Optimize by trying out a grid of configurations. More...

#include <shark/Algorithms/DirectSearch/GridSearch.h>

Inheritance diagram for shark::GridSearch:

Public Member Functions
	GridSearch ()

std::string	name () const
	From INameable: return the class name. More...

void	configure (size_t params, double min, double max, size_t numSections)

void	configure (const std::vector< double > &min, const std::vector< double > &max, const std::vector< size_t > &sections)

void	configure (double min1, double max1, size_t sections1, double min2, double max2, size_t sections2)

void	configure (double min1, double max1, size_t sections1)

void	configure (size_t params, const std::vector< double > &values)

void	configure (const std::vector< std::vector< double > > &values)

virtual void	read (InArchive &archive)
	Read the component from the supplied archive. More...

virtual void	write (OutArchive &archive) const
	Write the component to the supplied archive. More...

virtual void	init (ObjectiveFunctionType const &objectiveFunction, SearchPointType const &startingPoint)

void	assignLinearRange (size_t index, size_t noOfSections, double min, double max)

void	assignExponentialRange (size_t index, double factor, double exp_base, int min, int max)

void	step (ObjectiveFunctionType const &objectiveFunction)

Public Member Functions inherited from shark::AbstractSingleObjectiveOptimizer< RealVector >
std::size_t	numInitPoints () const
	By default most single objective optimizers only require a single point. More...

virtual void	init (ObjectiveFunctionType const &function, std::vector< SearchPointType > const &initPoints)
	Initialize the optimizer for the supplied objective function using a set of initialisation points. More...

virtual const SolutionType &	solution () const
	returns the current solution of the optimizer More...

Public Member Functions inherited from shark::AbstractOptimizer< RealVector, double, SingleObjectiveResultSet< RealVector > >
const Features &	features () const

virtual void	updateFeatures ()

bool	requiresValue () const

bool	requiresFirstDerivative () const

bool	requiresSecondDerivative () const

bool	canSolveConstrained () const

bool	requiresClosestFeasible () const

virtual	~AbstractOptimizer ()

virtual void	init (ObjectiveFunctionType const &function)
	Initialize the optimizer for the supplied objective function. More...

Public Member Functions inherited from shark::INameable
virtual	~INameable ()

Public Member Functions inherited from shark::ISerializable
virtual	~ISerializable ()
	Virtual d'tor. More...

void	load (InArchive &archive, unsigned int version)
	Versioned loading of components, calls read(...). More...

void	save (OutArchive &archive, unsigned int version) const
	Versioned storing of components, calls write(...). More...

	BOOST_SERIALIZATION_SPLIT_MEMBER ()

Protected Attributes
std::vector< std::vector< double > >	m_nodeValues
	The array columns contain the grid values for the corresponding parameter axis. More...

bool	m_configured

Protected Attributes inherited from shark::AbstractSingleObjectiveOptimizer< RealVector >
SolutionType	m_best
	Current solution of the optimizer. More...

Protected Attributes inherited from shark::AbstractOptimizer< RealVector, double, SingleObjectiveResultSet< RealVector > >
Features	m_features

Additional Inherited Members
Public Types inherited from shark::AbstractSingleObjectiveOptimizer< RealVector >
typedef base_type::SearchPointType	SearchPointType

typedef base_type::SolutionType	SolutionType

typedef base_type::ResultType	ResultType

typedef base_type::ObjectiveFunctionType	ObjectiveFunctionType

Public Types inherited from shark::AbstractOptimizer< RealVector, double, SingleObjectiveResultSet< RealVector > >
enum	Feature
	Models features that the optimizer requires from the objective function. More...

typedef RealVector	SearchPointType

typedef double	ResultType

typedef SingleObjectiveResultSet< RealVector >	SolutionType

typedef AbstractObjectiveFunction< RealVector, ResultType >	ObjectiveFunctionType

typedef TypedFlags< Feature >	Features

typedef TypedFeatureNotAvailableException< Feature >	FeatureNotAvailableException

Protected Member Functions inherited from shark::AbstractOptimizer< RealVector, double, SingleObjectiveResultSet< RealVector > >
void	checkFeatures (ObjectiveFunctionType const &objectiveFunction)
	Convenience function that checks whether the features of the supplied objective function match with the required features of the optimizer. More...

Detailed Description

Optimize by trying out a grid of configurations.

: The GridSearch class allows for the definition of a grid in parameter space. It does a simple one-step optimization over the grid by trying out every possible parameter combination. Please note that the computation effort grows exponentially with the number of parameters.

: If you only want to try a subset of the grid, consider using the PointSearch class instead. A more sophisticated (less exhaustive) grid search variant is available with the NestedGridSearch class.

Definition at line 63 of file GridSearch.h.

