diff --git a/commons-math-legacy/src/main/java/org/apache/commons/math4/legacy/optim/nonlinear/scalar/noderiv/CMAESOptimizer.java b/commons-math-legacy/src/main/java/org/apache/commons/math4/legacy/optim/nonlinear/scalar/noderiv/CMAESOptimizer.java index d6573306a..7606cee33 100644 --- a/commons-math-legacy/src/main/java/org/apache/commons/math4/legacy/optim/nonlinear/scalar/noderiv/CMAESOptimizer.java +++ b/commons-math-legacy/src/main/java/org/apache/commons/math4/legacy/optim/nonlinear/scalar/noderiv/CMAESOptimizer.java @@ -22,7 +22,6 @@ import java.util.Arrays; import java.util.List; import org.apache.commons.math4.legacy.exception.DimensionMismatchException; -import org.apache.commons.math4.legacy.exception.NotPositiveException; import org.apache.commons.math4.legacy.exception.NotStrictlyPositiveException; import org.apache.commons.math4.legacy.exception.OutOfRangeException; import org.apache.commons.math4.legacy.exception.TooManyEvaluationsException; @@ -35,6 +34,7 @@ import org.apache.commons.math4.legacy.optim.OptimizationData; import org.apache.commons.math4.legacy.optim.PointValuePair; import org.apache.commons.math4.legacy.optim.nonlinear.scalar.GoalType; import org.apache.commons.math4.legacy.optim.nonlinear.scalar.PopulationSize; +import org.apache.commons.math4.legacy.optim.nonlinear.scalar.Sigma; import org.apache.commons.math4.legacy.optim.nonlinear.scalar.MultivariateOptimizer; import org.apache.commons.rng.UniformRandomProvider; import org.apache.commons.statistics.distribution.ContinuousDistribution; @@ -78,14 +78,6 @@ import org.apache.commons.math4.core.jdkmath.JdkMath; *
  • Wikipedia
    • * * - *

    - * The {@link PopulationSize number of offsprings} is the primary strategy - * parameter. In the absence of better clues, a good default could be an integer - * close to {@code 4 + 3 ln(n)}, where {@code n} is the number of optimized - * parameters. Increasing the population size improves global search properties - * at the expense of speed (which in general decreases at most linearly with - * increasing population size). - * * @since 3.0 */ public class CMAESOptimizer @@ -278,44 +270,6 @@ public class CMAESOptimizer return statisticsDHistory; } - /** - * Input sigma values. - * They define the initial coordinate-wise standard deviations for - * sampling new search points around the initial guess. - * It is suggested to set them to the estimated distance from the - * initial to the desired optimum. - * Small values induce the search to be more local (and very small - * values are more likely to find a local optimum close to the initial - * guess). - * Too small values might however lead to early termination. - */ - public static class Sigma implements OptimizationData { - /** Sigma values. */ - private final double[] sigma; - - /** - * @param s Sigma values. - * @throws NotPositiveException if any of the array entries is smaller - * than zero. - */ - public Sigma(double[] s) { - for (int i = 0; i < s.length; i++) { - if (s[i] < 0) { - throw new NotPositiveException(s[i]); - } - } - - sigma = s.clone(); - } - - /** - * @return the sigma values. - */ - public double[] getSigma() { - return sigma.clone(); - } - } - /** * {@inheritDoc} * @@ -323,14 +277,32 @@ public class CMAESOptimizer * {@link MultivariateOptimizer#parseOptimizationData(OptimizationData[]) * MultivariateOptimizer}, this method will register the following data: *

