Module net.finmath.lib
Package net.finmath.optimizer

Class GoldenSectionSearch

  • public class GoldenSectionSearch
extends Object
    This class implements a Golden Section search algorithm, i.e., a minimization, implemented as a question-and-answer search algorithm. Example: 
     
                GoldenSectionSearch search = new GoldenSectionSearch(-1.0, 5.0);
                while(search.getAccuracy() > 1E-11 && !search.isDone()) {
                        double x = search.getNextPoint();

                        double y = (x - 0.656) * (x - 0.656);

                        search.setValue(y);
                }
    For an example on how to use this class see also its main method.
    Version:
    1.1
    Author:
    Christian Fries - http://www.christian-fries.de

    • Field Detail

      • GOLDEN_SECTION_RATIO

        public static final double GOLDEN_SECTION_RATIO

    • Constructor Detail

      • GoldenSectionSearch

        public GoldenSectionSearch​(double leftPoint,
                           double rightPoint)
        Parameters:
        leftPoint - left point of search interval
        rightPoint - right point of search interval

    • Method Detail

      • main

        public static void main​(String[] args)

      • getBestPoint

        public double getBestPoint()
        Returns:
        Returns the best point obtained so far.

      • getNextPoint

        public double getNextPoint()
        Returns the next point for which a valuation is requested.
        Returns:
        Returns the next point for which a value should be set using setValue.

      • setValue

        public void setValue​(double value)
        Set the value corresponding to the point returned by a previous call of getNextPoint(). If setValue is called without prior call to getNextPoint(), e.g., when called twice, a RuntimeException is thrown.
        Parameters:
        value - Value corresponding to point returned by previous getNextPoint() call.

      • value

        public double value​(double parameter)

      • getGoldenSection

        public static double getGoldenSection​(double left,
                                      double right)
        Get the golden section of an interval as gs * left + (1-gs) * right, where gs is GOLDEN_SECTION_RATIO.
        Parameters:
        left - Left point of the interval.
        right - Right point of the interval.
        Returns:
        Returns the golden section of an interval.

      • getNumberOfIterations

        public int getNumberOfIterations()
        Returns:
        Returns the number of iterations needed so far.

      • getAccuracy

        public double getAccuracy()
        Returns:
        Returns the accuracy obtained so far.

      • isDone

        public boolean isDone()
        Returns:
        Returns true if the solver is unable to improve further. This may be either due to reached accuracy or due to no solution existing.