NOX Development
Loading...
Searching...
No Matches
Public Member Functions | Protected Types | Protected Member Functions | Protected Attributes | List of all members
NOX::Solver::TensorBased Class Reference

Nonlinear solver based on a rank-1 tensor method. More...

#include <NOX_Solver_TensorBased.H>

Inheritance diagram for NOX::Solver::TensorBased:
Inheritance graph
[legend]
Collaboration diagram for NOX::Solver::TensorBased:
Collaboration graph
[legend]

Public Member Functions

 TensorBased (const Teuchos::RCP< NOX::Abstract::Group > &grp, const Teuchos::RCP< NOX::StatusTest::Generic > &tests, const Teuchos::RCP< Teuchos::ParameterList > &params)
 Constructor.
virtual ~TensorBased ()
 Destructor.
virtual void reset (const NOX::Abstract::Vector &initialGuess, const Teuchos::RCP< NOX::StatusTest::Generic > &tests)
 Resets the solver, sets a new status test, and sets a new initial guess.
virtual void reset (const NOX::Abstract::Vector &initialGuess)
 Resets the solver and sets a new initial guess.
virtual void reset ()
 Resets the solver for another solve. This resets the counters and status only. Uses the final solution from the last solve as the initial guess for the next solve.
virtual NOX::StatusTest::StatusType step ()
 Do one nonlinear step in the iteration sequence and return status.
virtual NOX::StatusTest::StatusType solve ()
 Solve the nonlinear problem and return final status.
virtual const NOX::Abstract::Group & getSolutionGroup () const
 Return a reference to the current solution group.
virtual const NOX::Abstract::Group & getPreviousSolutionGroup () const
 Return a reference to the previous solution group.
virtual NOX::StatusTest::StatusType getStatus () const
 Returns the current status of the solver.
virtual int getNumIterations () const
 Get number of iterations.
virtual const Teuchos::ParameterList & getList () const
 Return a reference to the solver parameters.
virtual Teuchos::RCP< const NOX::Abstract::Group > getSolutionGroupPtr () const
 Return a RCP to the solution group.
virtual Teuchos::RCP< const NOX::Abstract::Group > getPreviousSolutionGroupPtr () const
 Return a RCP to the previous solution group.
virtual Teuchos::RCP< const Teuchos::ParameterList > getListPtr () const
 Return a RCP to the solver parameters.
virtual Teuchos::RCP< const NOX::SolverStats > getSolverStatistics () const
 Return a RCP to the solver statistics.
Public Member Functions inherited from NOX::Solver::Generic
 Generic ()
 Constructor (does nothing).
virtual ~Generic ()
 Destructor (does nothing).

Protected Types

enum  StepType { TensorStep , NewtonStep }
 Types of steps.
enum  LineSearchType {
  Curvilinear , Standard , Dual , FullStep ,
  Newton
}
 Flag for the direction to be computed this iteration. More...
enum  ConvergenceCriteriaType { ArmijoGoldstein , AredPred , None }
 Algorithms used to determine convergence of the line search. More...
enum  LambdaSelectionType { Halving , Quadratic }
 Types of lambda selection.
enum  RecoveryStepType { Constant , LastComputedStep }
 Type of recovery step to use. More...

Protected Member Functions

virtual void init ()
 Print out initialization information.
virtual void printUpdate ()
 Prints the current iteration information.
virtual bool reset (const Teuchos::RCP< NOX::Abstract::Group > &grp, const Teuchos::RCP< NOX::StatusTest::Generic > &tests, const Teuchos::RCP< Teuchos::ParameterList > &params)
 Constructor initialization routine.
bool computeTensorDirection (NOX::Abstract::Group &soln, const NOX::Solver::Generic &solver)
 Subroutine for computing the tensor and Newton directions.
double calculateBeta (double qa, double qb, double qc, double &qval, double &lambdaBar, double lambda=1.0) const
 Subroutine for calculating beta.
bool computeCurvilinearStep (NOX::Abstract::Vector &dir, const NOX::Abstract::Group &soln, const NOX::Solver::Generic &s, double &lambda)
 Subroutine for computing the curvilinear step.
bool implementGlobalStrategy (NOX::Abstract::Group &newGrp, double &step, const NOX::Solver::Generic &s)
 Subroutine for executing the tensor linesearch.
bool performLinesearch (NOX::Abstract::Group &newsoln, double &step, const NOX::Abstract::Vector &lsDir, const NOX::Solver::Generic &s)
 Performs a standard tensor linesearch (tensor or Newton direction).
double getNormModelResidual (const NOX::Abstract::Vector &dir, const NOX::Abstract::Group &soln, bool isTensorModel) const
 Compute the residual norm of the local model.
void printDirectionInfo (std::string dirName, const NOX::Abstract::Vector &dir, const NOX::Abstract::Group &soln, bool isTensorModel) const
 Print pertinent information about the direction.
double getDirectionalDerivative (const NOX::Abstract::Vector &dir, const NOX::Abstract::Group &soln) const
 Calculate the directional derivative.
double selectLambda (double newf, double oldf, double oldfprime, double lambda)
 Select lambda for linesearch (quadratic or halving).
void throwError (const std::string &functionName, const std::string &errorMsg) const
 Throw an error with a method's name and error message.

