Ipopt::Vector Class Reference

Vector Base Class. More...

#include <IpVector.hpp>

Inheritance diagram for Ipopt::Vector:

[legend]

Collaboration diagram for Ipopt::Vector:

[legend]

List of all members.

Public Member Functions
Vector *	MakeNew () const
	Create new Vector of the same type with uninitialized data.
Vector *	MakeNewCopy () const
	Create new Vector of the same type and copy the data over.
bool	HasValidNumbers () const
	Method for determining if all stored numbers are valid (i.e., no Inf or Nan).
Constructor/Destructor

	Vector (const VectorSpace *owner_space)
	Constructor.
virtual	~Vector ()
	Destructor.
Standard BLAS-1 Operations
(derived classes do NOT overload these methods, instead, overload the protected versions of these methods).
void	Copy (const Vector &x)
	Copy the data of the vector x into this vector (DCOPY).
void	Scal (Number alpha)
	Scales the vector by scalar alpha (DSCAL).
void	Axpy (Number alpha, const Vector &x)
	Add the multiple alpha of vector x to this vector (DAXPY).
Number	Dot (const Vector &x) const
	Computes inner product of vector x with this (DDOT).
Number	Nrm2 () const
	Computes the 2-norm of this vector (DNRM2).
Number	Asum () const
	Computes the 1-norm of this vector (DASUM).
Number	Amax () const
	Computes the max-norm of this vector (based on IDAMAX).
Additional (Non-BLAS) Vector Methods
(derived classes do NOT overload these methods, instead, overload the protected versions of these methods).
void	Set (Number alpha)
	Set each element in the vector to the scalar alpha.
void	ElementWiseDivide (const Vector &x)
	Element-wise division $y_i \gets y_i/x_i$ .
void	ElementWiseMultiply (const Vector &x)
	Element-wise multiplication $y_i \gets y_i*x_i$ .
void	ElementWiseMax (const Vector &x)
	Element-wise max against entries in x.
void	ElementWiseMin (const Vector &x)
	Element-wise min against entries in x.
void	ElementWiseReciprocal ()
	Reciprocates the entries in the vector.
void	ElementWiseAbs ()
	Absolute values of the entries in the vector.
void	ElementWiseSqrt ()
	Element-wise square root of the entries in the vector.
void	ElementWiseSgn ()
	Replaces the vector values with their sgn values ( -1 if x_i < 0, 0 if x_i == 0, and 1 if x_i > 0).
void	AddScalar (Number scalar)
	Add scalar to every vector component.
Number	Max () const
	Returns the maximum value in the vector.
Number	Min () const
	Returns the minimum value in the vector.
Number	Sum () const
	Returns the sum of the vector entries.
Number	SumLogs () const
	Returns the sum of the logs of each vector entry.
Methods for specialized operations. A prototype
implementation is provided, but for efficient implementation those should be specially implemented.
void	AddOneVector (Number a, const Vector &v1, Number c)
	Add one vector, y = a * v1 + c * y.
void	AddTwoVectors (Number a, const Vector &v1, Number b, const Vector &v2, Number c)
	Add two vectors, y = a * v1 + b * v2 + c * y.
Number	FracToBound (const Vector &delta, Number tau) const
	Fraction to the boundary parameter.
void	AddVectorQuotient (Number a, const Vector &z, const Vector &s, Number c)
	Add the quotient of two vectors, y = a * z/s + c * y.
Accessor methods