    * @return {@inheritDoc} - * @throws TooManyEvaluationsException if the maximal number of - * evaluations is exceeded. - * @throws DimensionMismatchException if the initial guess, target, and weight - * arguments have inconsistent dimensions. + * @throws TooManyEvaluationsException if the maximal number of evaluations + * is exceeded. + * @throws DimensionMismatchException if the initial guess, target, and + * weight arguments have inconsistent dimensions. */ @Override public PointValuePair optimize(OptimizationData... optData) { diff --git a/commons-math-legacy/src/test/java/org/apache/commons/math4/legacy/optim/nonlinear/scalar/noderiv/CMAESOptimizerTest.java b/commons-math-legacy/src/test/java/org/apache/commons/math4/legacy/optim/nonlinear/scalar/noderiv/CMAESOptimizerTest.java index 830b75345..bf2a609df 100644 --- a/commons-math-legacy/src/test/java/org/apache/commons/math4/legacy/optim/nonlinear/scalar/noderiv/CMAESOptimizerTest.java +++ b/commons-math-legacy/src/test/java/org/apache/commons/math4/legacy/optim/nonlinear/scalar/noderiv/CMAESOptimizerTest.java @@ -18,7 +18,7 @@ package org.apache.commons.math4.legacy.optim.nonlinear.scalar.noderiv; import org.apache.commons.math4.legacy.analysis.MultivariateFunction; import org.apache.commons.math4.legacy.exception.DimensionMismatchException; -import org.apache.commons.math4.legacy.exception.NotPositiveException; +import org.apache.commons.math4.legacy.exception.NotStrictlyPositiveException; import org.apache.commons.math4.legacy.exception.NumberIsTooLargeException; import org.apache.commons.math4.legacy.exception.NumberIsTooSmallException; import org.apache.commons.math4.legacy.exception.OutOfRangeException; @@ -28,6 +28,7 @@ import org.apache.commons.math4.legacy.optim.PointValuePair; import org.apache.commons.math4.legacy.optim.SimpleBounds; import org.apache.commons.math4.legacy.optim.nonlinear.scalar.GoalType; import org.apache.commons.math4.legacy.optim.nonlinear.scalar.PopulationSize; +import org.apache.commons.math4.legacy.optim.nonlinear.scalar.Sigma; import org.apache.commons.math4.legacy.optim.nonlinear.scalar.ObjectiveFunction; import org.apache.commons.math4.legacy.optim.nonlinear.scalar.TestFunction; import org.apache.commons.rng.simple.RandomSource; @@ -81,7 +82,7 @@ public class CMAESOptimizerTest { 1e-13, 1e-6, 100000, expected); } - @Test(expected = NotPositiveException.class) + @Test(expected = NotStrictlyPositiveException.class) public void testInputSigmaNegative() { final int dim = 12; double[] startPoint = OptimTestUtils.point(dim, 0.5); @@ -175,7 +176,7 @@ public class CMAESOptimizerTest { SimpleBounds.unbounded(1), GoalType.MINIMIZE, new PopulationSize(5), - new CMAESOptimizer.Sigma(sigma), + new Sigma(sigma), new InitialGuess(start)).getPoint(); Assert.assertEquals(0, result[0], 1e-7); } @@ -396,7 +397,7 @@ public class CMAESOptimizerTest { new ObjectiveFunction(fitnessFunction), GoalType.MINIMIZE, new PopulationSize(5), - new CMAESOptimizer.Sigma(sigma), + new Sigma(sigma), new InitialGuess(start), new SimpleBounds(lower, upper)).getPoint(); Assert.assertTrue("Out of bounds (" + result[0] + " > " + upper[0] + ")", @@ -428,7 +429,7 @@ public class CMAESOptimizerTest { GoalType.MINIMIZE, SimpleBounds.unbounded(1), new PopulationSize(5), - new CMAESOptimizer.Sigma(new double[] { 1e-1 }), + new Sigma(new double[] { 1e-1 }), new InitialGuess(start)); final double resNoBound = result.getPoint()[0]; @@ -440,7 +441,7 @@ public class CMAESOptimizerTest { new ObjectiveFunction(fitnessFunction), GoalType.MINIMIZE, new PopulationSize(5), - new CMAESOptimizer.Sigma(sigma), + new Sigma(sigma), new InitialGuess(start), new SimpleBounds(lower, upper)); final double resNearLo = result.getPoint()[0]; @@ -452,7 +453,7 @@ public class CMAESOptimizerTest { new ObjectiveFunction(fitnessFunction), GoalType.MINIMIZE, new PopulationSize(5), - new CMAESOptimizer.Sigma(sigma), + new Sigma(sigma), new InitialGuess(start), new SimpleBounds(lower, upper)); final double resNearHi = result.getPoint()[0]; @@ -505,7 +506,7 @@ public class CMAESOptimizerTest { goal, new InitialGuess(startPoint), SimpleBounds.unbounded(dim), - new CMAESOptimizer.Sigma(inSigma), + new Sigma(inSigma), new PopulationSize(lambda)) : optim.optimize(new MaxEval(maxEvaluations), new ObjectiveFunction(func), @@ -513,7 +514,7 @@ public class CMAESOptimizerTest { new SimpleBounds(boundaries[0], boundaries[1]), new InitialGuess(startPoint), - new CMAESOptimizer.Sigma(inSigma), + new Sigma(inSigma), new PopulationSize(lambda)); Assert.assertEquals(expected.getValue(), result.getValue(), fTol);