Implements the projection operator within the symbolic operators family. More...

#include <projection.hpp>

Inheritance diagram for Projection:

Public Member Functions
virtual returnValue	AD_backward (int dim, VariableType varType, int component, Operator seed, Operator df, int &nNewIS, TreeProjection **newIS)
virtual returnValue	AD_backward (int number, double seed, double *df)
virtual returnValue	AD_backward2 (int number, double seed1, double seed2, double df, double ddf)
virtual Operator *	AD_forward (int dim, VariableType varType, int component, Operator seed, int &nNewIS, TreeProjection *newIS)
virtual returnValue	AD_forward (int number, double seed, double df)
virtual returnValue	AD_forward (int number, double x, double seed, double f, double df)
virtual returnValue	AD_forward2 (int number, double seed1, double seed2, double df, double ddf)
virtual returnValue	AD_symmetric (int dim, VariableType varType, int component, Operator l, Operator S, int dimS, Operator dfS, Operator ldf, Operator H, int &nNewLIS, TreeProjection newLIS, int &nNewSIS, TreeProjection newSIS, int &nNewHIS, TreeProjection **newHIS)
virtual returnValue	clearBuffer ()
virtual Operator *	clone () const
virtual Operator *	differentiate (int index)
virtual returnValue	enumerateVariables (SymbolicIndexList *indexList)
virtual returnValue	evaluate (int number, double x, double result)
virtual returnValue	evaluate (EvaluationBase *x)
virtual CurvatureType	getCurvature ()
virtual int	getGlobalIndex () const
virtual MonotonicityType	getMonotonicity ()
virtual OperatorName	getName ()
virtual double	getScale () const
VariableType	getType () const
virtual int	getVariableIndex () const
virtual BooleanType	isDependingOn (VariableType var) const
virtual BooleanType	isDependingOn (int dim, VariableType varType, int component, BooleanType *implicit_dep)
virtual BooleanType	isLinearIn (int dim, VariableType varType, int component, BooleanType *implicit_dep)
virtual NeutralElement	isOneOrZero () const
virtual BooleanType	isPolynomialIn (int dim, VariableType varType, int component, BooleanType *implicit_dep)
virtual BooleanType	isRationalIn (int dim, VariableType varType, int component, BooleanType *implicit_dep)
virtual BooleanType	isSymbolic () const
virtual BooleanType	isTrivial () const
virtual BooleanType	isVariable (VariableType &varType, int &component) const
virtual returnValue	loadIndices (SymbolicIndexList *indexList)
virtual std::ostream &	print (std::ostream &stream) const
	Projection ()
	Projection (const std::string &name_)
	Projection (VariableType variableType_, int vIndex_, const std::string &name_)
	Projection (const Projection &arg)
virtual returnValue	setCurvature (CurvatureType curvature_)
virtual returnValue	setMonotonicity (MonotonicityType monotonicity_)
virtual returnValue	setScale (const double &scale_)
virtual returnValue	setVariableExportName (const VariableType &_type, const std::vector< std::string > &_name)
virtual Operator *	substitute (int index, const Operator *sub)
virtual	~Projection ()
Protected Member Functions
virtual returnValue	ADbackwardProtected (int dim, VariableType varType, int component, Operator seed, Operator df, int &nNewIS, TreeProjection **newIS)
virtual Operator *	ADforwardProtected (int dim, VariableType varType, int component, Operator seed, int &nNewIS, TreeProjection *newIS)
virtual returnValue	ADsymmetricProtected (int dim, VariableType varType, int component, Operator l, Operator S, int dimS, Operator dfS, Operator ldf, Operator H, int &nNewLIS, TreeProjection newLIS, int &nNewSIS, TreeProjection newSIS, int &nNewHIS, TreeProjection **newHIS)
virtual void	copy (const Projection &arg)
Protected Attributes
CurvatureType	curvature
MonotonicityType	monotonicity
std::string	name
OperatorName	operatorName
double	scale
int	variableIndex
VariableType	variableType
int	vIndex
Friends
class	FunctionEvaluationTree

Detailed Description

Implements the projection operator within the symbolic operators family.