Protected Attributes

Teuchos::RCP< NOX::GlobalData > globalDataPtr
 Pointer to the global data object.
Teuchos::RCP< NOX::Utils > utilsPtr
 Printing Utils.
Teuchos::RCP< NOX::Abstract::Group > solnPtr
 Current solution.
Teuchos::RCP< NOX::Abstract::Group > oldSolnPtr
 Previous solution pointer.
Teuchos::RCP< NOX::Abstract::Vector > newtonVecPtr
 Current Newton direction (pointer).
Teuchos::RCP< NOX::Abstract::Vector > tensorVecPtr
 Current tensor direction (pointer).
Teuchos::RCP< NOX::Abstract::Vector > aVecPtr
 Current tensor term vector (pointer).
Teuchos::RCP< NOX::Abstract::Vector > sVecPtr
 Vector to previous point (pointer).
Teuchos::RCP< NOX::Abstract::Vector > tmpVecPtr
 Working vector (pointer).
Teuchos::RCP< NOX::Abstract::Vector > residualVecPtr
 Residual vector (pointer).
Teuchos::RCP< NOX::StatusTest::Generic > testPtr
 Stopping test.
NOX::StatusTest::CheckType checkType
 Type of check to use for status tests. See NOX::StatusTest for more details.
Teuchos::RCP< Teuchos::ParameterList > paramsPtr
 Input parameters.
Teuchos::ParameterList * linearParamsPtr
 Line Search parameters.
double stepSize
 Current step.
double beta
 Value of sc'*dt.
int nIter
 Number of nonlinear iterations.
NOX::StatusTest::StatusType status
 Status of nonlinear solver.
StepType requestedBaseStep
 Flag for the base direction to compute after the first iteration.
LineSearchType lsType
 Choice of line search.
ConvergenceCriteriaType convCriteria
 Choice of convergence criteria (currently unused).
LambdaSelectionType lambdaSelection
 Flag for lambda selection (Halving/Quadratic).
RecoveryStepType recoveryStepType
 Choice of the recovery step type; uses "Recovery Step Type" parameter.
bool useModifiedMethod
 Flag for using modifications that force quadratic to have real root.
bool isNewtonDirection
 Flag for Newton direction.
bool doRescue
 Flag for rescuing Linear Solver from a bad solve.
double minStep
 Minimum step length (i.e., when we give up).
double defaultStep
 Default step.
double recoveryStep
 Default step for linesearch failure.
int maxIters
 Maximum iterations.
double alpha
 Scaling factor for the Armijo-Goldstein condition.
double sTinvJF
 Value of s'*inv(J)*F.
double sTinvJa
 Value of s'*inv(J)*a.
Teuchos::RCP< NOX::LineSearch::Utils::Printing > print
 Common line search printing utilities.
NOX::LineSearchCounters * counter
 Common common counters for line searches.
NOX::LineSearch::Utils::Slope slopeObj
 Common slope calculations for line searches.
int numJvMults
 Counter for number of Jacobian-vector products.
int numJ2vMults
 Counter for number of "double" Jacobian-vector products.
Teuchos::RCP< NOX::Observer > observer
 Pointer to a user defined NOX::Observer object.

Detailed Description

Nonlinear solver based on a rank-1 tensor method.

Solves $ F(x)=0$ using a rank-1 tensor method and a linesearch globalization.

At the kth nonlinear iteration, the solver does the following:

  • Computes the tensor direction $ d_T $ by finding the root or smallest magnitude minimizer of the local model

    \‍[ M_T(x_k+d) = F_k + J_kd + a_k(s_k^Td)^2, \‍]

    where

    \‍[ a_k = 2(F_{k-1} - F_k - J_ks_k) / (s_k^Ts_k)^2 \‍]

    and

    \‍[ s_k = s_{k-1} - s_k. \‍]

  • Modifies the step according to a global strategy and updates $ x$ as $ x_{k+1} = x_k + d(\lambda) $ via a linesearch method, where $ d(\lambda) $ is some function of $ \lambda $. For instance, the curvilinear step $ d_{\lambda T} $ is a function of the linesearch parameter $ \lambda $ and is a parametric step that spans the directions of the tensor step and the Newton step. At $ \lambda=1 $, the curvilinear step equals the full tensor step, and as $ \lambda $ nears 0, the curvilinear step approaches the Newton direction. This step provides a monotonic decrease in the norm of the local tensor model as $ \lambda $ varies from 0 to 1.

