org.simulator.math.odes
Class RosenbrockSolver

java.lang.Object
  org.simulator.math.odes.AbstractDESSolver
      org.simulator.math.odes.AdaptiveStepsizeIntegrator
          org.simulator.math.odes.RosenbrockSolver

All Implemented Interfaces:

java.io.Serializable, org.apache.commons.math.ode.events.EventHandler, DelayValueHolder, DESSolver

public class RosenbrockSolver
extends AdaptiveStepsizeIntegrator

RosenbrockSolver implements the Rosenbrock method to approximate ODE solutions.

References: William H. Press, Saul A. Teukolsky, William T. Vetterling, and Brian P. Flannery. Numerical recipes in C. Cambridge Univ. Press Cambridge, 1992, pp. 738-747.

This solver has been adapted from ODE Toolkit: a free application for solving systems of ordinary differential equations.

Since:

0.9

Version:

$Rev: 146 $

Author:

Chris Moore, Roland Keller

See Also:

Serialized Form

Field Summary

static double	a21 Constants for solving
static double	a31 Constants for solving
static double	a32 Constants for solving
static double	a41 Constants for solving
static double	a42 Constants for solving
static double	a43 Constants for solving
static double	a51 Constants for solving
static double	a52 Constants for solving
static double	a53 Constants for solving
static double	a54 Constants for solving
static double	c2 Constants for solving
static double	c21 Constants for solving
static double	c3 Constants for solving
static double	c31 Constants for solving
static double	c32 Constants for solving
static double	c4 Constants for solving
static double	c41 Constants for solving
static double	c42 Constants for solving
static double	c43 Constants for solving
static double	c51 Constants for solving
static double	c52 Constants for solving
static double	c53 Constants for solving
static double	c54 Constants for solving
static double	c61 Constants for solving
static double	c62 Constants for solving
static double	c63 Constants for solving
static double	c64 Constants for solving
static double	c65 Constants for solving
static double	d1
static double	d2
static double	d3
static double	d4
(package private) double[]	DFDX helper array to hold intermediate values
(package private) double[][]	FAC helper array to hold intermediate values
static double	fac1 Constants used to adapt the stepsize according to the error in the last step (see rodas.f)
static double	fac2 Constants used to adapt the stepsize according to the error in the last step (see rodas.f)
(package private) double[]	g0 helper array to hold intermediate values
(package private) double[]	g1 helper array to hold intermediate values
(package private) double[]	g1x helper array to hold intermediate values
(package private) double[]	g2 helper array to hold intermediate values
(package private) double[]	g2x helper array to hold intermediate values
static double	gam Constants for solving
(package private) double	hAdap factor used for adjusting the step size, divide current step size by hAdap to get new step size
(package private) double[][]	I helper array to hold intermediate values
(package private) int[]	indx helper array to hold intermediate values
(package private) double[][]	JAC helper array to hold intermediate values
(package private) int	numEqn The number of equations
static double	PWR Constants used to adapt the stepsize according to the error in the last step (see rodas.f)
static double	RELMIN the minimum acceptable value of relTol - attempts to obtain higher accuracy than this are usually very expensive
static double	SAFETY Constants used to adapt the stepsize according to the error in the last step (see rodas.f)
(package private) double	sk factor for calculating error value used in adjusting step size
(package private) boolean	stop Keep track whether the thread is killed or not
(package private) double[]	ya helper array to hold intermediate values
(package private) double[]	yb helper array to hold intermediate values
(package private) double[]	yNew array that is y with approximated errors added on, used for comparing y to y+yerr
(package private) double[]	yTemp helper array to hold intermediate values

Fields inherited from class org.simulator.math.odes.AdaptiveStepsizeIntegrator

absTol, relTol

Fields inherited from class org.simulator.math.odes.AbstractDESSolver

listenerList

Fields inherited from interface org.apache.commons.math.ode.events.EventHandler

CONTINUE, RESET_DERIVATIVES, RESET_STATE, STOP

Constructor Summary

RosenbrockSolver()
default constructor

RosenbrockSolver(int size, double stepsize)