The class Projection implements the scalar-valued projection operator within the symbolic operators family.

Author:: Boris Houska, Hans Joachim Ferreau

Definition at line 55 of file projection.hpp.

Constructor & Destructor Documentation

BEGIN_NAMESPACE_ACADO Projection::Projection ( )

Default constructor

Definition at line 43 of file projection.cpp.

Projection::Projection ( const std::string & name_ )

Default constructor

Definition at line 55 of file projection.cpp.

Projection::Projection	(	VariableType	variableType_,
		int	vIndex_,
		const std::string &	name_
	)

Default constructor

Definition at line 68 of file projection.cpp.

Projection::Projection ( const Projection & arg )

Default constructor

Definition at line 136 of file projection.cpp.

Projection::~Projection ( ) [virtual]

Default destructor.

Definition at line 133 of file projection.cpp.

Member Function Documentation

returnValue Projection::AD_backward	(	int	dim,
		VariableType *	varType,
		int *	component,
		Operator *	seed,
		Operator **	df,
		int &	nNewIS,
		TreeProjection ***	newIS
	)		`[virtual]`

Automatic Differentiation in backward mode on the symbolic
level. This function generates an expression for a
backward derivative

Returns:: SUCCESSFUL_RETURN

Parameters:

dim	number of directions
varType	the variable types
component	and their components
seed	the backward seed
df	the result
nNewIS	the number of new IS
newIS	the new IS-pointer

Implements SmoothOperator.

Definition at line 203 of file projection.cpp.

returnValue Projection::AD_backward	(	int	number,
		double	seed,
		double *	df
	)		`[virtual]`

Automatic Differentiation in backward mode based on
buffered values

Returns:: SUCCESSFUL_RETURN
RET_NAN

Parameters:

number	the buffer position
seed	the seed
df	the derivative of the expression

Implements SmoothOperator.

Definition at line 353 of file projection.cpp.

returnValue Projection::AD_backward2	(	int	number,
		double	seed1,
		double	seed2,
		double *	df,
		double *	ddf
	)		`[virtual]`

Automatic Differentiation in backward mode for 2nd order
derivatives based on buffered values.

Returns:: SUCCESFUL_RETURN
RET_NAN

Parameters:

number	the buffer position
seed1	the seed1
seed2	the seed2
df	the 1st derivative of the expression
ddf	the 2nd derivative of the expression

Implements SmoothOperator.

Definition at line 368 of file projection.cpp.

Operator * Projection::AD_forward	(	int	dim,
		VariableType *	varType,
		int *	component,
		Operator **	seed,
		int &	nNewIS,
		TreeProjection ***	newIS
	)		`[virtual]`

Automatic Differentiation in forward mode on the symbolic
level. This function generates an expression for a
forward derivative

Returns:: SUCCESSFUL_RETURN

Parameters:

dim	dimension of the seed
varType	the variable types
component	and their components
seed	the forward seed
nNewIS	the number of new IS
newIS	the new IS-pointer

Implements SmoothOperator.

Definition at line 191 of file projection.cpp.

returnValue Projection::AD_forward	(	int	number,
		double *	seed,
		double *	df
	)		`[virtual]`

Automatic Differentiation in forward mode.
This function uses the intermediate
results from a buffer

Returns:: SUCCESFUL_RETURN
RET_NAN

Parameters:

number	storage position
seed	the seed
df	the derivative of the expression

Implements SmoothOperator.

Definition at line 346 of file projection.cpp.

returnValue Projection::AD_forward	(	int	number,
		double *	x,
		double *	seed,
		double *	f,
		double *	df
	)		`[virtual]`

Automatic Differentiation in forward mode.
This function stores the intermediate
results in a buffer (needed for 2nd order automatic
differentiation in backward mode)

Returns:: SUCCESSFUL_RETURN
RET_NAN

Parameters:

number	storage position
x	The evaluation point x
seed	the seed
f	the value of the expression at x
df	the derivative of the expression

Implements SmoothOperator.