The solver iterates until the status tests (see NOX::StatusTest) determine either failure or convergence.

Input Parameters

To use this solver, set the "Nonlinear Solver" parameter to be "Tensor Based". Then, specify the following sublists with the appropriate parameters as indicated below.

  • "Direction" - Sublist of the direction parameters, passed to the NOX::Direction::Factory constructor. Defaults to an empty list.

    • "Method" - Name of the direction to be computed in this solver. "Tensor" and "Newton" are the only two valid choices. A sublist by this name specifies all of the parameters to be passed to the linear solver. See below under "Linear Solver".

    • "Rescue Bad Newton Solve" (Boolean) - If the linear solve does not meet the tolerance specified by the forcing term, then use the step anyway. Defaults to true.

    • "Linear Solver" - Sublist for the specific linear solver parameters that are passed to NOX::Abstract::Group::computeNewton() and NOX::Abstract::Group::applyJacobianInverse(). "Linear Solver" is itself a sublist of the list specified in "Method" above (i.e., "Tensor" or "Newton"). Below is a partial list of standard parameters usually available in common linear solvers. Check with the specific linear solver being used for other parameters.

      • "Max Iterations" - Maximum number of Arnoldi iterations (also max Krylov space dimension)

      • "Tolerance" - Relative tolerance for solving local model [default = 1e-4]

      • "Output Frequency" - Print output at every number of iterations [default = 20]

  • "Line Search" - Sublist of the line search parameters. Because the tensor step is not guaranteed to be a descent direction on the function, not all "basic" line search approaches would be appropriate. Thus, the LineSearch classes available to Newton's method (e.g., Polynomial, More-Thuente) are not used here. Instead, this solver class approriately handles technical considerations for tensor methods with its own set of global strategies. The following parameters specify the specific options for this line search:

    • "Method" - Name of the line search available to tensor methods Valid choices are: 

       <ul>

 <li> "Curvilinear" - Backtrack along the "curvilinear"

      path that spans the tensor direction and the Newton direction and that maintains monotonicity on the tensor model. Recommended because it tends to be more robust and efficient than the other choices. [Default]

    • "Standard" - Backtrack along tensor direction unless it is not a descent direction, in which case backtrack along Newton direction.

    • "Dual" - Backtrack along both the Newton and tensor directions and choose the better of the two.

    • "Full Step" - Only use the full step and do not backtrack along both the Newton and tensor directions and choose the better of the two. 

      </ul>

    • "Lambda selection" - Flag for how to calculate the next linesearch parameter lambda. Valid choices are "Quadratic" and "Halving" (default). Quadratic constructs a quadratic interpolating polynomial from the last trial point and uses the minimum of this function as the next trial lambda (bounded by 0.1). Halving divides the linesearch parameter by 2 before each trial, which is simpler but tends to generate longer steps than quadratic.

    • "Default Step" - Starting value of the linesearch parameter (defaults to 1.0)

    • "Minimum Step" - Minimum acceptable linesearch parameter before the linesearch terminates (defaults to 1.0e-12). If there are many linesearch failures, then lowering this value is one thing to try.

    • "Recovery Step Type" - Determines the step size to take when the line search fails. Choices are:

      • "Constant" [default] - Uses a constant value set in "Recovery Step".

      • "Last Computed Step" - Uses the last value computed by the line search algorithm.

    • "Recovery Step" - Step parameter to take when the line search fails (defaults to value for "Default Step")

    • "Max Iters" - Maximum number of iterations (i.e., backtracks)

  • "Solver Options" - Sublist of general solver options.

    • "User Defined Pre/Post Operator" is supported. See NOX::Parameter::PrePostOperator for more details.

Output Parameters

Every time solve() is called, a sublist for output parameters called "Output" will be created and will contain the following parameters:

  • "Nonlinear Iterations" - Number of nonlinear iterations

  • "2-Norm of Residual" - L-2 norm of the final residual $ F(x_k) $.

References

  • B. W. Bader, Tensor-Krylov methods for solving large-scale systems of nonlinear equations, Ph.D. Thesis, 2003, University of Colorado, Boulder, Colorado.

  • B. W. Bader, Tensor-Krylov methods for solving large-scale systems of nonlinear equations, submitted to SIAM J. Numer. Anal.

  • B. W. Bader and R. B. Schnabel, Curvilinear linesearch for tensor methods, SISC, 25(2):604-622.

