Wrapper class for all ALS variants (dmrg etc.) More...

#include <als.h>

Classes
struct	ALSAlgorithmicData

Public Types
enum	Direction { Increasing, Decreasing }

using	LocalSolver = std::function< void(const TensorNetwork &, std::vector< Tensor > &, const TensorNetwork &, const ALSAlgorithmicData &)>
	the algorithm that is used to solve the local problems More...

Public Member Functions
	ALSVariant (uint _sites, size_t _numHalfSweeps, LocalSolver _localSolver, bool _assumeSPD, bool _useResidual=false)
	fully defining constructor. alternatively ALSVariants can be created by copying a predefined variant and modifying it More...

double	operator() (const TTOperator &_A, TTTensor &_x, const TTTensor &_b, value_t _convergenceEpsilon, PerformanceData &_perfData=NoPerfData) const

double	operator() (const TTOperator &_A, TTTensor &_x, const TTTensor &_b, size_t _numHalfSweeps, PerformanceData &_perfData=NoPerfData) const

double	operator() (const TTOperator &_A, TTTensor &_x, const TTTensor &_b, PerformanceData &_perfData=NoPerfData) const

double	operator() (TTTensor &_x, const TTTensor &_b, value_t _convergenceEpsilon, PerformanceData &_perfData=NoPerfData) const

double	operator() (TTTensor &_x, const TTTensor &_b, size_t _numHalfSweeps, PerformanceData &_perfData=NoPerfData) const

double	operator() (TTTensor &_x, const TTTensor &_b, PerformanceData &_perfData=NoPerfData) const

Static Public Member Functions
static void	ASD_solver (const TensorNetwork &_A, std::vector< Tensor > &_x, const TensorNetwork &_b, const ALSAlgorithmicData &_data)

static void	lapack_solver (const TensorNetwork &_A, std::vector< Tensor > &_x, const TensorNetwork &_b, const ALSAlgorithmicData &_data)
	local solver that calls the corresponding lapack routines (LU solver) More...

Public Attributes
bool	assumeSPD
	if true the operator A will be assumed to be symmetric positive definite More...

value_t	convergenceEpsilon
	default value for the change in the energy functional at which the ALS assumes it is converged More...

const size_t	FLAG_FINISHED_FULLSWEEP = 3

const size_t	FLAG_FINISHED_HALFSWEEP = 1

LocalSolver	localSolver

size_t	numHalfSweeps
	maximum number of sweeps to perform. set to 0 for infinite More...

bool	preserveCorePosition
	if true the core will be moved to its original position at the end More...

uint	sites
	the number of sites that are simultaneously optimized More...

bool	useResidualForEndCriterion
	calculates the residual to decide if the ALS converged. recommended if _perfdata is given. implied if assumeSPD = false More...

Protected Member Functions
bool	check_for_end_of_sweep (ALSAlgorithmicData &_data, size_t _numHalfSweeps, value_t _convergenceEpsilon, PerformanceData &_perfData) const

TensorNetwork	construct_local_operator (ALSAlgorithmicData &_data) const

TensorNetwork	construct_local_RHS (ALSAlgorithmicData &_data) const

double	solve (const TTOperator *_Ap, TTTensor &_x, const TTTensor &_b, size_t _numHalfSweeps, value_t _convergenceEpsilon, PerformanceData &_perfData=NoPerfData) const

Detailed Description

Wrapper class for all ALS variants (dmrg etc.)

By creating a new object of this class and modifying the member variables, the behaviour of the solver can be modified. The local solver will be ignored for all calls to the x=b variants (ie. without operator A)

Definition at line 37 of file als.h.

Member Typedef Documentation

◆ LocalSolver

using xerus::ALSVariant::LocalSolver = std::function<void(const TensorNetwork &, std::vector<Tensor> &, const TensorNetwork &, const ALSAlgorithmicData &)>

the algorithm that is used to solve the local problems

Definition at line 123 of file als.h.

Member Enumeration Documentation

◆ Direction

enum xerus::ALSVariant::Direction

Enumerator
Increasing
Decreasing

Definition at line 39 of file als.h.

Constructor & Destructor Documentation

◆ ALSVariant()

xerus::ALSVariant::ALSVariant	(	uint	_sites,
		size_t	_numHalfSweeps,
		LocalSolver	_localSolver,
		bool	_assumeSPD,
		bool	_useResidual = `false`
	)

inline

fully defining constructor. alternatively ALSVariants can be created by copying a predefined variant and modifying it

Definition at line 134 of file als.h.

