001 /*
002 * Java Genetic Algorithm Library (jenetics-1.6.0).
003 * Copyright (c) 2007-2014 Franz Wilhelmstötter
004 *
005 * Licensed under the Apache License, Version 2.0 (the "License");
006 * you may not use this file except in compliance with the License.
007 * You may obtain a copy of the License at
008 *
009 * http://www.apache.org/licenses/LICENSE-2.0
010 *
011 * Unless required by applicable law or agreed to in writing, software
012 * distributed under the License is distributed on an "AS IS" BASIS,
013 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
014 * See the License for the specific language governing permissions and
015 * limitations under the License.
016 *
017 * Author:
018 * Franz Wilhelmstötter (franz.wilhelmstoetter@gmx.at)
019 */
020 package org.jenetics;
021
022 import static java.lang.Math.pow;
023 import static java.lang.String.format;
024
025 import java.util.concurrent.atomic.AtomicInteger;
026
027 import org.jenetics.internal.util.HashBuilder;
028
029 import org.jenetics.util.IndexStream;
030 import org.jenetics.util.MSeq;
031
032
033 /**
034 * This class is for mutating a chromosomes of an given population. There are
035 * two distinct roles mutation plays
036 * <ul>
037 * <li>Exploring the search space. By making small moves mutation allows a
038 * population to explore the search space. This exploration is often slow
039 * compared to crossover, but in problems where crossover is disruptive this
040 * can be an important way to explore the landscape.
041 * </li>
042 * <li>Maintaining diversity. Mutation prevents a population from
043 * correlating. Even if most of the search is being performed by crossover,
044 * mutation can be vital to provide the diversity which crossover needs.
045 * </li>
046 * </ul>
047 *
048 * <p>
049 * The mutation probability is the parameter that must be optimized. The optimal
050 * value of the mutation rate depends on the role mutation plays. If mutation is
051 * the only source of exploration (if there is no crossover) then the mutation
052 * rate should be set so that a reasonable neighborhood of solutions is explored.
053 * </p>
054 * The mutation probability <i>P(m)</i> is the probability that a specific gene
055 * over the whole population is mutated. The number of available genes of an
056 * population is
057 * <p>
058 * <img src="doc-files/mutator-N_G.gif" alt="N_P N_{g}=N_P \sum_{i=0}^{N_{G}-1}N_{C[i]}" />
059 * </p>
060 * where <i>N<sub>P</sub></i> is the population size, <i>N<sub>g</sub></i> the
061 * number of genes of a genotype. So the (average) number of genes
062 * mutated by the mutation is
063 * <p>
064 * <img src="doc-files/mutator-mean_m.gif" alt="\hat{\mu}=N_{P}N_{g}\cdot P(m)" />
065 * </p>
066 *
067 * @author <a href="mailto:franz.wilhelmstoetter@gmx.at">Franz Wilhelmstötter</a>
068 * @since 1.0
069 * @version 1.0 — <em>$Date: 2014-02-27 $</em>
070 */
071 public class Mutator<G extends Gene<?, G>> extends AbstractAlterer<G> {
072
073 /**
074 * Construct a Mutation object which a given mutation probability.
075 *
076 * @param probability Mutation probability. The given probability is
077 * divided by the number of chromosomes of the genotype to form
078 * the concrete mutation probability.
079 * @throws IllegalArgumentException if the {@code probability} is not in the
080 * valid range of {@code [0, 1]}..
081 */
082 public Mutator(final double probability) {
083 super(probability);
084 }
085
086 /**
087 * Default constructor, with probability = 0.01.
088 */
089 public Mutator() {
090 this(0.01);
091 }
092
093 /**
094 * Concrete implementation of the alter method.
095 */
096 @Override
097 public <C extends Comparable<? super C>> int alter(
098 final Population<G, C> population,
099 final int generation
100 ) {
101 assert(population != null) : "Not null is guaranteed from base class.";
102
103 final double p = pow(_probability, 1.0/3.0);
104 final AtomicInteger alterations = new AtomicInteger(0);
105
106 final IndexStream stream = IndexStream.Random(population.size(), p);
107 for (int i = stream.next(); i != -1; i = stream.next()) {
108 final Phenotype<G, C> pt = population.get(i);
109
110 final Genotype<G> gt = pt.getGenotype();
111 final Genotype<G> mgt = mutate(gt, p, alterations);
112
113 final Phenotype<G, C> mpt = pt.newInstance(mgt, generation);
114 population.set(i, mpt);
115 }
116
117 return alterations.get();
118 }
119
120 private Genotype<G> mutate(
121 final Genotype<G> genotype,
122 final double p,
123 final AtomicInteger alterations
124 ) {
125 Genotype<G> gt = genotype;
126
127 final IndexStream stream = IndexStream.Random(genotype.length(), p);
128 final int start = stream.next();
129
130 if (start != -1) {
131 final MSeq<Chromosome<G>> chromosomes = genotype.toSeq().copy();
132
133 for (int i = start; i != -1; i = stream.next()) {
134 final Chromosome<G> chromosome = chromosomes.get(i);
135 final MSeq<G> genes = chromosome.toSeq().copy();
136
137 final int mutations = mutate(genes, p);
138 if (mutations > 0) {
139 alterations.addAndGet(mutations);
140 chromosomes.set(i, chromosome.newInstance(genes.toISeq()));
141 }
142 }
143
144 gt = genotype.newInstance(chromosomes.toISeq());
145 }
146
147 return gt;
148 }
149
150 /**
151 * <p>
152 * Template method which gives an (re)implementation of the mutation class
153 * the possibility to perform its own mutation operation, based on a
154 * writable gene array and the gene mutation probability <i>p</i>.
155 * </p>
156 * This implementation, for example, does it in this way:
157 * [code]
158 * protected int mutate(final MSeq〈G〉 genes, final double p) {
159 * final IndexStream stream = IndexStream.Random(genes.length(), p);
160 *
161 * int alterations = 0;
162 * for (int i = stream.next(); i != -1; i = stream.next()) {
163 * genes.set(i, genes.get(i).newInstance());
164 * ++alterations;
165 * }
166 * return alterations;
167 * }
168 * [/code]
169 *
170 * @param genes the genes to mutate.
171 * @param p the gene mutation probability.
172 */
173 protected int mutate(final MSeq<G> genes, final double p) {
174 final IndexStream stream = IndexStream.Random(genes.length(), p);
175
176 int alterations = 0;
177 for (int i = stream.next(); i != -1; i = stream.next()) {
178 genes.set(i, genes.get(i).newInstance());
179 ++alterations;
180 }
181
182 return alterations;
183 }
184
185 @Override
186 public int hashCode() {
187 return HashBuilder.of(getClass()).and(super.hashCode()).value();
188 }
189
190 @Override
191 public boolean equals(final Object obj) {
192 return obj == this || obj instanceof Mutator<?>;
193 }
194
195 @Override
196 public String toString() {
197 return format("%s[p=%f]", getClass().getSimpleName(), _probability);
198 }
199
200 }
|