  • R. B. Schnabel and P. D. Frank, Tensor methods for nonlinear equations, SIAM J. Numer. Anal., 21(5):815-843.

Author
Brett Bader (SNL 9233)

Member Enumeration Documentation

◆ ConvergenceCriteriaType

enum NOX::Solver::TensorBased::ConvergenceCriteriaType
protected

Algorithms used to determine convergence of the line search.

Enumerator
ArmijoGoldstein 

Sufficient decrease condition.

AredPred 

Ared/Pred condition.

None 

Just accept the first step.

◆ LineSearchType

enum NOX::Solver::TensorBased::LineSearchType
protected

Flag for the direction to be computed this iteration.

Enumerated list for each type of line search

◆ RecoveryStepType

enum NOX::Solver::TensorBased::RecoveryStepType
protected

Type of recovery step to use.

Enumerator
Constant 

Use a constant value.

LastComputedStep 

Use the last value computed in the line search algorithm.

Constructor & Destructor Documentation

◆ TensorBased()

NOX::Solver::TensorBased::TensorBased ( const Teuchos::RCP< NOX::Abstract::Group > & grp,
const Teuchos::RCP< NOX::StatusTest::Generic > & tests,
const Teuchos::RCP< Teuchos::ParameterList > & params )

Constructor.

See reset() for description.

References aVecPtr, beta, NOX::DeepCopy, linearParamsPtr, newtonVecPtr, oldSolnPtr, paramsPtr, reset(), residualVecPtr, NOX::ShapeCopy, solnPtr, sTinvJa, sTinvJF, sVecPtr, tensorVecPtr, testPtr, and tmpVecPtr.

◆ ~TensorBased()

NOX::Solver::TensorBased::~TensorBased ( )
virtual

Destructor.

References NOX::Utils::Details, numJ2vMults, numJvMults, and utilsPtr.

Member Function Documentation

◆ calculateBeta()

double NOX::Solver::TensorBased::calculateBeta ( double qa,
double qb,
double qc,
double & qval,
double & lambdaBar,
double lambda = 1.0 ) const
protected

Subroutine for calculating beta.

References NOX::Utils::Details, and utilsPtr.

Referenced by computeCurvilinearStep(), and computeTensorDirection().

◆ computeCurvilinearStep()

bool NOX::Solver::TensorBased::computeCurvilinearStep ( NOX::Abstract::Vector & dir,
const NOX::Abstract::Group & soln,
const NOX::Solver::Generic & s,
double & lambda )
protected

Subroutine for computing the curvilinear step.

References beta, calculateBeta(), NOX::Utils::Details, NOX::Abstract::Vector::innerProduct(), newtonVecPtr, sTinvJa, sTinvJF, tensorVecPtr, NOX::Abstract::Vector::update(), and utilsPtr.

Referenced by performLinesearch().

◆ computeTensorDirection()

bool NOX::Solver::TensorBased::computeTensorDirection ( NOX::Abstract::Group & soln,
const NOX::Solver::Generic & solver )
protected

Subroutine for computing the tensor and Newton directions.

References NOX::Abstract::Group::applyJacobian(), NOX::Abstract::Group::applyJacobianInverse(), aVecPtr, beta, calculateBeta(), NOX::Abstract::Group::computeF(), NOX::Abstract::Group::computeJacobian(), NOX::Abstract::Group::computeNewton(), NOX::Utils::Details, doRescue, NOX::Abstract::Group::getF(), NOX::Abstract::Group::getNewton(), NOX::Abstract::Group::getNormF(), NOX::Solver::Generic::getPreviousSolutionGroup(), NOX::Abstract::Group::getX(), newtonVecPtr, nIter, NOX::Abstract::Vector::norm(), numJ2vMults, numJvMults, NOX::Abstract::Group::Ok, paramsPtr, printDirectionInfo(), requestedBaseStep, residualVecPtr, sTinvJa, sTinvJF, sVecPtr, tensorVecPtr, throwError(), tmpVecPtr, NOX::Abstract::Vector::update(), useModifiedMethod, utilsPtr, and NOX::Utils::Warning.

Referenced by step().

◆ getDirectionalDerivative()

double NOX::Solver::TensorBased::getDirectionalDerivative ( const NOX::Abstract::Vector & dir,
const NOX::Abstract::Group & soln ) const
protected

Calculate the directional derivative.

References NOX::Abstract::Group::applyJacobian(), NOX::Abstract::Vector::clone(), NOX::Abstract::Group::getF(), numJvMults, and NOX::ShapeCopy.

Referenced by printDirectionInfo().

◆ getList()

const Teuchos::ParameterList & NOX::Solver::TensorBased::getList ( ) const
virtual

Return a reference to the solver parameters.

Implements NOX::Solver::Generic.

References paramsPtr.

