nl.tudelft.simulation.jstats.ode.integrators
Class CachingNumericalIntegrator

java.lang.Object
  nl.tudelft.simulation.jstats.ode.integrators.NumericalIntegrator
      nl.tudelft.simulation.jstats.ode.integrators.CachingNumericalIntegrator

Direct Known Subclasses:

Adams, Milne

public abstract class CachingNumericalIntegrator

extends NumericalIntegrator

The CachingNumericalIntegrator is the basis for an integrator that needs access to previously calculated values of y', e.g. y'_(k-1), y'_(k-2), etc.
(c) copyright 2003-2004 Delft University of Technology , the Netherlands.
See for project information www.simulation.tudelft.nl
License of use: General Public License (GPL) , no warranty

Version:

May 2, 2004

Author:

Alexander Verbraeck

Field Summary

protected NumericalIntegrator	startingIntegrator the primer integrator
protected int	startingSubSteps the substeps to use when starting the integrator

Fields inherited from class nl.tudelft.simulation.jstats.ode.integrators.NumericalIntegrator

ADAMS, DEFAULT_INTEGRATOR, equation, error, EULER, GILL, HEUN, MILNE, RUNGEKUTTA3, RUNGEKUTTA4, RUNGEKUTTACASHCARP, RUNGEKUTTAFEHLBERG, timeStep

Constructor Summary

CachingNumericalIntegrator(double timeStep, DifferentialEquationInterface equation, int cachePlaces, short integrationMethod, int startingSubSteps)
constructs a new CachingNumericalIntegrator with a fixed number of cache places.

Method Summary

double[]	getDY(int numberDown) get a cached dY-value,
double[]	getY(int numberDown) get a cached Y-value,
abstract double[]	next(double x) The integrators that extend the CachingNumericalIntegrator calculate the value of y(x+timeStep) just based on the x-value.
double[]	next(double x, double[] y) computes the next value
void	setTimeStep(double timeStep)

Methods inherited from class nl.tudelft.simulation.jstats.ode.integrators.NumericalIntegrator

add, add, add, add, add, getError, getTimeStep, multiply, resolve

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Field Detail

startingIntegrator

protected NumericalIntegrator startingIntegrator

the primer integrator

startingSubSteps

protected int startingSubSteps

the substeps to use when starting the integrator

Constructor Detail

CachingNumericalIntegrator

public CachingNumericalIntegrator(double timeStep,
                                  DifferentialEquationInterface equation,
                                  int cachePlaces,
                                  short integrationMethod,
                                  int startingSubSteps)

constructs a new CachingNumericalIntegrator with a fixed number of cache places.

Parameters:

timeStep - the timeStep

equation - the differentialEquation

cachePlaces - the number of cache places to store

integrationMethod - the primer integrator to use

startingSubSteps - the number of substeps per timestep during starting of the integrator

Method Detail

setTimeStep

public void setTimeStep(double timeStep)

Overrides:

setTimeStep in class NumericalIntegrator

Parameters:

timeStep - The timeStep to set.

See Also:

NumericalIntegrator.setTimeStep(double)

next

public double[] next(double x,
                     double[] y)

Description copied from class: NumericalIntegrator

computes the next value

Specified by:

next in class NumericalIntegrator

Parameters:

x - the x value corresponding to the last y-value computed

y - the last y value

Returns:

the new value

See Also:

NumericalIntegrator.next(double, double[])

getY

public double[] getY(int numberDown)

get a cached Y-value,

Parameters:

numberDown - the number of the previous value we want

Returns:

the corresponding Y-value

getDY

public double[] getDY(int numberDown)

get a cached dY-value,

Parameters:

numberDown - the number of the previous value we want

Returns:

the corresponding dY-value

next

public abstract double[] next(double x)

The integrators that extend the CachingNumericalIntegrator calculate the value of y(x+timeStep) just based on the x-value. They retrieve y(x), y(x-timeStep), etc. or y(k), y(k-1) all from the cache.

Parameters:

x - the x-value to use in the calculation

Returns:

the value of y(x+timeStep)