Base class for discretizing a DifferentialEquation for use in optimal control algorithms. More...

#include <dynamic_discretization.hpp>

Inheritance diagram for DynamicDiscretization:

Public Member Functions
virtual returnValue	addStage (const DynamicSystem &dynamicSystem_, const Grid &stageIntervals, const IntegratorType &integratorType_=INT_UNKNOWN)=0
virtual returnValue	addTransition (const Transition &transition_)=0
virtual returnValue	clear ()=0
virtual DynamicDiscretization *	clone () const =0
virtual returnValue	deleteAllSeeds ()
	DynamicDiscretization ()
	DynamicDiscretization (UserInteraction *_userInteraction)
	DynamicDiscretization (const DynamicDiscretization &rhs)
virtual returnValue	evaluate (OCPiterate &iter)=0
virtual returnValue	evaluateSensitivities ()=0
virtual returnValue	evaluateSensitivities (const BlockMatrix &seed, BlockMatrix &hessian)=0
virtual returnValue	evaluateSensitivitiesLifted ()=0
virtual returnValue	getBackwardSensitivities (BlockMatrix &D) const
virtual returnValue	getForwardSensitivities (BlockMatrix &D) const
int	getNumberOfIntervals () const
virtual returnValue	getResiduum (BlockMatrix &residuum_) const
virtual BooleanType	isAffine () const =0
DynamicDiscretization &	operator= (const DynamicDiscretization &rhs)
virtual returnValue	setBackwardSeed (const BlockMatrix &seed)
virtual returnValue	setForwardSeed (const BlockMatrix &xSeed_, const BlockMatrix &pSeed_, const BlockMatrix &uSeed_, const BlockMatrix &wSeed_)
virtual returnValue	setUnitBackwardSeed ()
virtual returnValue	setUnitForwardSeed ()
virtual returnValue	unfreeze ()=0
virtual	~DynamicDiscretization ()
Protected Member Functions
void	copy (const DynamicDiscretization &rhs)
uint	getNumEvaluationPoints () const
void	initializeVariables ()
virtual returnValue	setupLogging ()
virtual returnValue	setupOptions ()
Protected Attributes
BlockMatrix	bSeed
BlockMatrix	dBackward
BlockMatrix	dForward
BooleanType	freezeTraj
int	N
int	na
int	np
int	nu
int	nw
int	nx
PrintLevel	printLevel
BlockMatrix	pSeed
VariablesGrid	residuum
Grid	unionGrid
BlockMatrix	uSeed
BlockMatrix	wSeed
BlockMatrix	xSeed

Detailed Description

Base class for discretizing a DifferentialEquation for use in optimal control algorithms.

The class DynamicDiscretization serves as a base class for discretizing a DifferentialEquation for use in optimal control algorithms.

Author:: Boris Houska, Hans Joachim Ferreau

Definition at line 62 of file dynamic_discretization.hpp.

Constructor & Destructor Documentation

BEGIN_NAMESPACE_ACADO DynamicDiscretization::DynamicDiscretization ( )

Default constructor.

Definition at line 48 of file dynamic_discretization.cpp.

DynamicDiscretization::DynamicDiscretization ( UserInteraction * _userInteraction )

Definition at line 59 of file dynamic_discretization.cpp.

DynamicDiscretization::DynamicDiscretization ( const DynamicDiscretization & rhs )

Copy constructor (deep copy).

Definition at line 72 of file dynamic_discretization.cpp.

DynamicDiscretization::~DynamicDiscretization ( ) [virtual]

Destructor.

Definition at line 79 of file dynamic_discretization.cpp.

Member Function Documentation

virtual returnValue DynamicDiscretization::addStage	(	const DynamicSystem &	dynamicSystem_,
		const Grid &	stageIntervals,
		const IntegratorType &	integratorType_ = `INT_UNKNOWN`
	)		`[pure virtual]`

Set the Differential Equations stage by stage.

Implemented in CollocationMethod, and ShootingMethod.

virtual returnValue DynamicDiscretization::addTransition ( const Transition & transition_ ) [pure virtual]

Set the Transition stages.

Implemented in CollocationMethod, and ShootingMethod.

virtual returnValue DynamicDiscretization::clear ( ) [pure virtual]

Deletes all stages and transitions and resets the DynamicDiscretization.

Implemented in CollocationMethod, and ShootingMethod.

virtual DynamicDiscretization* DynamicDiscretization::clone ( ) const [pure virtual]

Clone constructor (deep copy).

Implemented in CollocationMethod, and ShootingMethod.

void DynamicDiscretization::copy ( const DynamicDiscretization & rhs ) [protected]

Definition at line 246 of file dynamic_discretization.cpp.

returnValue DynamicDiscretization::deleteAllSeeds ( ) [virtual]

Deletes all seeds that have been set with the methods above.
This function will also give the corresponding memory free.