◆ getListPtr()

virtual Teuchos::RCP< const Teuchos::ParameterList > NOX::Solver::TensorBased::getListPtr ( ) const
inlinevirtual

Return a RCP to the solver parameters.

Implements NOX::Solver::Generic.

References paramsPtr.

◆ getNormModelResidual()

double NOX::Solver::TensorBased::getNormModelResidual ( const NOX::Abstract::Vector & dir,
const NOX::Abstract::Group & soln,
bool isTensorModel ) const
protected

Compute the residual norm of the local model.

References NOX::Abstract::Group::applyJacobian(), aVecPtr, NOX::Abstract::Vector::clone(), NOX::Utils::Details, NOX::Abstract::Group::getF(), numJvMults, NOX::ShapeCopy, sVecPtr, and utilsPtr.

Referenced by printDirectionInfo().

◆ getNumIterations()

int NOX::Solver::TensorBased::getNumIterations ( ) const
virtual

Get number of iterations.

Implements NOX::Solver::Generic.

References nIter.

◆ getPreviousSolutionGroup()

const NOX::Abstract::Group & NOX::Solver::TensorBased::getPreviousSolutionGroup ( ) const
virtual

Return a reference to the previous solution group.

Implements NOX::Solver::Generic.

References oldSolnPtr.

◆ getPreviousSolutionGroupPtr()

virtual Teuchos::RCP< const NOX::Abstract::Group > NOX::Solver::TensorBased::getPreviousSolutionGroupPtr ( ) const
inlinevirtual

Return a RCP to the previous solution group.

Implements NOX::Solver::Generic.

References oldSolnPtr.

◆ getSolutionGroup()

const NOX::Abstract::Group & NOX::Solver::TensorBased::getSolutionGroup ( ) const
virtual

Return a reference to the current solution group.

Implements NOX::Solver::Generic.

References solnPtr.

◆ getSolutionGroupPtr()

virtual Teuchos::RCP< const NOX::Abstract::Group > NOX::Solver::TensorBased::getSolutionGroupPtr ( ) const
inlinevirtual

Return a RCP to the solution group.

Implements NOX::Solver::Generic.

References solnPtr.

◆ getSolverStatistics()

Teuchos::RCP< const NOX::SolverStats > NOX::Solver::TensorBased::getSolverStatistics ( ) const
virtual

Return a RCP to the solver statistics.

Implements NOX::Solver::Generic.

References globalDataPtr.

◆ getStatus()

NOX::StatusTest::StatusType NOX::Solver::TensorBased::getStatus ( ) const
virtual

Returns the current status of the solver.

Implements NOX::Solver::Generic.

References status.

◆ implementGlobalStrategy()

bool NOX::Solver::TensorBased::implementGlobalStrategy ( NOX::Abstract::Group & newGrp,
double & step,
const NOX::Solver::Generic & s )
protected

Subroutine for executing the tensor linesearch.

References NOX::Abstract::Group::computeF(), NOX::Abstract::Group::computeX(), counter, NOX::Abstract::Group::getNormF(), NOX::Solver::Generic::getPreviousSolutionGroup(), isNewtonDirection, lsType, newtonVecPtr, performLinesearch(), slopeObj, and tensorVecPtr.

Referenced by step().

◆ init()

void NOX::Solver::TensorBased::init ( )
protectedvirtual

Print out initialization information.

References counter, NOX::Utils::fill(), nIter, numJ2vMults, numJvMults, NOX::Utils::Parameters, paramsPtr, status, stepSize, NOX::StatusTest::Unconverged, and utilsPtr.

Referenced by reset(), reset(), and reset().

◆ performLinesearch()

bool NOX::Solver::TensorBased::performLinesearch ( NOX::Abstract::Group & newsoln,
double & step,
const NOX::Abstract::Vector & lsDir,
const NOX::Solver::Generic & s )
protected

Performs a standard tensor linesearch (tensor or Newton direction).

References alpha, computeCurvilinearStep(), NOX::Abstract::Group::computeF(), NOX::Abstract::Group::computeX(), Constant, counter, defaultStep, NOX::Utils::Details, NOX::Utils::fill(), NOX::Abstract::Group::getNormF(), NOX::Solver::Generic::getPreviousSolutionGroup(), NOX::Utils::InnerIteration, isNewtonDirection, lsType, maxIters, minStep, newtonVecPtr, numJvMults, paramsPtr, print, recoveryStep, recoveryStepType, selectLambda(), slopeObj, tmpVecPtr, and utilsPtr.

Referenced by implementGlobalStrategy().

◆ printDirectionInfo()

void NOX::Solver::TensorBased::printDirectionInfo ( std::string dirName,
const NOX::Abstract::Vector & dir,
const NOX::Abstract::Group & soln,
bool isTensorModel ) const
protected

Print pertinent information about the direction.