Definition at line 336 of file projection.cpp.

returnValue Projection::AD_forward2	(	int	number,
		double *	seed1,
		double *	seed2,
		double *	df,
		double *	ddf
	)		`[virtual]`

Automatic Differentiation in forward mode for
2nd derivatives.
This function uses intermediate
results from a buffer.

Returns:: SUCCESFUL_RETURN
RET_NAN

Parameters:

number	the buffer position
seed1	the seed
seed2	the seed for the first derivative
df	the derivative of the expression
ddf	the 2nd derivative of the expression

Implements SmoothOperator.

Definition at line 360 of file projection.cpp.

returnValue Projection::AD_symmetric	(	int	dim,
		VariableType *	varType,
		int *	component,
		Operator *	l,
		Operator **	S,
		int	dimS,
		Operator **	dfS,
		Operator **	ldf,
		Operator **	H,
		int &	nNewLIS,
		TreeProjection ***	newLIS,
		int &	nNewSIS,
		TreeProjection ***	newSIS,
		int &	nNewHIS,
		TreeProjection ***	newHIS
	)		`[virtual]`

Automatic Differentiation in symmetric mode on the symbolic
level. This function generates an expression for a
second order derivative.

Returns:: SUCCESSFUL_RETURN

Parameters:

dim	number of directions
varType	the variable types
component	and their components
l	the backward seed
S	forward seed matrix
dimS	dimension of forward seed
dfS	first order foward result
ldf	first order backward result
H	upper trianglular part of the Hessian
nNewLIS	the number of newLIS
newLIS	the new LIS-pointer
nNewSIS	the number of newSIS
newSIS	the new SIS-pointer
nNewHIS	the number of newHIS
newHIS	the new HIS-pointer

Implements SmoothOperator.

Definition at line 215 of file projection.cpp.

returnValue Projection::ADbackwardProtected	(	int	dim,
		VariableType *	varType,
		int *	component,
		Operator *	seed,
		Operator **	df,
		int &	nNewIS,
		TreeProjection ***	newIS
	)		`[protected, virtual]`

Automatic Differentiation in backward mode on the symbolic
level. This function generates an expression for a
backward derivative

Returns:: SUCCESSFUL_RETURN

Parameters:

dim	number of directions
varType	the variable types
component	and their components
seed	the backward seed
df	the result
nNewIS	the number of new IS
newIS	the new IS-pointer

Reimplemented in TreeProjection.

Definition at line 479 of file projection.cpp.

Operator * Projection::ADforwardProtected	(	int	dim,
		VariableType *	varType,
		int *	component,
		Operator **	seed,
		int &	nNewIS,
		TreeProjection ***	newIS
	)		`[protected, virtual]`

Automatic Differentiation in forward mode on the symbolic
level. This function generates an expression for a
forward derivative

Returns:: SUCCESSFUL_RETURN

Parameters:

dim	dimension of the seed
varType	the variable types
component	and their components
seed	the forward seed
nNewIS	the number of new IS
newIS	the new IS-pointer

Reimplemented in TreeProjection.

Definition at line 457 of file projection.cpp.

returnValue Projection::ADsymmetricProtected	(	int	dim,
		VariableType *	varType,
		int *	component,
		Operator *	l,
		Operator **	S,
		int	dimS,
		Operator **	dfS,
		Operator **	ldf,
		Operator **	H,
		int &	nNewLIS,
		TreeProjection ***	newLIS,
		int &	nNewSIS,
		TreeProjection ***	newSIS,
		int &	nNewHIS,
		TreeProjection ***	newHIS
	)		`[protected, virtual]`

Automatic Differentiation in symmetric mode on the symbolic
level. This function generates an expression for a
second order derivative.