Member Function Documentation

◆ ASD_solver()

void xerus::ALSVariant::ASD_solver	(	const TensorNetwork &	_A,
		std::vector< Tensor > &	_x,
		const TensorNetwork &	_b,
		const ALSAlgorithmicData &	_data
	)

static

Definition at line 73 of file als.cpp.

◆ check_for_end_of_sweep()

bool xerus::ALSVariant::check_for_end_of_sweep	(	ALSAlgorithmicData &	_data,
		size_t	_numHalfSweeps,
		value_t	_convergenceEpsilon,
		PerformanceData &	_perfData
	)		const

protected

Definition at line 426 of file als.cpp.

◆ construct_local_operator()

TensorNetwork xerus::ALSVariant::construct_local_operator ( ALSVariant::ALSAlgorithmicData & _data ) const

protected

Definition at line 383 of file als.cpp.

◆ construct_local_RHS()

TensorNetwork xerus::ALSVariant::construct_local_RHS ( ALSVariant::ALSAlgorithmicData & _data ) const

protected

Definition at line 404 of file als.cpp.

◆ lapack_solver()

void xerus::ALSVariant::lapack_solver	(	const TensorNetwork &	_A,
		std::vector< Tensor > &	_x,
		const TensorNetwork &	_b,
		const ALSAlgorithmicData &	_data
	)

static

local solver that calls the corresponding lapack routines (LU solver)

Definition at line 43 of file als.cpp.

◆ operator()() [1/6]

double xerus::ALSVariant::operator()	(	const TTOperator &	_A,
		TTTensor &	_x,
		const TTTensor &	_b,
		value_t	_convergenceEpsilon,
		PerformanceData &	_perfData = `NoPerfData`
	)		const

inline

call to solve \( A\cdot x = b\) for \( x \) (in a least-squares sense)

Parameters

	_A	operator to solve for
[in,out]	_x	in: initial guess, out: solution as found by the algorithm
	_b	right-hand side of the equation to be solved
	_convergenceEpsilon	minimum change in residual / energy under which the algorithm terminates
	_perfData	vector of performance data (residuals after every microiteration)

Returns: the residual \(|Ax-b|\) of the final _x

Definition at line 152 of file als.h.

◆ operator()() [2/6]

double xerus::ALSVariant::operator()	(	const TTOperator &	_A,
		TTTensor &	_x,
		const TTTensor &	_b,
		size_t	_numHalfSweeps,
		PerformanceData &	_perfData = `NoPerfData`
	)		const

inline

call to solve \( A\cdot x = b\) for \( x \) (in a least-squares sense)

Parameters

	_A	operator to solve for
[in,out]	_x	in: initial guess, out: solution as found by the algorithm
	_b	right-hand side of the equation to be solved
	_numHalfSweeps	maximum number of half-sweeps to perform
	_perfData	vector of performance data (residuals after every microiteration)

Returns: the residual \(|Ax-b|\) of the final _x

Definition at line 165 of file als.h.

◆ operator()() [3/6]

double xerus::ALSVariant::operator()	(	const TTOperator &	_A,
		TTTensor &	_x,
		const TTTensor &	_b,
		PerformanceData &	_perfData = `NoPerfData`
	)		const

inline

call to solve \( A\cdot x = b\) for \( x \) (in a least-squares sense)

Parameters

	_A	operator to solve for
[in,out]	_x	in: initial guess, out: solution as found by the algorithm
	_b	right-hand side of the equation to be solved
	_perfData	vector of performance data (residuals after every microiteration)

Returns: the residual \(|Ax-b|\) of the final _x

Definition at line 177 of file als.h.

◆ operator()() [4/6]

double xerus::ALSVariant::operator()	(	TTTensor &	_x,
		const TTTensor &	_b,
		value_t	_convergenceEpsilon,
		PerformanceData &	_perfData = `NoPerfData`
	)		const

inline

call to minimze \( \|x - b\|^2 \) for \( x \)

Parameters

[in,out]	_x	in: initial guess, out: solution as found by the algorithm
	_b	right-hand side of the equation to be solved
	_convergenceEpsilon	minimum change in residual / energy under which the algorithm terminates
	_perfData	vector of performance data (residuals after every microiteration)

Returns: the residual \(|x-b|\) of the final _x

Definition at line 189 of file als.h.