RosenbrockSolver(RosenbrockSolver solver)
clone constructor

Method Summary

AbstractDESSolver	clone()
double[]	computeChange(DESystem DES, double[] y2, double time, double currentStepSize, double[] change, boolean steadyState) Computes the change for a given system at the current time with the current setting for the integration step size.
java.lang.String	getName() This gives a human-readable name of this solver that can be displayed in a graphical user interface.
int	getNumEquations()
protected boolean	hasSolverEventProcessing()
double	step(DESystem DES) This function tries to make a time step.
double	unitRoundoff() Returns an approximation to the error involved with the current arithmetic implementation

Methods inherited from class org.simulator.math.odes.AdaptiveStepsizeIntegrator

getAbsTol, getRelTol, setAbsTol, setRelTol

Methods inherited from class org.simulator.math.odes.AbstractDESSolver

additionalResults, addPropertyChangeListener, checkSolution, checkSolution, computeDelayedValue, computeNextState, eventOccurred, firePropertyChange, g, getSerialversionuid, getStepSize, inBetweenSteps, initResultMatrix, initResultMatrix, initResultMatrix, isIncludeIntermediates, isNonnegative, isUnstable, numSteps, processEvents, processEventsAndRules, processRules, removePropertyChangeListener, reset, resetState, setIncludeIntermediates, setNonnegative, setStepSize, setUnstableFlag, solve, solve, solve, solve

Methods inherited from class java.lang.Object

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

Field Detail

SAFETY

public static final double SAFETY

Constants used to adapt the stepsize according to the error in the last step (see rodas.f)

See Also:

Constant Field Values

fac1

public static final double fac1

Constants used to adapt the stepsize according to the error in the last step (see rodas.f)

See Also:

Constant Field Values

fac2

public static final double fac2

Constants used to adapt the stepsize according to the error in the last step (see rodas.f)

See Also:

Constant Field Values

PWR

public static final double PWR

Constants used to adapt the stepsize according to the error in the last step (see rodas.f)

See Also:

Constant Field Values

c2

public static final double c2

Constants for solving

See Also:

Constant Field Values

c3

public static final double c3

Constants for solving

See Also:

Constant Field Values

c4

public static final double c4

Constants for solving

See Also:

Constant Field Values

a21

public static final double a21

Constants for solving

See Also:

Constant Field Values

a31

public static final double a31

Constants for solving

See Also:

Constant Field Values

a32

public static final double a32

Constants for solving

See Also:

Constant Field Values

a41

public static final double a41

Constants for solving

See Also:

Constant Field Values

a42

public static final double a42

Constants for solving

See Also:

Constant Field Values

a43

public static final double a43

Constants for solving

See Also:

Constant Field Values

a51

public static final double a51

Constants for solving

See Also:

Constant Field Values

a52

public static final double a52

Constants for solving

See Also:

Constant Field Values

a53

public static final double a53

Constants for solving

See Also:

Constant Field Values

a54

public static final double a54

Constants for solving

See Also:

Constant Field Values

gam

public static final double gam

Constants for solving

See Also:

Constant Field Values

c21

public static final double c21

Constants for solving

See Also:

Constant Field Values

c31

public static final double c31

Constants for solving

See Also:

Constant Field Values

c32

public static final double c32

Constants for solving

See Also:

Constant Field Values

c41

public static final double c41

Constants for solving

See Also:

Constant Field Values

c42

public static final double c42

Constants for solving

See Also:

Constant Field Values

c43

public static final double c43

Constants for solving

See Also:

Constant Field Values

c51

public static final double c51

Constants for solving

See Also:

Constant Field Values

c52

public static final double c52

Constants for solving

See Also:

Constant Field Values

c53

public static final double c53

Constants for solving

See Also:

Constant Field Values

c54

public static final double c54

Constants for solving

See Also:

Constant Field Values

c61

public static final double c61

Constants for solving

See Also:

Constant Field Values

c62

public static final double c62

Constants for solving

See Also:

Constant Field Values

c63

public static final double c63

Constants for solving

See Also:

Constant Field Values

c64

public static final double c64

Constants for solving

See Also:

Constant Field Values

c65

public static final double c65

Constants for solving

See Also:

Constant Field Values

d1

public static final double d1

See Also:

Constant Field Values

d2

public static final double d2

See Also:

Constant Field Values

d3

public static final double d3

See Also:

Constant Field Values

d4

public static final double d4

See Also:

Constant Field Values

RELMIN

public static final double RELMIN

the minimum acceptable value of relTol - attempts to obtain higher accuracy than this are usually very expensive

See Also:

Constant Field Values

sk

double sk

factor for calculating error value used in adjusting step size

hAdap

double hAdap

factor used for adjusting the step size, divide current step size by hAdap to get new step size

numEqn

int numEqn

The number of equations

yNew

double[] yNew

array that is y with approximated errors added on, used for comparing y to y+yerr

yTemp

double[] yTemp

helper array to hold intermediate values

ya

double[] ya

helper array to hold intermediate values

yb

double[] yb

helper array to hold intermediate values

g0

double[] g0

helper array to hold intermediate values

g1

double[] g1

helper array to hold intermediate values

g2

double[] g2

helper array to hold intermediate values

g1x

double[] g1x

helper array to hold intermediate values

g2x

double[] g2x

helper array to hold intermediate values

DFDX

double[] DFDX

helper array to hold intermediate values

indx

int[] indx

helper array to hold intermediate values

JAC

double[][] JAC

helper array to hold intermediate values

FAC

double[][] FAC

helper array to hold intermediate values

I

double[][] I

helper array to hold intermediate values

stop

boolean stop

Keep track whether the thread is killed or not

Constructor Detail

RosenbrockSolver

public RosenbrockSolver()

default constructor

Parameters:

size -

stepsize -

RosenbrockSolver

public RosenbrockSolver(int size,
                        double stepsize)

Parameters:

size -

stepsize -

RosenbrockSolver

public RosenbrockSolver(RosenbrockSolver solver)

clone constructor

Parameters:

solver -

Method Detail

clone

public AbstractDESSolver clone()

Specified by:

clone in interface DESSolver

Specified by:

clone in class AbstractDESSolver

Returns:

the cloned solver

step

public double step(DESystem DES)
            throws org.apache.commons.math.ode.DerivativeException

This function tries to make a time step.

Parameters:

the - differential equation system

Returns:

the error

Throws:

org.apache.commons.math.ode.DerivativeException

unitRoundoff

public double unitRoundoff()

Returns an approximation to the error involved with the current arithmetic implementation

Returns:

the approximation as described above

getName

public java.lang.String getName()

Description copied from class: AbstractDESSolver

This gives a human-readable name of this solver that can be displayed in a graphical user interface.

Specified by:

getName in class AbstractDESSolver

Returns:

A name that describes the underlying algorithm.

getNumEquations

public int getNumEquations()

Returns:

the number of equations in the system

computeChange

public double[] computeChange(DESystem DES,
                              double[] y2,
                              double time,
                              double currentStepSize,
                              double[] change,
                              boolean steadyState)
                       throws org.apache.commons.math.ode.DerivativeException

Description copied from class: AbstractDESSolver

Computes the change for a given system at the current time with the current setting for the integration step size.

Specified by:

computeChange in class AbstractDESSolver

Parameters:

DES - The system to be simulated.

y2 - The current state of the system.

time - The current simulation time.

change - The vector for the resulting change of the system.

Returns:

The change.

Throws:

org.apache.commons.math.ode.DerivativeException

hasSolverEventProcessing

protected boolean hasSolverEventProcessing()

Specified by:

hasSolverEventProcessing in class AbstractDESSolver

Returns:

Does the solver do the event processing itself?