Returns:: SUCCESSFUL_RETURN

Parameters:

dim	number of directions
varType	the variable types
component	and their components
l	the backward seed
S	forward seed matrix
dimS	dimension of forward seed
dfS	first order foward result
ldf	first order backward result
H	upper trianglular part of the Hessian
nNewLIS	the number of newLIS
newLIS	the new LIS-pointer
nNewSIS	the number of newSIS
newSIS	the new SIS-pointer
nNewHIS	the number of newHIS
newHIS	the new HIS-pointer

Reimplemented in TreeProjection.

Definition at line 518 of file projection.cpp.

returnValue Projection::clearBuffer ( ) [virtual]

Clears the buffer and resets the buffer size
to 1.

Returns:: SUCCESSFUL_RETURN

Implements SmoothOperator.

Definition at line 385 of file projection.cpp.

Operator * Projection::clone ( ) const [virtual]

Provides a deep copy of the expression.

Returns:: a clone of the expression.

Implements SmoothOperator.

Reimplemented in TreeProjection.

Definition at line 142 of file projection.cpp.

void Projection::copy ( const Projection & arg ) [protected, virtual]

Copy function.

Reimplemented in TreeProjection.

Definition at line 148 of file projection.cpp.

Operator * Projection::differentiate ( int index ) [virtual]

Returns the derivative of the expression with respect
to the variable var(index).

Returns:: The expression for the derivative.

Parameters:

index diff. index

Implements SmoothOperator.

Definition at line 180 of file projection.cpp.

returnValue Projection::enumerateVariables ( SymbolicIndexList * indexList ) [virtual]

Enumerates all variables based on a common
IndexList.

Returns:: SUCCESFUL_RETURN

Implements SmoothOperator.

Definition at line 418 of file projection.cpp.

returnValue Projection::evaluate	(	int	number,
		double *	x,
		double *	result
	)		`[virtual]`

Evaluates the expression and stores the intermediate
results in a buffer (needed for automatic differentiation
in backward mode)

Returns:: SUCCESSFUL_RETURN
RET_NAN

Parameters:

number	storage position
x	the input variable x
result	the result

Implements SmoothOperator.

Definition at line 166 of file projection.cpp.

returnValue Projection::evaluate ( EvaluationBase * x ) [virtual]

Evaluates the expression (templated version)

Implements SmoothOperator.

Definition at line 173 of file projection.cpp.

CurvatureType Projection::getCurvature ( ) [virtual]

Returns the curvature of the expression

Returns:: CT_CONSTANT
CT_AFFINE
CT_CONVEX
CT_CONCAVE

Implements SmoothOperator.

Definition at line 316 of file projection.cpp.

int Projection::getGlobalIndex ( ) const [virtual]

Returns the global type ID of this variable

Returns:: the requested global type ID.

Reimplemented from SmoothOperator.

Definition at line 432 of file projection.cpp.

MonotonicityType Projection::getMonotonicity ( ) [virtual]

Returns the monotonicity of the expression.

Returns:: MT_NONDECREASING
MT_NONINCREASING
MT_NONMONOTONIC

Implements SmoothOperator.

Definition at line 310 of file projection.cpp.

OperatorName Projection::getName ( ) [virtual]

Asks the expression for its name.

Returns:: the name of the expression.

Implements SmoothOperator.

Definition at line 391 of file projection.cpp.

double Projection::getScale ( ) const [virtual]

Asks the expression for its scale.

Returns:: the scale of the expression.

Definition at line 397 of file projection.cpp.

VariableType Projection::getType ( ) const

Asks the variable for its type.

Definition at line 438 of file projection.cpp.

int Projection::getVariableIndex ( ) const [virtual]

Asks the variable for its relative index.

Definition at line 426 of file projection.cpp.

BooleanType Projection::isDependingOn ( VariableType var ) const [virtual]

Asks the expression whether it is depending on a certian type of
variable.

Returns:: BT_TRUE if a dependency is detected,
BT_FALSE otherwise.

