net.sf.tweety.math.opt

Class NewtonRootFinder

java.lang.Object

net.sf.tweety.math.opt.RootFinder

net.sf.tweety.math.opt.NewtonRootFinder

public class NewtonRootFinder
extends RootFinder

This class implements the Newton method for finding zeros of a function.

Author:

Matthias Thimm

Field Summary

Fields

Modifier and Type	Field and Description
private java.util.List<java.util.List<Term>>	jacobian The jacobian of the function.
static int	MAX_FIX_ITERATIONS The maximum number of fixing iterations.
static double	PRECISION The precision of the approximation.

Constructor Summary

Constructors

Constructor and Description
NewtonRootFinder(java.util.List<Term> functions, java.util.Map<Variable,Term> startingPoint) Creates a new Newton root finder for the given starting point and the given (multi-dimensional) function
NewtonRootFinder(Term function, java.util.Map<Variable,Term> startingPoint) Creates a new Newton root finder for the given starting point and the given function

Method Summary

Modifier and Type	Method and Description
private java.util.List<java.lang.Double>	approximate(java.util.List<java.util.List<java.lang.Double>> currentJacobianValue, java.util.List<java.lang.Double> currentVector, java.util.List<java.lang.Double> currentValue) Solves the linear equation currentJacobianValue * (X-currentVector) = - currentValue.
private java.util.List<java.util.List<java.lang.Double>>	evaluateMatrix(java.util.List<java.util.List<Term>> functions, java.util.Map<Variable,? extends Term> mapping) Evaluates each function in the given matrix with the given values for variables.
private java.util.Map<Variable,Term>	midpoint(java.util.Map<Variable,Term> m1, java.util.Map<Variable,Term> m2) Computes the midpoint of the two maps
java.util.Map<Variable,Term>	randomRoot() Determines the values for the variables appearing in the function such the function evaluates to zero.

Methods inherited from class net.sf.tweety.math.opt.RootFinder

getFunctions, getStartingPoint, setStartingPoint

Methods inherited from class java.lang.Object

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

Field Detail

jacobian

private java.util.List<java.util.List<Term>> jacobian

The jacobian of the function.

PRECISION

public static final double PRECISION

The precision of the approximation. The actual used precision depends on the number of variables.

See Also:

Constant Field Values

MAX_FIX_ITERATIONS

public static final int MAX_FIX_ITERATIONS

The maximum number of fixing iterations.

See Also:

Constant Field Values

Constructor Detail

NewtonRootFinder

public NewtonRootFinder(Term function,
                        java.util.Map<Variable,Term> startingPoint)

Creates a new Newton root finder for the given starting point and the given function

Parameters:

startingPoint -

NewtonRootFinder

public NewtonRootFinder(java.util.List<Term> functions,
                        java.util.Map<Variable,Term> startingPoint)

Creates a new Newton root finder for the given starting point and the given (multi-dimensional) function

Parameters:

startingPoint -

Method Detail

randomRoot

public java.util.Map<Variable,Term> randomRoot()
                                        throws GeneralMathException

Description copied from class: RootFinder

Determines the values for the variables appearing in the function such the function evaluates to zero.

Specified by:

randomRoot in class RootFinder

Returns:

a map from variables to terms such that "function" evaluates to zero.

Throws:

GeneralMathException - if something went wrong.

midpoint

private java.util.Map<Variable,Term> midpoint(java.util.Map<Variable,Term> m1,
                                              java.util.Map<Variable,Term> m2)

Computes the midpoint of the two maps

Parameters:

m1 -

m2 -

approximate

private java.util.List<java.lang.Double> approximate(java.util.List<java.util.List<java.lang.Double>> currentJacobianValue,
                                                     java.util.List<java.lang.Double> currentVector,
                                                     java.util.List<java.lang.Double> currentValue)
                                              throws GeneralMathException

Solves the linear equation currentJacobianValue * (X-currentVector) = - currentValue.

Returns:

the next guess for the approximation.

Throws:

GeneralMathException

evaluateMatrix

private java.util.List<java.util.List<java.lang.Double>> evaluateMatrix(java.util.List<java.util.List<Term>> functions,
                                                                        java.util.Map<Variable,? extends Term> mapping)

Evaluates each function in the given matrix with the given values for variables.

Parameters:

functions -

mapping -