Index	Dim () const
	Dimension of the Vector.
SmartPtr< const VectorSpace >	OwnerSpace () const
	Return the owner VectorSpace.
Output methods
(derived classes do NOT overload these methods, instead, overload the protected versions of these methods).
void	Print (SmartPtr< const Journalist > jnlst, EJournalLevel level, EJournalCategory category, const std::string &name, Index indent=0, const std::string &prefix="") const
	Print the entire vector.
void	Print (const Journalist &jnlst, EJournalLevel level, EJournalCategory category, const std::string &name, Index indent=0, const std::string &prefix="") const
	Print the entire vector.
Protected Member Functions
implementation methods (derived classes MUST
overload these pure virtual protected methods. )
virtual void	CopyImpl (const Vector &x)=0
	Copy the data of the vector x into this vector (DCOPY).
virtual void	ScalImpl (Number alpha)=0
	Scales the vector by scalar alpha (DSCAL).
virtual void	AxpyImpl (Number alpha, const Vector &x)=0
	Add the multiple alpha of vector x to this vector (DAXPY).
virtual Number	DotImpl (const Vector &x) const =0
	Computes inner product of vector x with this (DDOT).
virtual Number	Nrm2Impl () const =0
	Computes the 2-norm of this vector (DNRM2).
virtual Number	AsumImpl () const =0
	Computes the 1-norm of this vector (DASUM).
virtual Number	AmaxImpl () const =0
	Computes the max-norm of this vector (based on IDAMAX).
virtual void	SetImpl (Number alpha)=0
	Set each element in the vector to the scalar alpha.
virtual void	ElementWiseDivideImpl (const Vector &x)=0
	Element-wise division $y_i \gets y_i/x_i$ .
virtual void	ElementWiseMultiplyImpl (const Vector &x)=0
	Element-wise multiplication $y_i \gets y_i*x_i$ .
virtual void	ElementWiseMaxImpl (const Vector &x)=0
	Element-wise max against entries in x.
virtual void	ElementWiseMinImpl (const Vector &x)=0
	Element-wise min against entries in x.
virtual void	ElementWiseReciprocalImpl ()=0
	Reciprocates the elements of the vector.
virtual void	ElementWiseAbsImpl ()=0
	Take elementwise absolute values of the elements of the vector.
virtual void	ElementWiseSqrtImpl ()=0
	Take elementwise square-root of the elements of the vector.
virtual void	ElementWiseSgnImpl ()=0
	Replaces entries with sgn of the entry.
virtual void	AddScalarImpl (Number scalar)=0
	Add scalar to every component of vector.
virtual Number	MaxImpl () const =0
	Max value in the vector.
virtual Number	MinImpl () const =0
	Min number in the vector.
virtual Number	SumImpl () const =0
	Sum of entries in the vector.
virtual Number	SumLogsImpl () const =0
	Sum of logs of entries in the vector.
virtual void	AddTwoVectorsImpl (Number a, const Vector &v1, Number b, const Vector &v2, Number c)
	Add two vectors (a * v1 + b * v2).
virtual Number	FracToBoundImpl (const Vector &delta, Number tau) const
	Fraction to boundary parameter.
virtual void	AddVectorQuotientImpl (Number a, const Vector &z, const Vector &s, Number c)
	Add the quotient of two vectors.
virtual bool	HasValidNumbersImpl () const
	Method for determining if all stored numbers are valid (i.e., no Inf or Nan).
virtual void	PrintImpl (const Journalist &jnlst, EJournalLevel level, EJournalCategory category, const std::string &name, Index indent, const std::string &prefix) const =0
	Print the entire vector.
Private Member Functions
Default Compiler Generated Methods
(Hidden to avoid implicit creation/calling). These methods are not implemented and we do not want the compiler to implement them for us, so we declare them private and do not define them. This ensures that they will not be implicitly created/called.
	Vector ()
	Default constructor.
	Vector (const Vector &)
	Copy constructor.
Vector &	operator= (const Vector &)
	Overloaded Equals Operator.
Private Attributes
const SmartPtr< const VectorSpace >	owner_space_
	Vector Space.
CachedResults data members

CachedResults< Number >	dot_cache_
	Cache for dot products.
TaggedObject::Tag	nrm2_cache_tag_
	Cache for dot products.
Number	cached_nrm2_
	Cache for dot products.
TaggedObject::Tag	asum_cache_tag_
	Cache for dot products.
Number	cached_asum_
	Cache for dot products.
TaggedObject::Tag	amax_cache_tag_
	Cache for dot products.
Number	cached_amax_
	Cache for dot products.
TaggedObject::Tag	max_cache_tag_
	Cache for dot products.
Number	cached_max_
	Cache for dot products.
TaggedObject::Tag	min_cache_tag_
	Cache for dot products.
Number	cached_min_
	Cache for dot products.
TaggedObject::Tag	sum_cache_tag_
	Cache for dot products.
Number	cached_sum_
	Cache for dot products.
TaggedObject::Tag	sumlogs_cache_tag_
	Cache for dot products.
Number	cached_sumlogs_
	Cache for dot products.
TaggedObject::Tag	valid_cache_tag_
	Cache for dot products.
bool	cached_valid_
	Cache for dot products.

Detailed Description

Vector Base Class.