References NOX::Utils::Details, getDirectionalDerivative(), NOX::Abstract::Group::getNormF(), getNormModelResidual(), NOX::Abstract::Vector::norm(), and utilsPtr.

Referenced by computeTensorDirection().

◆ printUpdate()

void NOX::Solver::TensorBased::printUpdate ( )
protectedvirtual

Prints the current iteration information.

References NOX::StatusTest::Converged, NOX::StatusTest::Failed, NOX::Utils::fill(), nIter, NOX::Utils::OuterIteration, NOX::Utils::OuterIterationStatusTest, solnPtr, status, stepSize, tensorVecPtr, testPtr, NOX::StatusTest::Unconverged, and utilsPtr.

Referenced by step().

◆ reset() [1/4]

void NOX::Solver::TensorBased::reset ( )
virtual

Resets the solver for another solve. This resets the counters and status only. Uses the final solution from the last solve as the initial guess for the next solve.

NOTE: All NOX solvers will call reset() automatically at teh beginning of the solve() method. We add the reset() method to the solver interface for the application to call in case the application needs to reset counters and status manually before the next call to solve() is made.

Implements NOX::Solver::Generic.

References counter, nIter, numJ2vMults, numJvMults, status, stepSize, and NOX::StatusTest::Unconverged.

◆ reset() [2/4]

void NOX::Solver::TensorBased::reset ( const NOX::Abstract::Vector & initial_guess)
virtual

Resets the solver and sets a new initial guess.

Implements NOX::Solver::Generic.

References init(), and solnPtr.

◆ reset() [3/4]

void NOX::Solver::TensorBased::reset ( const NOX::Abstract::Vector & initial_guess,
const Teuchos::RCP< NOX::StatusTest::Generic > & test )
virtual

Resets the solver, sets a new status test, and sets a new initial guess.

Implements NOX::Solver::Generic.

References init(), solnPtr, and testPtr.

Referenced by solve(), and TensorBased().

◆ reset() [4/4]

bool NOX::Solver::TensorBased::reset ( const Teuchos::RCP< NOX::Abstract::Group > & grp,
const Teuchos::RCP< NOX::StatusTest::Generic > & tests,
const Teuchos::RCP< Teuchos::ParameterList > & params )
protectedvirtual

Constructor initialization routine.

References alpha, ArmijoGoldstein, checkType, Constant, convCriteria, counter, defaultStep, doRescue, NOX::Utils::Error, globalDataPtr, init(), lambdaSelection, LastComputedStep, lsType, maxIters, minStep, observer, NOX::Utils::Parameters, paramsPtr, NOX::Solver::parseObserver(), NOX::Solver::parseStatusTestCheckType(), print, recoveryStep, recoveryStepType, requestedBaseStep, slopeObj, solnPtr, testPtr, useModifiedMethod, utilsPtr, and NOX::Solver::validateSolverOptionsSublist().

◆ selectLambda()

double NOX::Solver::TensorBased::selectLambda ( double newf,
double oldf,
double oldfprime,
double lambda )
protected

Select lambda for linesearch (quadratic or halving).

References lambdaSelection.

Referenced by performLinesearch().

◆ solve()

NOX::StatusTest::StatusType NOX::Solver::TensorBased::solve ( )
virtual

Solve the nonlinear problem and return final status.

By "solve", we call iterate() until the NOX::StatusTest value is either NOX::StatusTest::Converged or NOX::StatusTest::Failed.

Implements NOX::Solver::Generic.

References nIter, observer, paramsPtr, reset(), solnPtr, status, step(), and NOX::StatusTest::Unconverged.

◆ step()

NOX::StatusTest::StatusType NOX::Solver::TensorBased::step ( )
virtual

Do one nonlinear step in the iteration sequence and return status.

Implements NOX::Solver::Generic.

References NOX::StatusTest::Generic::checkStatus(), checkType, NOX::Abstract::Group::computeF(), computeTensorDirection(), NOX::StatusTest::Converged, NOX::Utils::Error, NOX::StatusTest::Failed, implementGlobalStrategy(), nIter, observer, NOX::Abstract::Group::Ok, oldSolnPtr, printUpdate(), solnPtr, status, stepSize, testPtr, NOX::StatusTest::Unconverged, utilsPtr, and NOX::Utils::Warning.

Referenced by solve().

◆ throwError()

void NOX::Solver::TensorBased::throwError ( const std::string & functionName,
const std::string & errorMsg ) const
protected

Throw an error with a method's name and error message.

References NOX::Utils::Error, and utilsPtr.

Referenced by computeTensorDirection().

