Class which represents a Builder for the EIG. Creates a EIG step by step from all the different components. More...

#include <EIGBuilder.h>

Public Member Functions
	operator unique_ptr< EIG< IP, IS, OP, S >> ()
	Builds a EIG object with all the components specified and returns a unique_ptr. More...

EIGBuilder< IP, IS, OP, S > &	toSolve (unique_ptr< IP > problem)
	Defines the AbstractDomain to solve. More...

EIGBuilder< IP, IS, OP, S > &	weights (const float &fW, const float &mW)
	Sets the weights that will be used in the Weighted Fitness Function. Notice that fW + nW = 1.0. More...

EIGBuilder< IP, IS, OP, S > &	with ()
	This method only returns *this and it is used to make the building process more natural. Completely optional. More...

EIGBuilder< IP, IS, OP, S > &	portfolio (vector< EA > &configs)
	Defines the portfolio of solvers to evaluate the instances during the evolutionary process. The first algorithm in the vector will be considered as the target for the instances. More...

EIGBuilder< IP, IS, OP, S > &	withSearch (NSType nsType, unique_ptr< Distance< float >> dist, const float &thres, const float &finalThresh, const int k, bool warmUp=false)
	Novelty search configuration for the EIG. Here we define the Novelty Approach (Features, Performance), the distance Metric, the threshold to include a solution in the NS archive and the 'k' value to use in the KNN method during the NS execution. More...

EIGBuilder< IP, IS, OP, S > &	evalWith (unique_ptr< InstanceFitness > evalMethod)
	Evaluation approach for generating instances. The instances must be considered easy to solve for the target or hard. More...

EIGBuilder< IP, IS, OP, S > &	repeating (const int &reps)
	Defines how many repetitions the algorithms in the portfolio must perform over each instance in each generation of the EIG. More...

EIGBuilder< IP, IS, OP, S > &	crossover (CXType cxType)
	Defines the crossover operator type to use with this Genetic Algorithm. Variants are available at CXType. More...

EIGBuilder< IP, IS, OP, S > &	withCrossRate (const float &crSsRate)
	Defines the crossover rate to apply the crossover operator. More...

EIGBuilder< IP, IS, OP, S > &	mutation (MutType mutation)
	Defines the mutation operator type to use with this Genetic Algorithm. Variants are available at MutType. More...

EIGBuilder< IP, IS, OP, S > &	withMutRate (const float &mutRate)
	Defines the mutation rate to apply in the mutation operator. More...

EIGBuilder< IP, IS, OP, S > &	selection (SelType selType)
	Defines the selection operator to use. Variants are available at SelType. More...

EIGBuilder< IP, IS, OP, S > &	replacement (ReplacementTypes reType)

EIGBuilder< IP, IS, OP, S > &	populationOf (const int &popsize)
	Defines the population of instances to use. This is the number of instances that the EIG will be evolving at each generation. More...

EIGBuilder< IP, IS, OP, S > &	runDuring (const int &generations)
	Defines the number of generations to perform by EIG. Notice that the EIG works by generations instead of evaluations as others algorithms in dignea do. More...

EIGBuilder< IP, IS, OP, S > &	setFeatures (const Features &f)
	Defines the Feature information for MapElites. More...

Static Public Member Functions
static EIGBuilder< IP, IS, OP, S >	create (GeneratorType type)
	Creates a EIGBuilder object. More...

Detailed Description

template<typename IP, typename IS, typename OP, typename S>
class EIGBuilder< IP, IS, OP, S >

Class which represents a Builder for the EIG. Creates a EIG step by step from all the different components.

Template Parameters

IP
IS
OP
S

Member Function Documentation

◆ create()

template<typename IP , typename IS , typename OP , typename S >

EIGBuilder< IP, IS, OP, S > EIGBuilder< IP, IS, OP, S >::create ( GeneratorType t )

static

Creates a EIGBuilder object.

