UniformCrossover.java

/*
 * Licensed to the Apache Software Foundation (ASF) under one or more
 * contributor license agreements.  See the NOTICE file distributed with
 * this work for additional information regarding copyright ownership.
 * The ASF licenses this file to You under the Apache License, Version 2.0
 * (the "License"); you may not use this file except in compliance with
 * the License.  You may obtain a copy of the License at
 *
 *      http://www.apache.org/licenses/LICENSE-2.0
 *
 * Unless required by applicable law or agreed to in writing, software
 * distributed under the License is distributed on an "AS IS" BASIS,
 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 * See the License for the specific language governing permissions and
 * limitations under the License.
 */
package org.apache.commons.math4.legacy.genetics;

import java.util.ArrayList;
import java.util.List;

import org.apache.commons.math4.legacy.exception.DimensionMismatchException;
import org.apache.commons.math4.legacy.exception.MathIllegalArgumentException;
import org.apache.commons.math4.legacy.exception.OutOfRangeException;
import org.apache.commons.math4.legacy.exception.util.LocalizedFormats;
import org.apache.commons.rng.UniformRandomProvider;

/**
 * Perform Uniform Crossover [UX] on the specified chromosomes. A fixed mixing
 * ratio is used to combine genes from the first and second parents, e.g. using a
 * ratio of 0.5 would result in approximately 50% of genes coming from each
 * parent. This is typically a poor method of crossover, but empirical evidence
 * suggests that it is more exploratory and results in a larger part of the
 * problem space being searched.
 * <p>
 * This crossover policy evaluates each gene of the parent chromosomes by choosing a
 * uniform random number {@code p} in the range [0, 1]. If {@code p} &lt; {@code ratio},
 * the parent genes are swapped. This means with a ratio of 0.7, 30% of the genes from the
 * first parent and 70% from the second parent will be selected for the first offspring (and
 * vice versa for the second offspring).
 * <p>
 * This policy works only on {@link AbstractListChromosome}, and therefore it
 * is parameterized by T. Moreover, the chromosomes must have same lengths.
 *
 * @see <a href="http://en.wikipedia.org/wiki/Crossover_%28genetic_algorithm%29">Crossover techniques (Wikipedia)</a>
 * @see <a href="http://www.obitko.com/tutorials/genetic-algorithms/crossover-mutation.php">Crossover (Obitko.com)</a>
 * @see <a href="http://www.tomaszgwiazda.com/uniformX.htm">Uniform crossover</a>
 * @param <T> generic type of the {@link AbstractListChromosome}s for crossover
 * @since 3.1
 */
public class UniformCrossover<T> implements CrossoverPolicy {

    /** The mixing ratio. */
    private final double ratio;

    /**
     * Creates a new {@link UniformCrossover} policy using the given mixing ratio.
     *
     * @param ratio the mixing ratio
     * @throws OutOfRangeException if the mixing ratio is outside the [0, 1] range
     */
    public UniformCrossover(final double ratio) throws OutOfRangeException {
        if (ratio < 0.0d || ratio > 1.0d) {
            throw new OutOfRangeException(LocalizedFormats.CROSSOVER_RATE, ratio, 0.0d, 1.0d);
        }
        this.ratio = ratio;
    }

    /**
     * Returns the mixing ratio used by this {@link CrossoverPolicy}.
     *
     * @return the mixing ratio
     */
    public double getRatio() {
        return ratio;
    }

    /**
     * {@inheritDoc}
     *
     * @throws MathIllegalArgumentException iff one of the chromosomes is
     *   not an instance of {@link AbstractListChromosome}
     * @throws DimensionMismatchException if the length of the two chromosomes is different
     */
    @Override
    @SuppressWarnings("unchecked")
    public ChromosomePair crossover(final Chromosome first, final Chromosome second)
        throws DimensionMismatchException, MathIllegalArgumentException {

        if (!(first instanceof AbstractListChromosome<?> && second instanceof AbstractListChromosome<?>)) {
            throw new MathIllegalArgumentException(LocalizedFormats.INVALID_FIXED_LENGTH_CHROMOSOME);
        }
        return mate((AbstractListChromosome<T>) first, (AbstractListChromosome<T>) second);
    }

    /**
     * Helper for {@link #crossover(Chromosome, Chromosome)}. Performs the actual crossover.
     *
     * @param first the first chromosome
     * @param second the second chromosome
     * @return the pair of new chromosomes that resulted from the crossover
     * @throws DimensionMismatchException if the length of the two chromosomes is different
     */
    private ChromosomePair mate(final AbstractListChromosome<T> first,
                                final AbstractListChromosome<T> second) throws DimensionMismatchException {
        final int length = first.getLength();
        if (length != second.getLength()) {
            throw new DimensionMismatchException(second.getLength(), length);
        }

        // array representations of the parents
        final List<T> parent1Rep = first.getRepresentation();
        final List<T> parent2Rep = second.getRepresentation();
        // and of the children
        final List<T> child1Rep = new ArrayList<>(length);
        final List<T> child2Rep = new ArrayList<>(length);

        final UniformRandomProvider random = GeneticAlgorithm.getRandomGenerator();

        for (int index = 0; index < length; index++) {

            if (random.nextDouble() < ratio) {
                // swap the bits -> take other parent
                child1Rep.add(parent2Rep.get(index));
                child2Rep.add(parent1Rep.get(index));
            } else {
                child1Rep.add(parent1Rep.get(index));
                child2Rep.add(parent2Rep.get(index));
            }
        }

        return new ChromosomePair(first.newFixedLengthChromosome(child1Rep),
                                  second.newFixedLengthChromosome(child2Rep));
    }
}