Member Data Documentation

◆ alpha

double NOX::Solver::TensorBased::alpha
protected

Scaling factor for the Armijo-Goldstein condition.

Referenced by performLinesearch(), and reset().

◆ aVecPtr

Teuchos::RCP<NOX::Abstract::Vector> NOX::Solver::TensorBased::aVecPtr
protected

Current tensor term vector (pointer).

We have both a pointer and a reference because we need to create a DERIVED object and then want to have a reference to it.

Referenced by computeTensorDirection(), getNormModelResidual(), and TensorBased().

◆ beta

double NOX::Solver::TensorBased::beta
protected

Value of sc'*dt.

Referenced by computeCurvilinearStep(), computeTensorDirection(), and TensorBased().

◆ checkType

NOX::StatusTest::CheckType NOX::Solver::TensorBased::checkType
protected

Type of check to use for status tests. See NOX::StatusTest for more details.

Referenced by reset(), and step().

◆ convCriteria

ConvergenceCriteriaType NOX::Solver::TensorBased::convCriteria
protected

Choice of convergence criteria (currently unused).

Referenced by reset().

◆ counter

NOX::LineSearchCounters* NOX::Solver::TensorBased::counter
protected

Common common counters for line searches.

Referenced by implementGlobalStrategy(), init(), performLinesearch(), reset(), and reset().

◆ defaultStep

double NOX::Solver::TensorBased::defaultStep
protected

Default step.

Referenced by performLinesearch(), and reset().

◆ doRescue

bool NOX::Solver::TensorBased::doRescue
protected

Flag for rescuing Linear Solver from a bad solve.

Referenced by computeTensorDirection(), and reset().

◆ globalDataPtr

Teuchos::RCP<NOX::GlobalData> NOX::Solver::TensorBased::globalDataPtr
protected

Pointer to the global data object.

Referenced by getSolverStatistics(), and reset().

◆ isNewtonDirection

bool NOX::Solver::TensorBased::isNewtonDirection
protected

Flag for Newton direction.

Referenced by implementGlobalStrategy(), and performLinesearch().

◆ lambdaSelection

LambdaSelectionType NOX::Solver::TensorBased::lambdaSelection
protected

Flag for lambda selection (Halving/Quadratic).

Referenced by reset(), and selectLambda().

◆ linearParamsPtr

Teuchos::ParameterList* NOX::Solver::TensorBased::linearParamsPtr
protected

Line Search parameters.

Direction parameters. Parameters for the Linear Solver of the local model.

Referenced by TensorBased().

◆ lsType

LineSearchType NOX::Solver::TensorBased::lsType
protected

Choice of line search.

Referenced by implementGlobalStrategy(), performLinesearch(), and reset().

◆ maxIters

int NOX::Solver::TensorBased::maxIters
protected

Maximum iterations.

Referenced by performLinesearch(), and reset().

◆ minStep

double NOX::Solver::TensorBased::minStep
protected

Minimum step length (i.e., when we give up).

Referenced by performLinesearch(), and reset().

◆ newtonVecPtr

Teuchos::RCP<NOX::Abstract::Vector> NOX::Solver::TensorBased::newtonVecPtr
protected

Current Newton direction (pointer).

We have both a pointer and a reference because we need to create a DERIVED object and then want to have a reference to it.

Referenced by computeCurvilinearStep(), computeTensorDirection(), implementGlobalStrategy(), performLinesearch(), and TensorBased().

◆ nIter

int NOX::Solver::TensorBased::nIter
protected

Number of nonlinear iterations.

Referenced by computeTensorDirection(), getNumIterations(), init(), printUpdate(), reset(), solve(), and step().

◆ numJ2vMults

int NOX::Solver::TensorBased::numJ2vMults
mutableprotected

Counter for number of "double" Jacobian-vector products.

Referenced by computeTensorDirection(), init(), reset(), and ~TensorBased().

◆ numJvMults

int NOX::Solver::TensorBased::numJvMults
mutableprotected

Counter for number of Jacobian-vector products.

Referenced by computeTensorDirection(), getDirectionalDerivative(), getNormModelResidual(), init(), performLinesearch(), reset(), and ~TensorBased().

◆ observer

Teuchos::RCP<NOX::Observer> NOX::Solver::TensorBased::observer
protected

Pointer to a user defined NOX::Observer object.

Referenced by reset(), solve(), and step().

◆ oldSolnPtr

Teuchos::RCP<NOX::Abstract::Group> NOX::Solver::TensorBased::oldSolnPtr
protected

Previous solution pointer.

We have both a pointer and a reference because we need to create a DERIVED object and then want to have a reference to it.

Referenced by getPreviousSolutionGroup(), getPreviousSolutionGroupPtr(), step(), and TensorBased().

