op::NLopt< T > Class Template Reference

A op::optimizer implementation for NLopt. More...

#include <nlopt_op.hpp>

Collaboration diagram for op::NLopt< T >:

Public Member Functions
	NLopt (op::Vector< std::vector< double >> &variables, NLoptOptions &o, std::optional< MPI_Comm > comm={}, std::optional< op::utility::CommPattern< T >> comm_pattern_info={})
	Constructor for our optimizer.
	num_local_owned_variables_ (0)
void	setObjective (op::Functional &o) override
	Sets the optimization objective. More...
void	addConstraint (op::Functional &o) override
	Adds a constraint for the optimization problem. More...
bool	variables_changed (const std::vector< double > &x)
	Method to see if variables changed, if they have set new x. More...
bool	isAdvanced ()
	returns whether NLopt is in "advanced" mode or not
auto	generateReducedLocalGradientFunction (std::function< std::vector< double >(const std::vector< double > &)> local_grad_func, std::function< double(const std::vector< double > &)> local_reduce_func)
	generates reduced local gradient using comm_pattern_
Public Member Functions inherited from op::Optimizer
	Optimizer ()
	Ctor has deferred initialization.
	iterate ([](){})
	save ([](){})
	final_obj (std::numeric_limits< double >::max())
void	Go ()
	Start the optimization.
void	UpdatedVariableCallback ()
	What to do when the variables are updated.
virtual double	Solution ()
	What to do when the solution is found. Return the objetive.
void	Iteration ()
	What to do at the end of an optimization iteration.
void	SaveState ()
	Saves the state of the optimizer.
virtual	~Optimizer ()=default
	Destructor.

Protected Attributes
MPI_Comm	comm_
std::vector< double >	global_variables_
op::Vector< std::vector < double > > &	variables_
std::unique_ptr< nlopt::opt >	nlopt_
NLoptOptions &	options_
std::vector< double >	previous_variables_
std::vector < detail::FunctionalInfo< T > >	obj_info_
std::vector < detail::FunctionalInfo< T > >	constraints_info_
std::vector< int >	owned_variables_per_rank_
std::vector< int >	owned_offsets_
std::optional < utility::CommPattern< T > >	comm_pattern_
std::optional < std::unordered_map< typename T::value_type, T > >	global_reduced_map_to_local_
std::size_t	num_local_owned_variables_
int	root_rank_ = 0
std::unique_ptr< WaitLoop >	waitloop_

Friends
double	NLoptFunctional (const std::vector< double > &x, std::vector< double > &grad, void *objective_and_optimizer)
	Takes in a op::Functional and computes the objective function and it's gradient as a nlopt function. More...

Additional Inherited Members
Public Attributes inherited from op::Optimizer
op::Go	go
	Go function to start optimization.
CallbackFn	update
	Update callback to compute before function calculations.
CallbackFn	iterate
	iterate callback to compute before
CallbackFn	save
	callback for saving current optimizer state
double	final_obj
	final objective value

Detailed Description

template<typename T>
class op::NLopt< T >

A op::optimizer implementation for NLopt.

Definition at line 15 of file nlopt_op.hpp.

Member Function Documentation

template<typename T >

void op::NLopt< T >::addConstraint ( op::Functional &  )

inlineoverridevirtual

Adds a constraint for the optimization problem.

Parameters

[in]

o

Constraint Functional

Reimplemented from op::Optimizer.

Definition at line 305 of file nlopt_op.hpp.

template<typename T >

void op::NLopt< T >::setObjective ( op::Functional &  o )

inlineoverridevirtual

Sets the optimization objective.

Parameters

[in]

o

Objective Functional

Implements op::Optimizer.

Definition at line 298 of file nlopt_op.hpp.

template<typename T >

bool op::NLopt< T >::variables_changed ( const std::vector< double > &  x )

inline

Method to see if variables changed, if they have set new x.

Parameters

[in]

x

Definition at line 331 of file nlopt_op.hpp.

Friends And Related Function Documentation

template<typename T >

double NLoptFunctional ( const std::vector< double > &  x, std::vector< double > &  grad, void *  objective_and_optimizer  )

friend

Takes in a op::Functional and computes the objective function and it's gradient as a nlopt function.

Has the same signature as nlopt::function so we can convert any op::Functional into a nlopt::function

Parameters

[in]	x	the optimization variables (on rank = 0 this is the actual global optimization variables, on other ranks it is the local-view of variables.data())
[in]	grad	the result of the gradient of the function w.r.t. x (on rank 0, this is the global gradient eval, on other ranks it is the owned-local gradient)
[in]	objective	Get FunctionalInfo into this call

for constraints g >= lower_bound, they need to be rewritten as -(g - lower_bound) <= 0

for constraints g <= upper_bound, they need to be rewritten as g - upper_bound < = 0

---

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

• /usr/workspace/jekel1/Repos/op/src/nlopt_op.hpp