gov.sandia.cognition.learning.algorithm.minimization.line.interpolator
Class LineBracketInterpolatorGoldenSection

java.lang.Object
  gov.sandia.cognition.util.AbstractCloneableSerializable
      gov.sandia.cognition.learning.algorithm.minimization.line.interpolator.AbstractLineBracketInterpolator<Evaluator<Double,Double>>
          gov.sandia.cognition.learning.algorithm.minimization.line.interpolator.LineBracketInterpolatorGoldenSection

All Implemented Interfaces:

LineBracketInterpolator<Evaluator<Double,Double>>, CloneableSerializable, Serializable, Cloneable

@PublicationReferences(references={@PublicationReference(author="Wikipedia",title="Golden section search",type=WebPage,url="http://en.wikipedia.org/wiki/Golden_section_search",year=2008),@PublicationReference(author={"Jeffrey Naisbitt","Michael Heath"},title="Golden Section Search",type=WebPage,url="http://www.cse.uiuc.edu/iem/optimization/GoldenSection/",year=2008),@PublicationReference(author={"William H. Press","Saul A. Teukolsky","William T. Vetterling","Brian P. Flannery"},title="Numerical Recipes in C, Second Edition",type=Book,year=1992,pages={401,402},url="http://www.nrbook.com/a/bookcpdf.php")})
public class LineBracketInterpolatorGoldenSection
extends AbstractLineBracketInterpolator<Evaluator<Double,Double>>

Interpolates between the two bound points of a LineBracket using the golden-section step rule, if that step fails, then the interpolator uses a linear (secant) interpolation. Neither of the bound points in the LineBracket need slope information.

Since:

2.1

Author:

Kevin R. Dixon

See Also:

Serialized Form

Field Summary

static double	GOLDEN_RATIO Golden ratio, from the Fibonacci sequence, 1.618034
static double	GOLDEN_RATIO_CONJUGATE The Golden Ratio conjugate, 0.618034

Fields inherited from class gov.sandia.cognition.learning.algorithm.minimization.line.interpolator.AbstractLineBracketInterpolator

DEFAULT_TOLERANCE

Constructor Summary

LineBracketInterpolatorGoldenSection()
Creates a new instance of LineBracketInterpolatorGoldenSection

Method Summary

double	findMinimum(LineBracket bracket, double minx, double maxx, Evaluator<Double,Double> function) Finds the minimum of the bracket using the interpolation/extrapolation routine, where the minimum must lie between the minx and maxx values on the x-axis.
boolean	hasSufficientPoints(LineBracket bracket) Golden section step only needs the bounds to operate, so we'll just check to make sure they're not null.
static double	step(InputOutputPair<Double,Double> a, InputOutputPair<Double,Double> b, double tolerance) Takes a Golden Section step between the two points

Methods inherited from class gov.sandia.cognition.learning.algorithm.minimization.line.interpolator.AbstractLineBracketInterpolator

getTolerance, setTolerance

Methods inherited from class gov.sandia.cognition.util.AbstractCloneableSerializable

clone

Methods inherited from class java.lang.Object

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

Methods inherited from interface gov.sandia.cognition.util.CloneableSerializable

clone

Field Detail

GOLDEN_RATIO

public static final double GOLDEN_RATIO

Golden ratio, from the Fibonacci sequence, 1.618034

See Also:

Constant Field Values

GOLDEN_RATIO_CONJUGATE

public static final double GOLDEN_RATIO_CONJUGATE

The Golden Ratio conjugate, 0.618034

See Also:

Constant Field Values

Constructor Detail

LineBracketInterpolatorGoldenSection

public LineBracketInterpolatorGoldenSection()

Creates a new instance of LineBracketInterpolatorGoldenSection

Method Detail

findMinimum

public double findMinimum(LineBracket bracket,
                          double minx,
                          double maxx,
                          Evaluator<Double,Double> function)

Description copied from interface: LineBracketInterpolator

Finds the minimum of the bracket using the interpolation/extrapolation routine, where the minimum must lie between the minx and maxx values on the x-axis.

Parameters:

bracket - Bracket of points

minx - Minimum x-axis value to search for the minimum

maxx - Maximum x-axis value to search for the minimum

function - Function to consider. The function, or gradient, may be evaluated to generate enough output/slope samples to interpolate.

Returns:

Minimum value of the function between minx and maxx.

step

public static double step(InputOutputPair<Double,Double> a,
                          InputOutputPair<Double,Double> b,
                          double tolerance)

Takes a Golden Section step between the two points

Parameters:

a - First point

b - Second point

tolerance - Minimum below which to consider the points identical.

Returns:

Interpolated Golden Section step between the two points

hasSufficientPoints

public boolean hasSufficientPoints(LineBracket bracket)

Golden section step only needs the bounds to operate, so we'll just check to make sure they're not null.

Parameters:

bracket - LineBracket to consider

Returns:

True if both bounds are not null, false otherwise