Implements SmoothOperator.

Definition at line 256 of file projection.cpp.

BooleanType Projection::isDependingOn	(	int	dim,
		VariableType *	varType,
		int *	component,
		BooleanType *	implicit_dep
	)		`[virtual]`

Checks whether the expression is depending on a variable

Returns:: BT_FALSE if no dependence is detected
BT_TRUE otherwise

Parameters:

dim	number of directions
varType	the variable types
component	and their components
implicit_dep	implicit dependencies

Implements SmoothOperator.

Reimplemented in TreeProjection.

Definition at line 265 of file projection.cpp.

BooleanType Projection::isLinearIn	(	int	dim,
		VariableType *	varType,
		int *	component,
		BooleanType *	implicit_dep
	)		`[virtual]`

Checks whether the expression is linear in
(or not depending on) a variable

Returns:: BT_FALSE if no linearity is
detected
BT_TRUE otherwise

Parameters:

dim	number of directions
varType	the variable types
component	and their components
implicit_dep	implicit dependencies

Implements SmoothOperator.

Reimplemented in TreeProjection.

Definition at line 283 of file projection.cpp.

NeutralElement Projection::isOneOrZero ( ) const [virtual]

Checks whether the expression is zero or one

Returns:: NE_ZERO
NE_ONE
NE_NEITHER_ONE_NOR_ZERO

Implements SmoothOperator.

Reimplemented in TreeProjection.

Definition at line 248 of file projection.cpp.

BooleanType Projection::isPolynomialIn	(	int	dim,
		VariableType *	varType,
		int *	component,
		BooleanType *	implicit_dep
	)		`[virtual]`

Checks whether the expression is polynomial in
the specified variables

Returns:: BT_FALSE if the expression is not polynomial
BT_TRUE otherwise

Parameters:

dim	number of directions
varType	the variable types
component	and their components
implicit_dep	implicit dependencies

Implements SmoothOperator.

Reimplemented in TreeProjection.

Definition at line 292 of file projection.cpp.

BooleanType Projection::isRationalIn	(	int	dim,
		VariableType *	varType,
		int *	component,
		BooleanType *	implicit_dep
	)		`[virtual]`

Checks whether the expression is rational in
the specified variables

Returns:: BT_FALSE if the expression is not rational
BT_TRUE otherwise

Parameters:

dim	number of directions
varType	the variable types
component	and their components
implicit_dep	implicit dependencies

Implements SmoothOperator.

Reimplemented in TreeProjection.

Definition at line 301 of file projection.cpp.

BooleanType Projection::isSymbolic ( ) const [virtual]

Asks whether all elements are purely symbolic.

Returns:: BT_TRUE if the complete tree is symbolic.
BT_FALSE otherwise (e.g. if C functions are linked).

Implements SmoothOperator.

Definition at line 451 of file projection.cpp.

BooleanType Projection::isTrivial ( ) const [virtual]

Reimplemented from Operator.

Reimplemented in TreeProjection.

Definition at line 579 of file projection.cpp.

BooleanType Projection::isVariable	(	VariableType &	varType,
		int &	component
	)		const `[virtual]`

Asks the expression whether it is a variable.

Returns:: The answer.

Implements SmoothOperator.

Definition at line 410 of file projection.cpp.

returnValue Projection::loadIndices ( SymbolicIndexList * indexList ) [virtual]

The function loadIndices passes an IndexList through
the whole expression tree. Whenever a variable gets the
IndexList it tries to make an entry. However if a
variable recognices that it has already been added
before, it will not be allowed to make a second entry.
Note that all variables, in paticular the intermediate
states, will keep in mind whether they were allowed
to make an entry or not. This guarantees that
intermediate states are never evaluated twice if they
occur at several knots of the tree.

THIS FUNCTION IS FOR INTERNAL USE ONLY.

PLEASE CALL THIS FUNTION AT MOST ONCE FOR AN EXPRESSION
AS A KIND OF INIT ROUTINE.