Template Parameters

IP
IS
OP
S

Returns: EIGBuilder<IP, IS, OP, S>

◆ crossover()

template<typename IP , typename IS , typename OP , typename S >

EIGBuilder< IP, IS, OP, S > & EIGBuilder< IP, IS, OP, S >::crossover ( CXType cxType )

Defines the crossover operator type to use with this Genetic Algorithm. Variants are available at CXType.

Template Parameters

IP
IS
OP
S

Parameters

cxType

Returns: EIGBuilder<IP, IS, OP, S>&

◆ evalWith()

template<typename IP , typename IS , typename OP , typename S >

EIGBuilder< IP, IS, OP, S > & EIGBuilder< IP, IS, OP, S >::evalWith ( unique_ptr< InstanceFitness > evalMethod )

Evaluation approach for generating instances. The instances must be considered easy to solve for the target or hard.

Template Parameters

IP
IS
OP
S

Parameters

evalMethod

Returns: EIGBuilder<IP, IS, OP, S>&

◆ mutation()

template<typename IP , typename IS , typename OP , typename S >

EIGBuilder< IP, IS, OP, S > & EIGBuilder< IP, IS, OP, S >::mutation ( MutType mutation )

Defines the mutation operator type to use with this Genetic Algorithm. Variants are available at MutType.

Template Parameters

IP
IS
OP
S

Parameters

mutation

Returns: EIGBuilder<IP, IS, OP, S>&

◆ operator unique_ptr< EIG< IP, IS, OP, S >>()

template<typename IP , typename IS , typename OP , typename S >

EIGBuilder< IP, IS, OP, S >::operator unique_ptr< EIG< IP, IS, OP, S >>

Builds a EIG object with all the components specified and returns a unique_ptr.

Template Parameters

IP
IS
OP
S

Returns: unique_ptr<EIG<IP, IS, OP, S>>

◆ populationOf()

template<typename IP , typename IS , typename OP , typename S >

EIGBuilder< IP, IS, OP, S > & EIGBuilder< IP, IS, OP, S >::populationOf ( const int & popsize )

Defines the population of instances to use. This is the number of instances that the EIG will be evolving at each generation.

Template Parameters

IP
IS
OP
S

Parameters

popsize

Returns: EIGBuilder<IP, IS, OP, S>&

◆ portfolio()

template<typename IP , typename IS , typename OP , typename S >

EIGBuilder< IP, IS, OP, S > & EIGBuilder< IP, IS, OP, S >::portfolio ( vector< EA > & configs )

Defines the portfolio of solvers to evaluate the instances during the evolutionary process. The first algorithm in the vector will be considered as the target for the instances.

Template Parameters

IP
IS
OP
S

Parameters

configs

Returns: EIGBuilder<IP, IS, OP, S>&

◆ repeating()

template<typename IP , typename IS , typename OP , typename S >

EIGBuilder< IP, IS, OP, S > & EIGBuilder< IP, IS, OP, S >::repeating ( const int & reps )

Defines how many repetitions the algorithms in the portfolio must perform over each instance in each generation of the EIG.

Template Parameters

IP
IS
OP
S

Parameters

reps

Returns: EIGBuilder<IP, IS, OP, S>&

◆ runDuring()

template<typename IP , typename IS , typename OP , typename S >

EIGBuilder< IP, IS, OP, S > & EIGBuilder< IP, IS, OP, S >::runDuring ( const int & generations )

Defines the number of generations to perform by EIG. Notice that the EIG works by generations instead of evaluations as others algorithms in dignea do.

Template Parameters

IP
IS
OP
S

Parameters

generations

Returns: EIGBuilder<IP, IS, OP, S>&

◆ selection()

template<typename IP , typename IS , typename OP , typename S >

EIGBuilder< IP, IS, OP, S > & EIGBuilder< IP, IS, OP, S >::selection ( SelType selType )

Defines the selection operator to use. Variants are available at SelType.