This is the base class for all derived vector types. Those vectors are meant to store entities like iterates, Lagrangian multipliers, constraint values etc. The implementation of a vector type depends on the computational environment (e.g. just a double array on a shared memory machine, or distributed double arrays for a distributed memory machine.)

Deriving from Vector: This class inherits from tagged object to implement an advanced caching scheme. Because of this, the TaggedObject method ObjectChanged() must be called each time the Vector changes. If you overload the XXXX_Impl protected methods, this taken care of (along with caching if possible) for you. If you have additional methods in your derived class that change the underlying data (vector values), you MUST remember to call ObjectChanged() AFTER making the change!

Definition at line 42 of file IpVector.hpp.

Constructor & Destructor Documentation

Ipopt::Vector::Vector ( const VectorSpace * owner_space ) [inline]

Constructor.

It has to be given a pointer to the corresponding VectorSpace.

Definition at line 410 of file IpVector.hpp.

Ipopt::Vector::~Vector ( ) [inline, virtual]

Destructor.

Definition at line 406 of file IpVector.hpp.

Ipopt::Vector::Vector ( ) [private]

Default constructor.

Ipopt::Vector::Vector ( const Vector & ) [private]

Copy constructor.

Member Function Documentation

Vector * Ipopt::Vector::MakeNew ( ) const [inline]

Create new Vector of the same type with uninitialized data.

Definition at line 428 of file IpVector.hpp.

Vector * Ipopt::Vector::MakeNewCopy ( ) const [inline]

Create new Vector of the same type and copy the data over.

Definition at line 434 of file IpVector.hpp.

void Ipopt::Vector::Copy ( const Vector & x ) [inline]

Copy the data of the vector x into this vector (DCOPY).

Definition at line 443 of file IpVector.hpp.

void Ipopt::Vector::Scal ( Number alpha )

Scales the vector by scalar alpha (DSCAL).

void Ipopt::Vector::Axpy	(	Number	alpha,
		const Vector &	x
	)			`[inline]`

Add the multiple alpha of vector x to this vector (DAXPY).

Definition at line 481 of file IpVector.hpp.

Number Ipopt::Vector::Dot ( const Vector & x ) const [inline]

Computes inner product of vector x with this (DDOT).

Definition at line 488 of file IpVector.hpp.

Number Ipopt::Vector::Nrm2 ( ) const [inline]

Computes the 2-norm of this vector (DNRM2).

Definition at line 507 of file IpVector.hpp.

Number Ipopt::Vector::Asum ( ) const [inline]

Computes the 1-norm of this vector (DASUM).

Definition at line 517 of file IpVector.hpp.

Number Ipopt::Vector::Amax ( ) const [inline]

Computes the max-norm of this vector (based on IDAMAX).

Definition at line 527 of file IpVector.hpp.

void Ipopt::Vector::Set ( Number alpha ) [inline]

Set each element in the vector to the scalar alpha.

Definition at line 564 of file IpVector.hpp.

void Ipopt::Vector::ElementWiseDivide ( const Vector & x ) [inline]

Element-wise division $y_i \gets y_i/x_i$ .

Definition at line 572 of file IpVector.hpp.

void Ipopt::Vector::ElementWiseMultiply ( const Vector & x ) [inline]

Element-wise multiplication $y_i \gets y_i*x_i$ .

Definition at line 579 of file IpVector.hpp.

void Ipopt::Vector::ElementWiseMax ( const Vector & x ) [inline]

Element-wise max against entries in x.

Definition at line 593 of file IpVector.hpp.

void Ipopt::Vector::ElementWiseMin ( const Vector & x ) [inline]

Element-wise min against entries in x.

Definition at line 601 of file IpVector.hpp.

void Ipopt::Vector::ElementWiseReciprocal ( ) [inline]

Reciprocates the entries in the vector.

Definition at line 586 of file IpVector.hpp.

void Ipopt::Vector::ElementWiseAbs ( ) [inline]

Absolute values of the entries in the vector.

Definition at line 609 of file IpVector.hpp.

void Ipopt::Vector::ElementWiseSqrt ( ) [inline]

Element-wise square root of the entries in the vector.

Definition at line 617 of file IpVector.hpp.

void Ipopt::Vector::ElementWiseSgn ( ) [inline]

Replaces the vector values with their sgn values ( -1 if x_i < 0, 0 if x_i == 0, and 1 if x_i > 0).

