math.fun

Class NumericalDiffDMultiFunction

java.lang.Object

math.fun.NumericalDiffDMultiFunction

All Implemented Interfaces:

DiffDMultiFunction, DMultiFunction, Gradient

public abstract class NumericalDiffDMultiFunction
extends Object
implements DiffDMultiFunction

Finite difference numerical gradient calculation using the forward difference approximation f'(x) = (f(x + h) - f(x)) / h.

Scaling of h is taken into account by each individual h being based upon the absolute magnitude of the corresponding element in the vector x.

Field Summary

Fields

Modifier and Type	Field and Description
protected double	diffScale The scaling factor to use for scaling of the individual h

Constructor Summary

Constructors

Constructor and Description
NumericalDiffDMultiFunction() Creates a new NumericalDiffDMultiFunction with a default scaling factor.
NumericalDiffDMultiFunction(double diffScale) Creates a new NumericalDiffDMultiFunction with the provided scaling factor.

Method Summary

Modifier and Type	Method and Description
void	derivativeAt(double[] x, double[] grad) The first-derivative vector (a.k.a. gradient) of a double-valued function over a double[] array evaluated at the input location x gets stored into the output vector grad.

Methods inherited from class java.lang.Object

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

Methods inherited from interface math.fun.DMultiFunction

apply

Field Detail

diffScale

protected final double diffScale

The scaling factor to use for scaling of the individual h

Constructor Detail

NumericalDiffDMultiFunction

public NumericalDiffDMultiFunction()

Creates a new NumericalDiffDMultiFunction with a default scaling factor.

NumericalDiffDMultiFunction

public NumericalDiffDMultiFunction(double diffScale)

Creates a new NumericalDiffDMultiFunction with the provided scaling factor.

Parameters:

diffScale - scaling factor to use for h

Method Detail

derivativeAt

public final void derivativeAt(double[] x,
                               double[] grad)

The first-derivative vector (a.k.a. gradient) of a double-valued function over a double[] array evaluated at the input location x gets stored into the output vector grad.

Specified by:

derivativeAt in interface Gradient

Parameters:

x - a double[] input vector (not modified)

grad - a double[] output vector containing the gradient at location x (modified)