All Downloads are FREE. Search and download functionalities are using the official Maven repository.

MOEAFramework-2.11.test.org.moeaframework.core.indicator.AdditiveEpsilonIndicatorTest Maven / Gradle / Ivy

/* Copyright 2009-2016 David Hadka
 *
 * This file is part of the MOEA Framework.
 *
 * The MOEA Framework is free software: you can redistribute it and/or modify
 * it under the terms of the GNU Lesser General Public License as published by
 * the Free Software Foundation, either version 3 of the License, or (at your
 * option) any later version.
 *
 * The MOEA Framework is distributed in the hope that it will be useful, but
 * WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY
 * or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU Lesser General Public
 * License for more details.
 *
 * You should have received a copy of the GNU Lesser General Public License
 * along with the MOEA Framework.  If not, see .
 */
package org.moeaframework.core.indicator;

import java.io.IOException;

import org.junit.Assert;
import org.junit.Test;
import org.moeaframework.TestUtils;
import org.moeaframework.core.NondominatedPopulation;
import org.moeaframework.core.Problem;
import org.moeaframework.core.Settings;
import org.moeaframework.core.Solution;
import org.moeaframework.core.spi.ProblemFactory;

/**
 * Tests the {@link AdditiveEpsilonIndicator} class against the JMetal
 * implementation.
 */
public class AdditiveEpsilonIndicatorTest extends IndicatorTest {
	
	/**
	 * Tests if an exception is thrown when using an empty reference set.
	 */
	@Test(expected = IllegalArgumentException.class)
	public void testEmptyReferenceSet() {
		Problem problem = ProblemFactory.getInstance().getProblem("DTLZ2_2");
		NondominatedPopulation referenceSet = new NondominatedPopulation();
		NondominatedPopulation approximationSet = ProblemFactory.getInstance()
				.getReferenceSet("DTLZ2_2");

		AdditiveEpsilonIndicator aei = new AdditiveEpsilonIndicator(problem, 
				referenceSet);
		aei.evaluate(approximationSet);
	}
	
	/**
	 * Tests if an empty approximation set returns a distance of infinity.
	 */
	@Test
	public void testEmptyApproximationSet() {
		Problem problem = ProblemFactory.getInstance().getProblem("DTLZ2_2");
		NondominatedPopulation referenceSet = ProblemFactory.getInstance()
				.getReferenceSet("DTLZ2_2");
		NondominatedPopulation approximationSet = new NondominatedPopulation();

		AdditiveEpsilonIndicator aei = new AdditiveEpsilonIndicator(problem, 
				referenceSet);
		Assert.assertEquals(Double.POSITIVE_INFINITY, 
				aei.evaluate(approximationSet), Settings.EPS);
	}
	
	/**
	 * Tests if infeasible solutions are properly ignored.
	 */
	@Test
	public void testInfeasibleApproximationSet() {
		Problem problem = ProblemFactory.getInstance().getProblem("CF1");
		NondominatedPopulation referenceSet = ProblemFactory.getInstance()
				.getReferenceSet("CF1");
		NondominatedPopulation approximationSet = new NondominatedPopulation();
		
		Solution solution = problem.newSolution();
		solution.setObjectives(new double[] { 0.5, 0.5 });
		solution.setConstraints(new double[] { 10.0 });
		approximationSet.add(solution);
		
		AdditiveEpsilonIndicator aei = new AdditiveEpsilonIndicator(problem, 
				referenceSet);
		Assert.assertEquals(Double.POSITIVE_INFINITY, 
				aei.evaluate(approximationSet), Settings.EPS);
	}
	
	/**
	 * Runs through some simple cases to ensure the generational distance is
	 * computed correctly.
	 */
	@Test
	public void testSimple() {
		Problem problem = ProblemFactory.getInstance().getProblem("DTLZ2_2");
		
		NondominatedPopulation referenceSet = new NondominatedPopulation();
		referenceSet.add(TestUtils.newSolution(0.0, 1.0));
		referenceSet.add(TestUtils.newSolution(1.0, 0.0));
		
		AdditiveEpsilonIndicator aei = new AdditiveEpsilonIndicator(problem, 
				referenceSet);
		
		NondominatedPopulation approximationSet = new NondominatedPopulation();
		
