Bringmann-Friedrich approximation method. More...

#include <bf_approx.h>

Inheritance diagram for pagmo::util::hv_algorithm::bf_approx:

Public Member Functions
	bf_approx (const bool use_exact=true, const unsigned int trivial_subcase_size=1, const double eps=1e-2, const double delta=1e-6, const double delta_multiplier=0.775, const double m_alpha=0.2, const double initial_delta_coeff=0.1, const double gamma=0.25)
	Constructor. More...

double	compute (std::vector< fitness_vector > &, const fitness_vector &) const
	Compute hypervolume. More...

unsigned int	least_contributor (std::vector< fitness_vector > &, const fitness_vector &) const
	Least contributor method. More...

unsigned int	greatest_contributor (std::vector< fitness_vector > &, const fitness_vector &) const
	Greatest contributor method. More...

void	verify_before_compute (const std::vector< fitness_vector > &, const fitness_vector &) const
	Verify before compute method. More...

base_ptr	clone () const
	Clone method.

std::string	get_name () const
	Algorithm name.

Public Member Functions inherited from pagmo::util::hv_algorithm::base
virtual double	exclusive (const unsigned int, std::vector< fitness_vector > &, const fitness_vector &) const
	Exclusive hypervolume method. More...

virtual std::vector< double >	contributions (std::vector< fitness_vector > &, const fitness_vector &) const
	Contributions method. More...

virtual	~base ()
	Destructor required for pure virtual methods.

Friends
class	boost::serialization::access

Additional Inherited Members
Static Public Member Functions inherited from pagmo::util::hv_algorithm::base
static double	volume_between (const fitness_vector &, const fitness_vector &, unsigned int=0)
	Compute volume between two points. More...

Protected Types inherited from pagmo::util::hv_algorithm::base
enum	{ DOM_CMP_B_DOMINATES_A = 1, DOM_CMP_A_DOMINATES_B = 2, DOM_CMP_A_B_EQUAL = 3, DOM_CMP_INCOMPARABLE = 4 }

Protected Member Functions inherited from pagmo::util::hv_algorithm::base
void	assert_minimisation (const std::vector< fitness_vector > &, const fitness_vector &) const
	Assert that reference point dominates every other point from the set. More...

virtual unsigned int	extreme_contributor (std::vector< fitness_vector > &, const fitness_vector &, bool(*)(double, double)) const
	compute the extreme contributor More...

Static Protected Member Functions inherited from pagmo::util::hv_algorithm::base
static bool	cmp_least (const double, const double)
	Comparison method for the least contributor. More...

static bool	cmp_greatest (const double, const double)
	Comparison method for the least contributor. More...

static double	volume_between (double , double , unsigned int)
	Compute volume between two points. More...

static int	dom_cmp (double , double , unsigned int)
	Dominance comparison method. More...

static int	dom_cmp (const fitness_vector &, const fitness_vector &, unsigned int=0)
	Dominance comparison method. More...

Detailed Description

Bringmann-Friedrich approximation method.

This is the class containing the implementation of the Bringmann-Friedrich approximation method for the computation of the least contributor to the hypervolume. Default values for the parameters of the algorithm were obtained from the shark implementation of the algorithm (http://image.diku.dk/shark/doxygen_pages/html/_least_contributor_approximator_8hpp_source.html)

See also: "Approximating the least hypervolume contributor: NP-hard in general, but fast in practice", Karl Bringmann, Tobias Friedrich.

Author: Krzysztof Nowak (kn@ki.nosp@m.ryx..nosp@m.net)

Definition at line 50 of file bf_approx.h.

Constructor & Destructor Documentation

pagmo::util::hv_algorithm::bf_approx::bf_approx	(	const bool	use_exact = `true`,
		const unsigned int	trivial_subcase_size = `1`,
		const double	eps = `1e-2`,
		const double	delta = `1e-6`,
		const double	delta_multiplier = `0.775`,
		const double	alpha = `0.2`,
		const double	initial_delta_coeff = `0.1`,
		const double	gamma = `0.25`
	)

Constructor.

Constructs an instance of the algorithm

Parameters

[in]	use_exact	boolean flag stating whether algorithm is allowed to use exact algorithms for the computation
[in]	trivial_subcase_size	size of the sub-front (points overlapping the bounding box) for which algorithm skips to the exact computation right away
[in]	eps	accuracy of the approximation
[in]	delta	confidence of the approximation
[in]	gamma	constant used for computation of delta for each of the points during the sampling
[in]	delta_multiplier	factor with which delta diminishes each round
[in]	initial_delta_coeff	initial coefficient multiplied by the delta at round 0
[in]	alpha	coefficicient stating how accurately current lowest contributor should be sampled

Definition at line 43 of file bf_approx.cpp.

Member Function Documentation

double pagmo::util::hv_algorithm::bf_approx::compute	(	std::vector< fitness_vector > &	points,
		const fitness_vector &	r_point
	)		const

virtual

Compute hypervolume.

This method is overloaded to throw an exception in case a hypervolume indicator computation is requested.

Parameters

[in]	points	vector of points containing the 3-dimensional points for which we compute the hypervolume
[in]	r_point	reference point for the points

Returns: hypervolume.

Implements pagmo::util::hv_algorithm::base.

Definition at line 426 of file bf_approx.cpp.

unsigned int pagmo::util::hv_algorithm::bf_approx::greatest_contributor	(	std::vector< fitness_vector > &	points,
		const fitness_vector &	r_point
	)		const

virtual

Greatest contributor method.

This method establishes the individual that contributes the most to the hypervolume (approximately withing given epsilon and delta).

Parameters

[in]	points	vector of fitness_vectors for which the hypervolume is computed
[in]	r_point	distinguished "reference point".

Returns: index of the greatest contributing point

Reimplemented from pagmo::util::hv_algorithm::base.

Definition at line 86 of file bf_approx.cpp.

unsigned int pagmo::util::hv_algorithm::bf_approx::least_contributor	(	std::vector< fitness_vector > &	points,
		const fitness_vector &	r_point
	)		const

virtual

Least contributor method.

This method establishes the individual that contributes the least to the hypervolume (approximately withing given epsilon and delta).

Parameters

[in]	points	vector of fitness_vectors for which the hypervolume is computed
[in]	r_point	distinguished "reference point".

Returns: index of the least contributing point

Reimplemented from pagmo::util::hv_algorithm::base.

Definition at line 72 of file bf_approx.cpp.

void pagmo::util::hv_algorithm::bf_approx::verify_before_compute	(	const std::vector< fitness_vector > &	points,
		const fitness_vector &	r_point
	)		const

virtual

Verify before compute method.

Verifies whether given algorithm suits the requested data.

Parameters

[in]	points	vector of points containing the d dimensional points for which we compute the hypervolume
[in]	r_point	reference point for the vector of points

Exceptions

value_error when trying to compute the hypervolume for the non-maximal reference point

Implements pagmo::util::hv_algorithm::base.

Definition at line 412 of file bf_approx.cpp.

Friends And Related Function Documentation

friend class boost::serialization::access

friend

End of 'least_contributor' method variables section

Definition at line 145 of file bf_approx.h.

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

Public Member Functions

Friends

Additional Inherited Members

Detailed Description

Constructor & Destructor Documentation

Member Function Documentation

Friends And Related Function Documentation