Definition at line 557 of file IpVector.hpp.

void Ipopt::Vector::AddScalar ( Number scalar ) [inline]

Add scalar to every vector component.

Definition at line 624 of file IpVector.hpp.

Number Ipopt::Vector::Max ( ) const [inline]

Returns the maximum value in the vector.

Definition at line 632 of file IpVector.hpp.

Number Ipopt::Vector::Min ( ) const [inline]

Returns the minimum value in the vector.

Definition at line 642 of file IpVector.hpp.

Number Ipopt::Vector::Sum ( ) const [inline]

Returns the sum of the vector entries.

Definition at line 537 of file IpVector.hpp.

Number Ipopt::Vector::SumLogs ( ) const [inline]

Returns the sum of the logs of each vector entry.

Definition at line 547 of file IpVector.hpp.

void Ipopt::Vector::AddOneVector	(	Number	a,
		const Vector &	v1,
		Number	c
	)			`[inline]`

Add one vector, y = a * v1 + c * y.

This is automatically reduced to call AddTwoVectors.

Definition at line 652 of file IpVector.hpp.

void Ipopt::Vector::AddTwoVectors	(	Number	a,
		const Vector &	v1,
		Number	b,
		const Vector &	v2,
		Number	c
	)			`[inline]`

Add two vectors, y = a * v1 + b * v2 + c * y.

Here, this vector is y

Definition at line 658 of file IpVector.hpp.

Number Ipopt::Vector::FracToBound	(	const Vector &	delta,
		Number	tau
	)			const `[inline]`

Fraction to the boundary parameter.

Computes $\alpha = \max\{\bar\alpha\in(0,1] : x + \bar\alpha \Delta \geq (1-\tau)x\}$

Definition at line 666 of file IpVector.hpp.

void Ipopt::Vector::AddVectorQuotient	(	Number	a,
		const Vector &	z,
		const Vector &	s,
		Number	c
	)			`[inline]`

Add the quotient of two vectors, y = a * z/s + c * y.

Definition at line 686 of file IpVector.hpp.

bool Ipopt::Vector::HasValidNumbers ( ) const [inline]

Method for determining if all stored numbers are valid (i.e., no Inf or Nan).

Definition at line 694 of file IpVector.hpp.

Index Ipopt::Vector::Dim ( ) const [inline]

Dimension of the Vector.

Definition at line 704 of file IpVector.hpp.

SmartPtr< const VectorSpace > Ipopt::Vector::OwnerSpace ( ) const [inline]

Return the owner VectorSpace.

Definition at line 710 of file IpVector.hpp.

void Ipopt::Vector::Print	(	SmartPtr< const Journalist >	jnlst,
		EJournalLevel	level,
		EJournalCategory	category,
		const std::string &	name,
		Index	indent = `0`,
		const std::string &	prefix = `""`
	)			const

Print the entire vector.

void Ipopt::Vector::Print	(	const Journalist &	jnlst,
		EJournalLevel	level,
		EJournalCategory	category,
		const std::string &	name,
		Index	indent = `0`,
		const std::string &	prefix = `""`
	)			const

Print the entire vector.

virtual void Ipopt::Vector::CopyImpl ( const Vector & x ) [protected, pure virtual]

Copy the data of the vector x into this vector (DCOPY).

virtual void Ipopt::Vector::ScalImpl ( Number alpha ) [protected, pure virtual]

Scales the vector by scalar alpha (DSCAL).

Implemented in Ipopt::CompoundVector, and Ipopt::DenseVector.

virtual void Ipopt::Vector::AxpyImpl	(	Number	alpha,
		const Vector &	x
	)			`[protected, pure virtual]`

Add the multiple alpha of vector x to this vector (DAXPY).

virtual Number Ipopt::Vector::DotImpl ( const Vector & x ) const [protected, pure virtual]

Computes inner product of vector x with this (DDOT).

virtual Number Ipopt::Vector::Nrm2Impl ( ) const [protected, pure virtual]

Computes the 2-norm of this vector (DNRM2).

Implemented in Ipopt::CompoundVector, and Ipopt::DenseVector.

virtual Number Ipopt::Vector::AsumImpl ( ) const [protected, pure virtual]

Computes the 1-norm of this vector (DASUM).

