net.sourceforge.cilib.math.random.GaussianDistribution Maven / Gradle / Ivy

Go to download
Show more of this group Show more artifacts with this name
Show all versions of cilib-library Show documentation
A library of composable components enabling simpler Computational Intelligence
There is a newer version: 0.8
/**           __  __
 *    _____ _/ /_/ /_    Computational Intelligence Library (CIlib)
 *   / ___/ / / / __ \   (c) CIRG @ UP
 *  / /__/ / / / /_/ /   http://cilib.net
 *  \___/_/_/_/_.___/
 */
package net.sourceforge.cilib.math.random;

import static com.google.common.base.Preconditions.checkArgument;
import net.sourceforge.cilib.controlparameter.ConstantControlParameter;
import net.sourceforge.cilib.controlparameter.ControlParameter;
import net.sourceforge.cilib.math.random.generator.Rand;

public class GaussianDistribution implements ProbabilityDistributionFunction {
    private ControlParameter mean;
    private ControlParameter deviation;

    /**
     * Default constructor.
     */
    public GaussianDistribution() {
        mean = ConstantControlParameter.of(0.0);
        deviation = ConstantControlParameter.of(1.0);
    }

    /**
     * Get a Gaussian number with currently set mean and deviation.
     * @return A Gaussian number with mean 0.0 and deviation 1.0.
     */
    @Override
    public double getRandomNumber() {
        return getRandomNumber(mean.getParameter(), deviation.getParameter());
    }

    /**
     * Return a random number with the mean of mean and a deviation of
     * deviation. Based on the formula:
s*U(0, 1) + m == U(m, s)
     *
     * Two parameters are required. The first specifies the location, the second
     * specifies the scale.
     *
     * 
     * ALGORITHM 712, COLLECTED ALGORITHMS FROM ACM.

     * THIS WORK PUBLISHED IN TRANSACTIONS ON MATHEMATICAL SOFTWARE,

     * VOL. 18, NO. 4, DECEMBER, 1992, PP. 434-435.

     * The function returns a normally distributed pseudo-random number

     * with a given mean and standard deviation.  Calls are made to a

     * function subprogram which must return independent random

     * numbers uniform in the interval (0,1).
     * 
     * The algorithm uses the ratio of uniforms method of A.J. Kinderman
     * and J.F. Monahan augmented with quadratic bounding curves.
     *
     * @param location The mean to use.
     * @param scale The deviation to use.
     * @return A Gaussian number with mean location and deviation scale
     */
    @Override
    public double getRandomNumber(double... locationScale) {
        checkArgument(locationScale.length == 2, "The Gaussian distribution requires two parameters. The first specifies the mean, the second specifies the deviation.");
        double q, u, v, x, y;

        /*
        Generate P = (u,v) uniform in rect. enclosing acceptance region
        Make sure that any random numbers <= 0 are rejected, since
        gaussian() requires uniforms > 0, but nextDouble() delivers >= 0.
         */
        do {
            u = Rand.nextDouble();
            v = Rand.nextDouble();

            if (u <= 0.0 || v <= 0.0) {
                u = 1.0;
                v = 1.0;
            }
            v = 1.7156 * (v - 0.5);

            /*  Evaluate the quadratic form */
            x = u - 0.449871;
            y = Math.abs(v) + 0.386595;
            q = x * x + y * (0.19600 * y - 0.25472 * x);

            /* Accept P if inside inner ellipse */
            if (q < 0.27597) {
                break;
            }

            /*  Reject P if outside outer ellipse, or outside acceptance region */
        } while ((q > 0.27846) || (v * v > -4.0 * Math.log(u) * u * u));

        /*  Return ratio of P's coordinates as the normal deviate */
        return (locationScale[0] + locationScale[1] * v / u);
    }

    public void setDeviation(ControlParameter deviation) {
        this.deviation = deviation;
    }

    public ControlParameter getDeviation() {
        return deviation;
    }

    public void setMean(ControlParameter mean) {
        this.mean = mean;
    }

    public ControlParameter getMean() {
        return mean;
    }
}