◆ operator()() [5/6]

double xerus::ALSVariant::operator()	(	TTTensor &	_x,
		const TTTensor &	_b,
		size_t	_numHalfSweeps,
		PerformanceData &	_perfData = `NoPerfData`
	)		const

inline

call to minimze \( \|x - b\|^2 \) for \( x \)

Parameters

[in,out]	_x	in: initial guess, out: solution as found by the algorithm
	_b	right-hand side of the equation to be solved
	_numHalfSweeps	maximum number of half-sweeps to perform
	_perfData	vector of performance data (residuals after every microiteration)

Returns: the residual \(|x-b|\) of the final _x

Definition at line 201 of file als.h.

◆ operator()() [6/6]

double xerus::ALSVariant::operator()	(	TTTensor &	_x,
		const TTTensor &	_b,
		PerformanceData &	_perfData = `NoPerfData`
	)		const

inline

Definition at line 205 of file als.h.

◆ solve()

double xerus::ALSVariant::solve	(	const TTOperator *	_Ap,
		TTTensor &	_x,
		const TTTensor &	_b,
		size_t	_numHalfSweeps,
		value_t	_convergenceEpsilon,
		PerformanceData &	_perfData = `NoPerfData`
	)		const

protected

Definition at line 483 of file als.cpp.

Member Data Documentation

◆ assumeSPD

bool xerus::ALSVariant::assumeSPD

if true the operator A will be assumed to be symmetric positive definite

Definition at line 118 of file als.h.

◆ convergenceEpsilon

value_t xerus::ALSVariant::convergenceEpsilon

default value for the change in the energy functional at which the ALS assumes it is converged

Definition at line 114 of file als.h.

◆ FLAG_FINISHED_FULLSWEEP

const size_t xerus::ALSVariant::FLAG_FINISHED_FULLSWEEP = 3

Definition at line 110 of file als.h.

◆ FLAG_FINISHED_HALFSWEEP

const size_t xerus::ALSVariant::FLAG_FINISHED_HALFSWEEP = 1

Definition at line 109 of file als.h.

◆ localSolver

LocalSolver xerus::ALSVariant::localSolver

Definition at line 124 of file als.h.

◆ numHalfSweeps

size_t xerus::ALSVariant::numHalfSweeps

maximum number of sweeps to perform. set to 0 for infinite

Definition at line 113 of file als.h.

◆ preserveCorePosition

bool xerus::ALSVariant::preserveCorePosition

if true the core will be moved to its original position at the end

Definition at line 117 of file als.h.

◆ sites

uint xerus::ALSVariant::sites

the number of sites that are simultaneously optimized

Definition at line 112 of file als.h.

◆ useResidualForEndCriterion

bool xerus::ALSVariant::useResidualForEndCriterion

calculates the residual to decide if the ALS converged. recommended if _perfdata is given. implied if assumeSPD = false

Definition at line 116 of file als.h.

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

/home/gitlab-runner/builds/9071116c/0/xerus/xerus/include/xerus/algorithms/als.h
/home/gitlab-runner/builds/9071116c/0/xerus/xerus/src/xerus/algorithms/als.cpp

Classes

Public Types

Public Member Functions

Static Public Member Functions

Public Attributes

Protected Member Functions

Detailed Description

Member Typedef Documentation

◆ LocalSolver

Member Enumeration Documentation

◆ Direction

Constructor & Destructor Documentation

◆ ALSVariant()

Member Function Documentation

◆ ASD_solver()

◆ check_for_end_of_sweep()

◆ construct_local_operator()

◆ construct_local_RHS()

◆ lapack_solver()

◆ operator()() [1/6]

◆ operator()() [2/6]

◆ operator()() [3/6]

◆ operator()() [4/6]

◆ operator()() [5/6]

◆ operator()() [6/6]

◆ solve()

Member Data Documentation

◆ assumeSPD

◆ convergenceEpsilon

◆ FLAG_FINISHED_FULLSWEEP

◆ FLAG_FINISHED_HALFSWEEP

◆ localSolver

◆ numHalfSweeps

◆ preserveCorePosition

◆ sites

◆ useResidualForEndCriterion