Constructor & Destructor Documentation

◆ GridSearch()

shark::GridSearch::GridSearch ( )

inline

Definition at line 66 of file GridSearch.h.

References m_configured.

Member Function Documentation

◆ assignExponentialRange()

void shark::GridSearch::assignExponentialRange	(	size_t	index,
		double	factor,
		double	exp_base,
		int	min,
		int	max
	)

inline

Set exponentially progressing grid values for one certain parameter only. This is especially useful if one parameter needs special treatment. The grid points will be filled with values $factor \cdot expbase ^i$ , where i does integer steps between min and max.

Parameters

index	the index of the parameter that gets new grid values
factor	the value that the exponential base grid should be multiplied by
exp_base	the exponential grid will progress on this base (e.g. 2, 10)
min	the smallest exponent for exp_base
max	the largest exponent for exp_base

Definition at line 263 of file GridSearch.h.

◆ assignLinearRange()

void shark::GridSearch::assignLinearRange	(	size_t	index,
		size_t	noOfSections,
		double	min,
		double	max
	)

inline

Assign linearly progressing grid values to one certain parameter only. This is especially useful if one parameter needs special treatment

Parameters

index	the index of the parameter to which grid values are assigned
noOfSections	how many grid points should be assigned to that parameter
min	smallest value for that parameter
max	largest value for that parameter

Definition at line 237 of file GridSearch.h.

◆ configure() [1/6]

void shark::GridSearch::configure	(	size_t	params,
		double	min,
		double	max,
		size_t	numSections
	)

inline

uniform initialization for all parameters

Parameters

params	number of model parameters to optimize
min	smallest parameter value
max	largest parameter value
numSections	total number of values in the interval

Definition at line 81 of file GridSearch.h.

Referenced by init(), and shark::NestedGridSearch::init().

◆ configure() [2/6]

void shark::GridSearch::configure	(	const std::vector< double > &	min,
		const std::vector< double > &	max,
		const std::vector< size_t > &	sections
	)

inline

individual definition for every parameter

Parameters

min	smallest value for every parameter
max	largest value for every parameter
sections	total number of values for every parameter

Definition at line 100 of file GridSearch.h.

◆ configure() [3/6]

void shark::GridSearch::configure	(	double	min1,
		double	max1,
		size_t	sections1,
		double	min2,
		double	max2,
		size_t	sections2
	)

inline

special case for 2D grid, individual definition for every parameter

Parameters

min1	smallest value for first parameter
max1	largest value for first parameter
sections1	total number of values for first parameter
min2	smallest value for second parameter
max2	largest value for second parameter
sections2	total number of values for second parameter

Definition at line 134 of file GridSearch.h.

References m_nodeValues, and RANGE_CHECK.

◆ configure() [4/6]

void shark::GridSearch::configure	(	double	min1,
		double	max1,
		size_t	sections1
	)

inline

special case for line search

Parameters

min1	smallest value for first parameter
max1	largest value for first parameter
sections1	total number of values for first parameter

Definition at line 162 of file GridSearch.h.

References m_nodeValues, and RANGE_CHECK.

◆ configure() [5/6]

void shark::GridSearch::configure	(	size_t	params,
		const std::vector< double > &	values
	)

inline

uniform definition of the values to test for all parameters

Parameters

params	number of model parameters to optimize
values	values used for every coordinate

Definition at line 181 of file GridSearch.h.

References m_configured, m_nodeValues, and SIZE_CHECK.

◆ configure() [6/6]

void shark::GridSearch::configure ( const std::vector< std::vector< double > > & values )

inline

individual definition for every parameter

Parameters

values values used. The first dimension is the parameter, the second dimension is the node.

Definition at line 193 of file GridSearch.h.

References m_configured, m_nodeValues, and SIZE_CHECK.

◆ init()

virtual void shark::GridSearch::init	(	ObjectiveFunctionType const &	objectiveFunction,
		SearchPointType const &	startingPoint
	)

inlinevirtual

If Gridsearch wasn't configured before calling this method, it is default constructed as a net spanning the range [-1,1] in all dimensions with 5 searchpoints (-1,-0.5,0,0.5,1). so don't forget to scale the parameter-ranges of the objective function! The startingPoint can actually be anything, only its dimension has to be correct.

Implements shark::AbstractSingleObjectiveOptimizer< RealVector >.

Definition at line 222 of file GridSearch.h.

References shark::AbstractOptimizer< RealVector, double, SingleObjectiveResultSet< RealVector > >::checkFeatures(), configure(), shark::AbstractSingleObjectiveOptimizer< RealVector >::m_best, m_configured, m_nodeValues, shark::ResultSet< SearchPointT, ResultT >::point, and SIZE_CHECK.