		Assert.assertEquals(Double.POSITIVE_INFINITY, 
				aei.evaluate(approximationSet), Settings.EPS);
		
		approximationSet.add(TestUtils.newSolution(0.0, 1.0));
		Assert.assertEquals(1.0, aei.evaluate(approximationSet), Settings.EPS);
		
		approximationSet.clear();
		approximationSet.add(TestUtils.newSolution(1.0, 1.0));
		Assert.assertEquals(1.0, aei.evaluate(approximationSet), Settings.EPS);
		
		approximationSet.clear();
		approximationSet.add(TestUtils.newSolution(2.0, 2.0));
		Assert.assertEquals(2.0, aei.evaluate(approximationSet), Settings.EPS);
		
		approximationSet.clear();
		approximationSet.add(TestUtils.newSolution(0.0, 0.0));
		Assert.assertEquals(0.0, aei.evaluate(approximationSet), Settings.EPS);
		
		approximationSet.clear();
		approximationSet.add(TestUtils.newSolution(0.0, 1.0));
		approximationSet.add(TestUtils.newSolution(1.0, 0.0));
		Assert.assertEquals(0.0, aei.evaluate(approximationSet), Settings.EPS);

		approximationSet.clear();
		approximationSet.add(TestUtils.newSolution(2.0, 0.0));
		approximationSet.add(TestUtils.newSolution(0.0, 2.0));
		Assert.assertEquals(1.0, aei.evaluate(approximationSet), Settings.EPS);
	}

	/**
	 * Tests the ε-indicator calculation on a 2D continuous Pareto
	 * front.
	 */
	@Test
	public void testDTLZ2_2D() {
		test("DTLZ2_2");
	}

	/**
	 * Tests the ε-indicator calculation on a 4D continuous Pareto
	 * front.
	 */
	@Test
	public void testDTLZ2_4D() {
		test("DTLZ2_4");
	}

	/**
	 * Tests the ε-indicator calculation on a 6D continuous Pareto
	 * front.
	 */
	@Test
	public void testDTLZ2_6D() {
		test("DTLZ2_6");
	}

	/**
	 * Tests the ε-indicator calculation on a 8D continuous Pareto
	 * front.
	 */
	@Test
	public void testDTLZ2_8D() {
		test("DTLZ2_8");
	}

	/**
	 * Tests the ε-indicator calculation on a 2D disconnected Pareto
	 * front.
	 */
	@Test
	public void testDTLZ7_2D() {
		test("DTLZ7_2");
	}

	/**
	 * Tests the ε-indicator calculation on a 4D disconnected Pareto
	 * front.
	 */
	@Test
	public void testDTLZ7_4D() {
		test("DTLZ7_4");
	}

	/**
	 * Tests the ε-indicator calculation on a 6D disconnected Pareto
	 * front.
	 */
	@Test
	public void testDTLZ7_6D() {
		test("DTLZ7_6");
	}

	/**
	 * Tests the ε-indicator calculation on a 8D disconnected Pareto
	 * front.
	 */
	@Test
	public void testDTLZ7_8D() {
		test("DTLZ7_8");
	}

	/**
	 * Generates a random approximation set and tests if the ε-indicator
	 * is computed correctly.
	 * 
	 * @param problemName the problem being tested
	 * @throws IOException should not occur
	 */
	protected void test(String problemName) {
		Problem problem = ProblemFactory.getInstance().getProblem(problemName);
		NondominatedPopulation referenceSet = ProblemFactory.getInstance()
				.getReferenceSet(problemName);
		NondominatedPopulation approximationSet = generateApproximationSet(
				problemName, 100);

		AdditiveEpsilonIndicator myIndicator = new AdditiveEpsilonIndicator(
				problem, referenceSet);
		jmetal.qualityIndicator.Epsilon theirIndicator = 
				new jmetal.qualityIndicator.Epsilon();
		
		Normalizer normalizer = new Normalizer(problem, referenceSet);

		// test against random reference set
		double actual = myIndicator.evaluate(approximationSet);
		double expected = theirIndicator.epsilon(
				toArray(normalizer.normalize(approximationSet)),
				toArray(normalizer.normalize(referenceSet)), 
				problem.getNumberOfObjectives());

		Assert.assertEquals(expected, actual, Settings.EPS);

		// test against reference set
		actual = myIndicator.evaluate(referenceSet);

		Assert.assertEquals(0.0, actual, Settings.EPS);
	}

}




© 2015 - 2024 Weber Informatics LLC | Privacy Policy