Returns:: SUCCESSFUL_RETURN
RET_NO_SEED_ALLOCATED

Reimplemented in CollocationMethod, and ShootingMethod.

Definition at line 176 of file dynamic_discretization.cpp.

virtual returnValue DynamicDiscretization::evaluate ( OCPiterate & iter ) [pure virtual]

Evaluates the descretized DifferentialEquation at a specified
VariablesGrid. The results are written into the residuum of the
type VariablesGrid. This routine is for a simple evaluation only.
If sensitivities are needed use one of the routines below
instead.

Returns:: SUCCESSFUL_RETURN
RET_INVALID_ARGUMENTS
or a specific error message form an underlying
discretization instance.

Implemented in CollocationMethod, and ShootingMethod.

virtual returnValue DynamicDiscretization::evaluateSensitivities ( ) [pure virtual]

Evaluates the sensitivities.

Returns:: SUCCESSFUL_RETURN
RET_NOT_FROZEN

Implemented in ShootingMethod, and CollocationMethod.

virtual returnValue DynamicDiscretization::evaluateSensitivities	(	const BlockMatrix &	seed,
		BlockMatrix &	hessian
	)		`[pure virtual]`

Evaluates the sensitivities and the hessian.

Returns:: SUCCESSFUL_RETURN
RET_NOT_FROZEN

Implemented in ShootingMethod, and CollocationMethod.

virtual returnValue DynamicDiscretization::evaluateSensitivitiesLifted ( ) [pure virtual]

Evaluates the sensitivities.

Returns:: SUCCESSFUL_RETURN
RET_NOT_FROZEN

Implemented in ShootingMethod, and CollocationMethod.

returnValue DynamicDiscretization::getBackwardSensitivities ( BlockMatrix & D ) const [virtual]

Returns the result for the backward sensitivities in BlockMatrix form.

Returns:: SUCCESSFUL_RETURN
RET_INPUT_OUT_OF_RANGE
Returns the total number of intervals.

Parameters:

D	the result for the forward sensitivi- ties

Definition at line 169 of file dynamic_discretization.cpp.

returnValue DynamicDiscretization::getForwardSensitivities ( BlockMatrix & D ) const [virtual]

Returns the result for the forward sensitivities in BlockMatrix form.

Returns:: SUCCESSFUL_RETURN
RET_INPUT_OUT_OF_RANGE

Parameters:

D	the result for the forward sensitivi- ties

Definition at line 162 of file dynamic_discretization.cpp.

int DynamicDiscretization::getNumberOfIntervals ( ) const [inline]

uint DynamicDiscretization::getNumEvaluationPoints ( ) const [protected]

Definition at line 271 of file dynamic_discretization.cpp.

returnValue DynamicDiscretization::getResiduum ( BlockMatrix & residuum_ ) const [virtual]

Returns the result for the residuum.

Returns:: SUCCESSFUL_RETURN

Parameters:

residuum_ the residuum

Definition at line 144 of file dynamic_discretization.cpp.

void DynamicDiscretization::initializeVariables ( ) [protected]

Definition at line 234 of file dynamic_discretization.cpp.

virtual BooleanType DynamicDiscretization::isAffine ( ) const [pure virtual]

Implemented in ShootingMethod, and CollocationMethod.

DynamicDiscretization & DynamicDiscretization::operator= ( const DynamicDiscretization & rhs )

Assignment operator (deep copy).

Definition at line 82 of file dynamic_discretization.cpp.

returnValue DynamicDiscretization::setBackwardSeed ( const BlockMatrix & seed ) [virtual]

Define a backward seed in form of a block matrix.
Here, the block matrix should have N block columns,
where N is the number of points of the union grid.
The i-th column is associated with the i-th grid
point in the union grid. Note that the directions
can itself be organized in sub-blocks as long as
all dimensions fit together.

Returns:: SUCCESFUL_RETURN
RET_INPUT_OUT_OF_RANGE

Parameters:

seed	the seed matrix

Definition at line 125 of file dynamic_discretization.cpp.

returnValue DynamicDiscretization::setForwardSeed	(	const BlockMatrix &	xSeed_,
		const BlockMatrix &	pSeed_,
		const BlockMatrix &	uSeed_,
		const BlockMatrix &	wSeed_
	)		`[virtual]`

Define a forward seed in form of a block matrix.
Here, the block matrix should have N block rows,
where N is the number of points of the union grid.
The i-th row is associated with the i-th grid
point in the union grid. Note that the direction
can itself be organized in sub-blocks as long as
all dimensions fit together.