com.imsl.math
Class ZeroSystem

java.lang.Object
  com.imsl.math.ZeroSystem

All Implemented Interfaces:

Cloneable, Serializable

public class ZeroSystem

extends Object

implements Serializable, Cloneable

Solves a system of n nonlinear equations f(x) = 0 using a modified Powell hybrid algorithm.

ZeroSystem is based on the MINPACK subroutine HYBRD1, which uses a modification of M.J.D. Powell's hybrid algorithm. This algorithm is a variation of Newton's method, which uses a finite-difference approximation to the Jacobian and takes precautions to avoid large step sizes or increasing residuals. For further description, see More et al. (1980).

A finite-difference method is used to estimate the Jacobian. Whenever the exact Jacobian can be easily provided, objectF should implement ZeroSystem.Jacobian.

See Also:

Example, Serialized Form

Nested Class Summary

static class	ZeroSystem.DidNotConvergeException The iteration did not converge.
static interface	ZeroSystem.Function Public interface for user supplied function to ZeroSystem object.
static interface	ZeroSystem.Jacobian Public interface for user supplied function to ZeroSystem object.
static class	ZeroSystem.ToleranceTooSmallException Tolerance too small
static class	ZeroSystem.TooManyIterationsException Too many iterations.

Constructor Summary

ZeroSystem(int n)
Creates an object to find the zeros of a system of n equations.

Method Summary

void	setGuess(double[] xguess) Sets the initial guess for the array x.
void	setMaxIterations(int maxIterations) Sets the maximum number of iterations allowed.
void	setRelativeError(double errorRelative) Sets the relative error tolerance.
double[]	solve(ZeroSystem.Function objectF) Solve a system of nonlinear equations using the Levenberg-Marquardt algorithm

Methods inherited from class java.lang.Object

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

Constructor Detail

ZeroSystem

public ZeroSystem(int n)

Creates an object to find the zeros of a system of n equations.

Parameters:

n - is the number of equations that the solver handles

Method Detail

setGuess

public void setGuess(double[] xguess)

Sets the initial guess for the array x. The default is to set x to all zeros.

Parameters:

xguess - a double array containing the initial guess

setMaxIterations

public void setMaxIterations(int maxIterations)

Sets the maximum number of iterations allowed. The default value is 200.

Parameters:

maxIterations - an int specifying the maximum number of iterations allowed

Throws:

IllegalArgumentException - is thrown if maxIterations is less than or equal to zero.

setRelativeError

public void setRelativeError(double errorRelative)

Sets the relative error tolerance. The root is accepted if the relative error between two successive approximations to this root is within errorRelative. The default is the square root of the precision, about 1.0e-08.

Parameters:

errorRelative - a double specifying the relative error tolerance

Throws:

IllegalArgumentException - is thrown if errorRelative is less than 0 or greater than 1.

solve

public double[] solve(ZeroSystem.Function objectF)
               throws ZeroSystem.TooManyIterationsException,
                      ZeroSystem.ToleranceTooSmallException,
                      ZeroSystem.DidNotConvergeException

Solve a system of nonlinear equations using the Levenberg-Marquardt algorithm

Parameters:

objectF - defines the function whose zero is to be found. If objectF implements a Jacobian then its Jacobian is used. Otherwise a finite difference is computed.

Returns:

a double array containing the solution

Throws:

ZeroSystem.TooManyIterationsException - is thrown if the maximum number of iterations is exceeded

ZeroSystem.ToleranceTooSmallException - is thrown if the error tolerance is too small

ZeroSystem.DidNotConvergeException - is thrown if the algorithm does not converge

See Also:

ZeroSystem.Jacobian