Template Parameters

IP
IS
OP
S

Parameters

selType

Returns: EIGBuilder<IP, IS, OP, S>&

◆ setFeatures()

template<typename IP , typename IS , typename OP , typename S >

EIGBuilder< IP, IS, OP, S > & EIGBuilder< IP, IS, OP, S >::setFeatures ( const Features & f )

Defines the Feature information for MapElites.

Template Parameters

IP
IS
OP
S

Parameters

generations

Returns: EIGBuilder<IP, IS, OP, S>&

◆ toSolve()

template<typename IP , typename IS , typename OP , typename S >

EIGBuilder< IP, IS, OP, S > & EIGBuilder< IP, IS, OP, S >::toSolve ( unique_ptr< IP > problem )

Defines the AbstractDomain to solve.

Template Parameters

IP
IS
OP
S

Parameters

problem

Returns: EIGBuilder<IP, IS, OP, S>&

◆ weights()

template<typename IP , typename IS , typename OP , typename S >

EIGBuilder< IP, IS, OP, S > & EIGBuilder< IP, IS, OP, S >::weights	(	const float &	fW,
		const float &	nW
	)

Sets the weights that will be used in the Weighted Fitness Function. Notice that fW + nW = 1.0.

Template Parameters

IP
IS
OP
S

Parameters

fW
nW

Returns: EIGBuilder<IP, IS, OP, S>&

◆ with()

template<typename IP , typename IS , typename OP , typename S >

EIGBuilder<IP, IS, OP, S>& EIGBuilder< IP, IS, OP, S >::with ( )

inline

This method only returns *this and it is used to make the building process more natural. Completely optional.

Returns: EIGBuilder<IP, IS, OP, S>&

◆ withCrossRate()

template<typename IP , typename IS , typename OP , typename S >

EIGBuilder< IP, IS, OP, S > & EIGBuilder< IP, IS, OP, S >::withCrossRate ( const float & crSsRate )

Defines the crossover rate to apply the crossover operator.

Template Parameters

IP
IS
OP
S

Parameters

crSsRate

Returns: EIGBuilder<IP, IS, OP, S>&

◆ withMutRate()

template<typename IP , typename IS , typename OP , typename S >

EIGBuilder< IP, IS, OP, S > & EIGBuilder< IP, IS, OP, S >::withMutRate ( const float & mutRate )

Defines the mutation rate to apply in the mutation operator.

Template Parameters

IP
IS
OP
S

Parameters

mutRate

Returns: EIGBuilder<IP, IS, OP, S>&

◆ withSearch()

template<typename IP , typename IS , typename OP , typename S >

EIGBuilder< IP, IS, OP, S > & EIGBuilder< IP, IS, OP, S >::withSearch	(	NSType	nsType,
		unique_ptr< Distance< float >>	dist,
		const float &	thres,
		const float &	finalThresh,
		const int	k,
		bool	warmUp = `false`
	)

Novelty search configuration for the EIG. Here we define the Novelty Approach (Features, Performance), the distance Metric, the threshold to include a solution in the NS archive and the 'k' value to use in the KNN method during the NS execution.

Template Parameters

IP
IS
OP
S

Parameters

nsType
dist
thres
k

Returns: EIGBuilder<IP, IS, OP, S>&

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

include/dignea/builders/EIGBuilder.h

Public Member Functions

Static Public Member Functions

Detailed Description

template<typename IP, typename IS, typename OP, typename S> class EIGBuilder< IP, IS, OP, S >

Member Function Documentation

◆ create()

◆ crossover()

◆ evalWith()

◆ mutation()

◆ operator unique_ptr< EIG< IP, IS, OP, S >>()

◆ populationOf()

◆ portfolio()

◆ repeating()

◆ runDuring()

◆ selection()

◆ setFeatures()

◆ toSolve()

◆ weights()

◆ with()

◆ withCrossRate()

◆ withMutRate()

◆ withSearch()

template<typename IP, typename IS, typename OP, typename S>
class EIGBuilder< IP, IS, OP, S >