BayesOpt

Main module for Bayesian optimization. More...

Classes
class	bayesopt::ContinuousModel
	Bayesian optimization for functions in continuous input spaces. More...
class	bayesopt::DiscreteModel
	Bayesian optimization for functions in discrete spaces. More...
class	bayesopt::BayesOptBase
	Abstract module for Bayesian optimization. More...
class	bayesopt::EmpiricalBayes
	Bayesian optimization using different non-parametric processes as distributions over surrogate functions. More...
class	bayesopt::PosteriorFixed
	Bayesian optimization using different non-parametric processes as distributions over surrogate functions. More...
class	bayesopt::PosteriorModel
	Bayesian optimization using different non-parametric processes as distributions over surrogate functions. More...

Typedefs
typedef double(*	eval_func) (unsigned int n, const double *x, double *gradient, void *func_data)

Functions
BAYESOPT_API int	bayes_optimization (int nDim, eval_func f, void *f_data, const double *lb, const double *ub, double *x, double *minf, bopt_params parameters)
	C wrapper for the Bayesian optimization algorithm. More...
BAYESOPT_API int	bayes_optimization_disc (int nDim, eval_func f, void *f_data, double *valid_x, size_t n_points, double *x, double *minf, bopt_params parameters)
	C wrapper for the Bayesian optimization algorithm. More...
BAYESOPT_API int	bayes_optimization_categorical (int nDim, eval_func f, void *f_data, int *categories, double *x, double *minf, bopt_params parameters)
	C wrapper for the Bayesian optimization algorithm. More...
static PosteriorModel *	bayesopt::PosteriorModel::create (size_t dim, Parameters params, randEngine &eng)
	bayesopt::PosteriorModel::PosteriorModel (size_t dim, Parameters params, randEngine &eng)
	Constructor. More...
virtual	bayesopt::PosteriorModel::~PosteriorModel ()
	Default destructor.
virtual void	bayesopt::PosteriorModel::updateHyperParameters ()=0
virtual void	bayesopt::PosteriorModel::fitSurrogateModel ()=0
virtual void	bayesopt::PosteriorModel::updateSurrogateModel ()=0
virtual double	bayesopt::PosteriorModel::evaluateCriteria (const vectord &query)=0
virtual void	bayesopt::PosteriorModel::updateCriteria (const vectord &query)=0
virtual bool	bayesopt::PosteriorModel::criteriaRequiresComparison ()=0
virtual void	bayesopt::PosteriorModel::setFirstCriterium ()=0
virtual bool	bayesopt::PosteriorModel::setNextCriterium (const vectord &prevResult)=0
virtual std::string	bayesopt::PosteriorModel::getBestCriteria (vectord &best)=0
void	bayesopt::PosteriorModel::setSamples (const matrixd &x, const vectord &y)
void	bayesopt::PosteriorModel::setSamples (const matrixd &x)
void	bayesopt::PosteriorModel::setSamples (const vectord &y)
void	bayesopt::PosteriorModel::setSample (const vectord &x, double y)
void	bayesopt::PosteriorModel::addSample (const vectord &x, double y)
double	bayesopt::PosteriorModel::getValueAtMinimum ()
vectord	bayesopt::PosteriorModel::getPointAtMinimum ()
void	bayesopt::PosteriorModel::plotDataset (TLogLevel level)
const Dataset *	bayesopt::PosteriorModel::getData ()
virtual ProbabilityDistribution *	bayesopt::PosteriorModel::getPrediction (const vectord &query)=0

Variables
Parameters	bayesopt::PosteriorModel::mParameters
	Configuration parameters.
size_t	bayesopt::PosteriorModel::mDims
	Number of dimensions.
Dataset	bayesopt::PosteriorModel::mData
	Dataset (x-> inputs, y-> labels/output)
MeanModel	bayesopt::PosteriorModel::mMean

Detailed Description

Main module for Bayesian optimization.

Function Documentation

bayes_optimization()

BAYESOPT_API int bayes_optimization	(	int	nDim,
		eval_func	f,
		void *	f_data,
		const double *	lb,
		const double *	ub,
		double *	x,
		double *	minf,
		bopt_params	parameters
	)

C wrapper for the Bayesian optimization algorithm.

This function assumes continuous optimization.

Parameters

nDim	number of input dimensions
f	pointer to the function to optimize
f_data	pointer to extra data to be used by f
lb	array of lower bounds
ub	array of upper bounds
x	input: initial query, output: result (minimum)
minf	value of the function at the minimum
parameters	parameters for the Bayesian optimization.

Returns

error code

Definition at line 98 of file bayesoptwpr.cpp.

bayes_optimization_categorical()

BAYESOPT_API int bayes_optimization_categorical	(	int	nDim,
		eval_func	f,
		void *	f_data,
		int *	categories,
		double *	x,
		double *	minf,
		bopt_params	parameters
	)

C wrapper for the Bayesian optimization algorithm.

This assumes the input variables are categories. Use of the Hamming kernel is highly recommended.

Parameters

nDim	number of input dimensions
f	pointer to the function to optimize
f_data	pointer to extra data to be used by f
categories	number of categories on each dimension
x	input: initial query, output: result (minimum)
minf	value of the function at the minimum
parameters	parameters for the Bayesian optimization.

Returns

error code

Definition at line 203 of file bayesoptwpr.cpp.

bayes_optimization_disc()

BAYESOPT_API int bayes_optimization_disc	(	int	nDim,
		eval_func	f,
		void *	f_data,
		double *	valid_x,
		size_t	n_points,
		double *	x,
		double *	minf,
		bopt_params	parameters
	)

C wrapper for the Bayesian optimization algorithm.

This function assumes discrete optimization.

Parameters

nDim	number of input dimensions
f	pointer to the function to optimize
f_data	pointer to extra data to be used by f
valid_x	set of possible discrete points
n_points	number of possible discrete points
x	input: initial query, output: result (minimum)
minf	value of the function at the minimum
parameters	parameters for the Bayesian optimization.

Returns

error code

Definition at line 143 of file bayesoptwpr.cpp.

Referenced by ExampleDisc::checkReachability().

PosteriorModel()

bayesopt::PosteriorModel::PosteriorModel	(	size_t	dim,
		Parameters	params,
		randEngine &	eng
	)

Constructor.

Parameters

params

set of parameters (see parameters.hpp)

Definition at line 46 of file posteriormodel.cpp.