◆ paramsPtr

Teuchos::RCP<Teuchos::ParameterList> NOX::Solver::TensorBased::paramsPtr
protected

Input parameters.

Referenced by computeTensorDirection(), getList(), getListPtr(), init(), performLinesearch(), reset(), solve(), and TensorBased().

◆ print

Teuchos::RCP<NOX::LineSearch::Utils::Printing> NOX::Solver::TensorBased::print
protected

Common line search printing utilities.

Referenced by performLinesearch(), and reset().

◆ recoveryStep

double NOX::Solver::TensorBased::recoveryStep
protected

Default step for linesearch failure.

Referenced by performLinesearch(), and reset().

◆ recoveryStepType

RecoveryStepType NOX::Solver::TensorBased::recoveryStepType
protected

Choice of the recovery step type; uses "Recovery Step Type" parameter.

Referenced by performLinesearch(), and reset().

◆ requestedBaseStep

StepType NOX::Solver::TensorBased::requestedBaseStep
protected

Flag for the base direction to compute after the first iteration.

Referenced by computeTensorDirection(), and reset().

◆ residualVecPtr

Teuchos::RCP<NOX::Abstract::Vector> NOX::Solver::TensorBased::residualVecPtr
protected

Residual vector (pointer).

We have both a pointer and a reference because we need to create a DERIVED object and then want to have a reference to it.

Referenced by computeTensorDirection(), and TensorBased().

◆ slopeObj

NOX::LineSearch::Utils::Slope NOX::Solver::TensorBased::slopeObj
protected

Common slope calculations for line searches.

Referenced by implementGlobalStrategy(), performLinesearch(), and reset().

◆ solnPtr

Teuchos::RCP<NOX::Abstract::Group> NOX::Solver::TensorBased::solnPtr
protected

Current solution.

Referenced by getSolutionGroup(), getSolutionGroupPtr(), printUpdate(), reset(), reset(), reset(), solve(), step(), and TensorBased().

◆ status

NOX::StatusTest::StatusType NOX::Solver::TensorBased::status
protected

Status of nonlinear solver.

Referenced by getStatus(), init(), printUpdate(), reset(), solve(), and step().

◆ stepSize

double NOX::Solver::TensorBased::stepSize
protected

Current step.

Referenced by init(), printUpdate(), reset(), and step().

◆ sTinvJa

double NOX::Solver::TensorBased::sTinvJa
protected

Value of s'*inv(J)*a.

Referenced by computeCurvilinearStep(), computeTensorDirection(), and TensorBased().

◆ sTinvJF

double NOX::Solver::TensorBased::sTinvJF
protected

Value of s'*inv(J)*F.

Referenced by computeCurvilinearStep(), computeTensorDirection(), and TensorBased().

◆ sVecPtr

Teuchos::RCP<NOX::Abstract::Vector> NOX::Solver::TensorBased::sVecPtr
protected

Vector to previous point (pointer).

We have both a pointer and a reference because we need to create a DERIVED object and then want to have a reference to it.

Referenced by computeTensorDirection(), getNormModelResidual(), and TensorBased().

◆ tensorVecPtr

Teuchos::RCP<NOX::Abstract::Vector> NOX::Solver::TensorBased::tensorVecPtr
protected

Current tensor direction (pointer).

We have both a pointer and a reference because we need to create a DERIVED object and then want to have a reference to it.

Referenced by computeCurvilinearStep(), computeTensorDirection(), implementGlobalStrategy(), printUpdate(), and TensorBased().

◆ testPtr

Teuchos::RCP<NOX::StatusTest::Generic> NOX::Solver::TensorBased::testPtr
protected

Stopping test.

Referenced by printUpdate(), reset(), reset(), step(), and TensorBased().

◆ tmpVecPtr

Teuchos::RCP<NOX::Abstract::Vector> NOX::Solver::TensorBased::tmpVecPtr
protected

Working vector (pointer).

We have both a pointer and a reference because we need to create a DERIVED object and then want to have a reference to it.

Referenced by computeTensorDirection(), performLinesearch(), and TensorBased().

◆ useModifiedMethod

bool NOX::Solver::TensorBased::useModifiedMethod
protected

Flag for using modifications that force quadratic to have real root.

Referenced by computeTensorDirection(), and reset().

◆ utilsPtr

Teuchos::RCP<NOX::Utils> NOX::Solver::TensorBased::utilsPtr
protected

Printing Utils.

Referenced by calculateBeta(), computeCurvilinearStep(), computeTensorDirection(), getNormModelResidual(), init(), performLinesearch(), printDirectionInfo(), printUpdate(), reset(), step(), throwError(), and ~TensorBased().

The documentation for this class was generated from the following files:
Generated by doxygen 1.18.0