slp::Filter< Scalar > Class Template Reference

#include </home/runner/work/Sleipnir/Sleipnir/include/sleipnir/optimization/solver/util/filter.hpp>

Public Member Functions
constexpr	Filter (Scalar initial_constraint_violation=Scalar(0))
void	reset ()
	Resets the filter.
bool	try_reset (const FilterEntry< Scalar > &current_entry)
bool	try_add (const FilterEntry< Scalar > &current_entry, const FilterEntry< Scalar > &trial_entry, Scalar D_ϕ, Scalar α)
void	add (const FilterEntry< Scalar > &entry)
bool	is_acceptable (const FilterEntry< Scalar > &entry) const
Scalar	min_step_size (const FilterEntry< Scalar > &current_entry, Scalar D_ϕ) const

Public Attributes
Scalar	min_constraint_violation
	The minimum constraint violation.
Scalar	max_constraint_violation
	The maximum constraint violation.

Detailed Description

template<typename Scalar>
class slp::Filter< Scalar >

Step filter.

See the section on filters in chapter 15 of [1].

Template Parameters

Scalar

Scalar type.

Constructor & Destructor Documentation

Filter()

template<typename Scalar >

constexpr slp::Filter< Scalar >::Filter ( Scalar  initial_constraint_violation = Scalar(0) )

inlineexplicitconstexpr

Constructs an empty filter.

Parameters

initial_constraint_violation

The optimization problem's initial constraint violation.

Member Function Documentation

add()

template<typename Scalar >

void slp::Filter< Scalar >::add ( const FilterEntry< Scalar > &  entry )

inline

Adds the given entry to the filter.

The entry is shifted by the sufficient decrease margins, so iterates that don't sufficiently decrease the cost or constraint violation relative to it are rejected afterward.

Parameters

entry

The entry to add to the filter.

is_acceptable()

template<typename Scalar >

bool slp::Filter< Scalar >::is_acceptable ( const FilterEntry< Scalar > &  entry ) const

inline

Returns true if the given entry is acceptable to the filter.

Parameters

entry

The entry.

Returns

True if the given entry is acceptable to the filter.

min_step_size()

template<typename Scalar >

Scalar slp::Filter< Scalar >::min_step_size ( const FilterEntry< Scalar > &  current_entry, Scalar  D_ϕ  ) const

inline

Returns the minimum step size of the backtracking line search.

Below this step size, no trial iterate can satisfy the sufficient decrease condition or both the switching condition and Armijo condition, so feasibility restoration should be invoked.

Parameters

current_entry	The entry corresponding to the current iterate.
D_ϕ	Cost directional derivative along the search direction.

Returns

The minimum step size.

try_add()

template<typename Scalar >

bool slp::Filter< Scalar >::try_add ( const FilterEntry< Scalar > &  current_entry, const FilterEntry< Scalar > &  trial_entry, Scalar  D_ϕ, Scalar  α  )

inline

Returns true if the given trial entry is acceptable to the filter.

Parameters

current_entry	The entry corresponding to the current iterate.
trial_entry	The entry corresponding to the trial iterate.
D_ϕ	Cost directional derivative along the search direction.
α	The step size (0, 1].

Returns

True if the given entry is acceptable to the filter.

try_reset()

template<typename Scalar >

bool slp::Filter< Scalar >::try_reset ( const FilterEntry< Scalar > &  current_entry )

inline

Resets the filter and lowers the maximum constraint violation if the filter may be impeding progress.

Parameters

current_entry

The entry corresponding to the current iterate.

Returns

True if the filter was reset.

---

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

• /home/runner/work/Sleipnir/Sleipnir/include/sleipnir/optimization/solver/util/filter.hpp