Implemented in Ipopt::CompoundVector, and Ipopt::DenseVector.

virtual Number Ipopt::Vector::AmaxImpl ( ) const [protected, pure virtual]

Computes the max-norm of this vector (based on IDAMAX).

Implemented in Ipopt::CompoundVector, and Ipopt::DenseVector.

virtual void Ipopt::Vector::SetImpl ( Number alpha ) [protected, pure virtual]

Set each element in the vector to the scalar alpha.

Implemented in Ipopt::CompoundVector, and Ipopt::DenseVector.

virtual void Ipopt::Vector::ElementWiseDivideImpl ( const Vector & x ) [protected, pure virtual]

Element-wise division $y_i \gets y_i/x_i$ .

virtual void Ipopt::Vector::ElementWiseMultiplyImpl ( const Vector & x ) [protected, pure virtual]

Element-wise multiplication $y_i \gets y_i*x_i$ .

virtual void Ipopt::Vector::ElementWiseMaxImpl ( const Vector & x ) [protected, pure virtual]

Element-wise max against entries in x.

virtual void Ipopt::Vector::ElementWiseMinImpl ( const Vector & x ) [protected, pure virtual]

Element-wise min against entries in x.

virtual void Ipopt::Vector::ElementWiseReciprocalImpl ( ) [protected, pure virtual]

Reciprocates the elements of the vector.

Implemented in Ipopt::CompoundVector, and Ipopt::DenseVector.

virtual void Ipopt::Vector::ElementWiseAbsImpl ( ) [protected, pure virtual]

Take elementwise absolute values of the elements of the vector.

Implemented in Ipopt::CompoundVector, and Ipopt::DenseVector.

virtual void Ipopt::Vector::ElementWiseSqrtImpl ( ) [protected, pure virtual]

Take elementwise square-root of the elements of the vector.

Implemented in Ipopt::CompoundVector, and Ipopt::DenseVector.

virtual void Ipopt::Vector::ElementWiseSgnImpl ( ) [protected, pure virtual]

Replaces entries with sgn of the entry.

Implemented in Ipopt::CompoundVector, and Ipopt::DenseVector.

virtual void Ipopt::Vector::AddScalarImpl ( Number scalar ) [protected, pure virtual]

Add scalar to every component of vector.

Implemented in Ipopt::CompoundVector, and Ipopt::DenseVector.

virtual Number Ipopt::Vector::MaxImpl ( ) const [protected, pure virtual]

Max value in the vector.

Implemented in Ipopt::CompoundVector, and Ipopt::DenseVector.

virtual Number Ipopt::Vector::MinImpl ( ) const [protected, pure virtual]

Min number in the vector.

Implemented in Ipopt::CompoundVector, and Ipopt::DenseVector.

virtual Number Ipopt::Vector::SumImpl ( ) const [protected, pure virtual]

Sum of entries in the vector.

Implemented in Ipopt::CompoundVector, and Ipopt::DenseVector.

virtual Number Ipopt::Vector::SumLogsImpl ( ) const [protected, pure virtual]

Sum of logs of entries in the vector.

Implemented in Ipopt::CompoundVector, and Ipopt::DenseVector.

virtual void Ipopt::Vector::AddTwoVectorsImpl	(	Number	a,
		const Vector &	v1,
		Number	b,
		const Vector &	v2,
		Number	c
	)			`[protected, virtual]`

Add two vectors (a * v1 + b * v2).

Result is stored in this vector.

virtual Number Ipopt::Vector::FracToBoundImpl	(	const Vector &	delta,
		Number	tau
	)			const `[protected, virtual]`

Fraction to boundary parameter.

virtual void Ipopt::Vector::AddVectorQuotientImpl	(	Number	a,
		const Vector &	z,
		const Vector &	s,
		Number	c
	)			`[protected, virtual]`

Add the quotient of two vectors.

virtual bool Ipopt::Vector::HasValidNumbersImpl ( ) const [protected, virtual]

Method for determining if all stored numbers are valid (i.e., no Inf or Nan).

A default implementation using Asum is provided.

Reimplemented in Ipopt::CompoundVector.

virtual void Ipopt::Vector::PrintImpl	(	const Journalist &	jnlst,
		EJournalLevel	level,
		EJournalCategory	category,
		const std::string &	name,
		Index	indent,
		const std::string &	prefix
	)			const `[protected, pure virtual]`