mathcomp.assemble.minimalsurface
Class NewtonAssembler

java.lang.Object
  mathcomp.monitor.DefaultReportable
      mathcomp.monitor.DefaultAlgorithm
          mathcomp.assemble.Assembler
              mathcomp.assemble.LagrangeLinearAssembler
                  mathcomp.assemble.minimalsurface.MinimalSurfaceAssembler
                      mathcomp.assemble.minimalsurface.NewtonAssembler

All Implemented Interfaces:

java.lang.Runnable, Algorithm, Reportable, IteratedAssembler

public class NewtonAssembler
extends MinimalSurfaceAssembler
implements IteratedAssembler

This assembler is used to solve the minimal suface problem using Newtons's iterative method.

Field Summary

Fields inherited from class mathcomp.assemble.LagrangeLinearAssembler

gradVector

Fields inherited from class mathcomp.assemble.Assembler

dirichletBoundaryValues, solScale

Fields inherited from interface mathcomp.run.IteratedAssembler

tolerance

Constructor Summary

NewtonAssembler()

Method Summary

void	assembleGlobal() The main assemble method generally called from outside.
ErrorCalculator	getErrorCalculator()
double	getIt() if x(k+1)=x(k)+g(x(k)) returns 0 (NewtonAssembler), if x(k+1)=g(x(k)) returns -1 (FixedPointAssembler);
double	getOldArea()
void	init()
int	initializeFP() this method is used to decide if some iterations should be done using fixed point iteration before using the Newton's method, since this one only converges if the first iteration is enough close to the exact solution.
boolean	isDamped()
boolean	isToContinue(Vector x) this method defines if the iterative solver stops or continues.
void	setOldSolution(Vector x, Vector yy) This method is used to set the previous iteration that is used to calculate the new iteration
java.lang.String	toString() Set the name of your assembler by overwriting this method in your assembler class.

Methods inherited from class mathcomp.assemble.minimalsurface.MinimalSurfaceAssembler

evalApproxGradient, evalExactGradient, evalExactInitial, evalExactSolution, evalF, getPreferredGridFileName

Methods inherited from class mathcomp.assemble.LagrangeLinearAssembler

createDof, createDofExport, evalApproxGradient, evalApproxSolution, evalBasisFuncOnRef, evalGradXBasisFuncOnRef, evalGradYBasisFuncOnRef

Methods inherited from class mathcomp.assemble.Assembler

addDirichletNodes, calcDirichletBoundary, createSurfaces, createSurfaces, getDof, getElementMatrix, getEpsilon, getExactPartialX, getExactPartialY, getGrid, getInfoUrl, getM, getSolScale, getY, report, run, setEpsilon, setGrid, setM, setTauIndex, setY

Methods inherited from class mathcomp.monitor.DefaultAlgorithm

algoAbortIfNeeded, algoCancel, algoClear, algoCount, algoIncrement, algoMax, isAlgoCancel, setAlgoCount, setAlgoMax

Methods inherited from class mathcomp.monitor.DefaultReportable

getReport, setReport

Methods inherited from class java.lang.Object

equals, getClass, hashCode, notify, notifyAll, wait, wait, wait

Methods inherited from interface mathcomp.monitor.Reportable

getReport, setReport

Constructor Detail

NewtonAssembler

public NewtonAssembler()

Method Detail

assembleGlobal

public void assembleGlobal()

Description copied from class: Assembler

The main assemble method generally called from outside. It once calculates the boundary, the global DOFs and the Dirichlet values. Then Matrix m and right side Vector y are created. Finally an iteration over all elements calls assemble(Grid g, Triangle t, Matrix m) for each element.

Overrides:

assembleGlobal in class Assembler

init

public void init()

Specified by:

init in interface IteratedAssembler

setOldSolution

public void setOldSolution(Vector x,
                           Vector yy)

Description copied from interface: IteratedAssembler

This method is used to set the previous iteration that is used to calculate the new iteration

Specified by:

setOldSolution in interface IteratedAssembler

Parameters:

x - the previous iteration

yy - it's only used in damped Newton's method

getOldArea

public double getOldArea()

Specified by:

getOldArea in interface IteratedAssembler

toString

public java.lang.String toString()

Description copied from class: Assembler

Set the name of your assembler by overwriting this method in your assembler class.

Specified by:

toString in class Assembler

Returns:

the name of the assembler

isToContinue

public boolean isToContinue(Vector x)

Description copied from interface: IteratedAssembler

this method defines if the iterative solver stops or continues. The tolerance is used in this method

Specified by:

isToContinue in interface IteratedAssembler

Parameters:

x - solution of the system of linear equations

Returns:

true if it should contine or false if it should stop

getIt

public double getIt()

Description copied from interface: IteratedAssembler

if x(k+1)=x(k)+g(x(k)) returns 0 (NewtonAssembler), if x(k+1)=g(x(k)) returns -1 (FixedPointAssembler);

Specified by:

getIt in interface IteratedAssembler

isDamped

public boolean isDamped()

Specified by:

isDamped in interface IteratedAssembler

Returns:

true if the iterative solver is damped and false if is not

initializeFP

public int initializeFP()

Description copied from interface: IteratedAssembler

this method is used to decide if some iterations should be done using fixed point iteration before using the Newton's method, since this one only converges if the first iteration is enough close to the exact solution.

Specified by:

initializeFP in interface IteratedAssembler

Returns:

0 (if no iterations should be done using fixed point iteration) or the number of iterations that should be done to initialize the iterative solver

getErrorCalculator

public ErrorCalculator getErrorCalculator()

Overrides:

getErrorCalculator in class Assembler