

◆ tol_gradient_norm()

template<typename Derived , concepts::line_searcher LineSearcher>

auto num_collect::opt::descent_method_base< Derived, LineSearcher >::tol_gradient_norm ( const value_type & value ) -> Derived&

inline

Set tolerance of norm of gradient.

Parameters

[in] value Value.

Returns: This object.

Definition at line 155 of file descent_method_base.h.

Member Data Documentation

◆ default_tol_grad_norm

template<typename Derived , concepts::line_searcher LineSearcher>

const auto num_collect::opt::descent_method_base< Derived, LineSearcher >::default_tol_grad_norm

inlinestaticprivate

Initial value:

=

static_cast<value_type>(1e-3)

num_collect::opt::descent_method_base::value_type

typename objective_function_type::value_type value_type

Type of function values.

Definition descent_method_base.h:49

Default tolerance of norm of gradient.

Definition at line 182 of file descent_method_base.h.

◆ iterations_

template<typename Derived , concepts::line_searcher LineSearcher>

index_type num_collect::opt::descent_method_base< Derived, LineSearcher >::iterations_ {0}

private

Number of iterations.

Definition at line 179 of file descent_method_base.h.

◆ line_searcher_

template<typename Derived , concepts::line_searcher LineSearcher>

line_searcher_type num_collect::opt::descent_method_base< Derived, LineSearcher >::line_searcher_

private

Object to perform line search.

Definition at line 176 of file descent_method_base.h.

◆ tol_grad_norm_

template<typename Derived , concepts::line_searcher LineSearcher>

value_type num_collect::opt::descent_method_base< Derived, LineSearcher >::tol_grad_norm_ {default_tol_grad_norm}

private

Tolerance of norm of gradient.

Definition at line 186 of file descent_method_base.h.

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

include/num_collect/opt/descent_method_base.h

Public Types

Public Member Functions

Protected Member Functions

Private Attributes

Static Private Attributes

Detailed Description

Member Typedef Documentation

◆ line_searcher_type

◆ objective_function_type

◆ value_type

◆ variable_type

Constructor & Destructor Documentation

◆ descent_method_base()

Member Function Documentation

◆ calc_direction()

◆ evaluations()

◆ gradient()

◆ gradient_norm()

◆ init()

◆ is_stop_criteria_satisfied()

◆ iterate()

◆ iterations()

◆ line_searcher() [1/2]

◆ line_searcher() [2/2]

◆ opt_value()

◆ opt_variable()

◆ tol_gradient_norm()

Member Data Documentation

◆ default_tol_grad_norm

◆ iterations_

◆ line_searcher_

◆ tol_grad_norm_