optk::optimiser Class Reference

#include <optimiser.hpp>

Inheritance diagram for optk::optimiser:

Public Member Functions
	optimiser (std::string name)
	~optimiser ()
virtual void	clear ()
std::string	get_name ()
virtual void	update_search_space (sspace::sspace_t *space)
virtual inst::set	generate_parameters (int param_id)
virtual void	receive_trial_results (int param_id, inst::set params, double value)
void	accept (optimisers *o)

Protected Member Functions
void	add_to_trials (int param_id, inst::node *n)

Protected Attributes
std::unordered_map< int, inst::set >	trials

Private Attributes
std::string	m_name
uint	stepidx
double	best_val
inst::set *	best_params

Detailed Description

This is the base class that every optimisation algorithm will inherit.

Constructor & Destructor Documentation

optimiser()

optk::optimiser::optimiser

(

std::string

name

)

The constructor

Parameters

name	The identifiying name of this optimisation algorithm.

~optimiser()

optk::optimiser::~optimiser

(

)

The destructor will free any generated parameter instances not freed by explicit calls to receive_trial_results.

Member Function Documentation

accept()

void optk::optimiser::accept

(

optimisers *

o

)

Uses the visitor pattern to register an optimiser with the optimisers class.

Parameters

o	A pointer to the optimisers class with which to register this optimiser.

Todo:

is this method necessary?

add_to_trials()

void optk::optimiser::add_to_trials ( int  param_id, inst::node *  n  )

protected

This will add a parameter configuration allocated on the heap to the optimiser::trials map, so that it may be freed later by the optimiser's destructor. If a trial configuration already exists with the same param_id, those values will be freed, and the value in the map will be replaced.

Parameters

param_id	The identifier for this trial.
n	The root node of the concrete parameter settings.

clear()

void optk::optimiser::clear ( )

virtual

Clears / resets this optimiser. This should be the same as calling the destructor and then the constructor.

Reimplemented in gridsearch.

generate_parameters()

virtual inst::set optk::optimiser::generate_parameters ( int  param_id )

inlinevirtual

This is the method invoked to generate the next trial's parameters for the benchmark or program being optimised.

Parameters

param_id

An identifier for this parameter trial; this may or may not be used by the optimistaion algorithm.

Returns

A pointer to an inst::node, or a NULL pointer if there are no more parameters to return.

Reimplemented in gridsearch, gp_opt, and random_search.

get_name()

std::string optk::optimiser::get_name ( )

inline

Returns

the optimiser's name

receive_trial_results()

virtual void optk::optimiser::receive_trial_results ( int  param_id, inst::set  params, double  value  )

inlinevirtual

Allows the benchmark or program being optimised to communicate back the value of the cost function for a parameter set.

Parameters

param_id	This parameter id is used to identify the parameters used to generate the result.
parameters	For increased flexibility, the actual parameters used to generate the result are also copied back.
value	The actual value of the objective function.

Reimplemented in gridsearch, gp_opt, and random_search.

update_search_space()

virtual void optk::optimiser::update_search_space ( sspace::sspace_t *  space )

inlinevirtual

This updates the search space defined by the optimiser.

Parameters

space

The new search space to use

Reimplemented in gridsearch, gp_opt, and random_search.

Member Data Documentation

best_params

inst::set* optk::optimiser::best_params

private

The parameters which gave rise to the best_val

best_val

double optk::optimiser::best_val

private

The best found value of the function

m_name

std::string optk::optimiser::m_name

private

The optimisation algorithm's name

stepidx

uint optk::optimiser::stepidx

private

stepidx is used to generate sequential parameter IDs during step

trials

std::unordered_map<int, inst::set> optk::optimiser::trials

protected

This is used to keep track of previously generated instances or settings of the parameters, identified by the parameter id passed to the generate_parameters method. Contains elements stored on the heap.

---

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

• includes/optk/optimiser.hpp
• src/optimiser.cpp