001/*
002 * Licensed to the Apache Software Foundation (ASF) under one or more
003 * contributor license agreements.  See the NOTICE file distributed with
004 * this work for additional information regarding copyright ownership.
005 * The ASF licenses this file to You under the Apache License, Version 2.0
006 * (the "License"); you may not use this file except in compliance with
007 * the License.  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
018package org.apache.commons.math3.optimization.univariate;
019
020import java.util.Arrays;
021import java.util.Comparator;
022
023import org.apache.commons.math3.analysis.UnivariateFunction;
024import org.apache.commons.math3.exception.MathIllegalStateException;
025import org.apache.commons.math3.exception.NotStrictlyPositiveException;
026import org.apache.commons.math3.exception.NullArgumentException;
027import org.apache.commons.math3.exception.util.LocalizedFormats;
028import org.apache.commons.math3.random.RandomGenerator;
029import org.apache.commons.math3.optimization.GoalType;
030import org.apache.commons.math3.optimization.ConvergenceChecker;
031
032/**
033 * Special implementation of the {@link UnivariateOptimizer} interface
034 * adding multi-start features to an existing optimizer.
035 *
036 * This class wraps a classical optimizer to use it several times in
037 * turn with different starting points in order to avoid being trapped
038 * into a local extremum when looking for a global one.
039 *
040 * @param <FUNC> Type of the objective function to be optimized.
041 *
042 * @deprecated As of 3.1 (to be removed in 4.0).
043 * @since 3.0
044 */
045@Deprecated
046public class UnivariateMultiStartOptimizer<FUNC extends UnivariateFunction>
047    implements BaseUnivariateOptimizer<FUNC> {
048    /** Underlying classical optimizer. */
049    private final BaseUnivariateOptimizer<FUNC> optimizer;
050    /** Maximal number of evaluations allowed. */
051    private int maxEvaluations;
052    /** Number of evaluations already performed for all starts. */
053    private int totalEvaluations;
054    /** Number of starts to go. */
055    private int starts;
056    /** Random generator for multi-start. */
057    private RandomGenerator generator;
058    /** Found optima. */
059    private UnivariatePointValuePair[] optima;
060
061    /**
062     * Create a multi-start optimizer from a single-start optimizer.
063     *
064     * @param optimizer Single-start optimizer to wrap.
065     * @param starts Number of starts to perform. If {@code starts == 1},
066     * the {@code optimize} methods will return the same solution as
067     * {@code optimizer} would.
068     * @param generator Random generator to use for restarts.
069     * @throws NullArgumentException if {@code optimizer} or {@code generator}
070     * is {@code null}.
071     * @throws NotStrictlyPositiveException if {@code starts < 1}.
072     */
073    public UnivariateMultiStartOptimizer(final BaseUnivariateOptimizer<FUNC> optimizer,
074                                             final int starts,
075                                             final RandomGenerator generator) {
076        if (optimizer == null ||
077                generator == null) {
078                throw new NullArgumentException();
079        }
080        if (starts < 1) {
081            throw new NotStrictlyPositiveException(starts);
082        }
083
084        this.optimizer = optimizer;
085        this.starts = starts;
086        this.generator = generator;
087    }
088
089    /**
090     * {@inheritDoc}
091     */
092    public ConvergenceChecker<UnivariatePointValuePair> getConvergenceChecker() {
093        return optimizer.getConvergenceChecker();
094    }
095
096    /** {@inheritDoc} */
097    public int getMaxEvaluations() {
098        return maxEvaluations;
099    }
100
101    /** {@inheritDoc} */
102    public int getEvaluations() {
103        return totalEvaluations;
104    }
105
106    /**
107     * Get all the optima found during the last call to {@link
108     * #optimize(int,UnivariateFunction,GoalType,double,double) optimize}.
109     * The optimizer stores all the optima found during a set of
110     * restarts. The {@link #optimize(int,UnivariateFunction,GoalType,double,double) optimize}
111     * method returns the best point only. This method returns all the points
112     * found at the end of each starts, including the best one already
113     * returned by the {@link #optimize(int,UnivariateFunction,GoalType,double,double) optimize}
114     * method.
115     * <br/>
116     * The returned array as one element for each start as specified
117     * in the constructor. It is ordered with the results from the
118     * runs that did converge first, sorted from best to worst
119     * objective value (i.e in ascending order if minimizing and in
120     * descending order if maximizing), followed by {@code null} elements
121     * corresponding to the runs that did not converge. This means all
122     * elements will be {@code null} if the {@link
123     * #optimize(int,UnivariateFunction,GoalType,double,double) optimize}
124     * method did throw an exception.
125     * This also means that if the first element is not {@code null}, it is
126     * the best point found across all starts.
127     *
128     * @return an array containing the optima.
129     * @throws MathIllegalStateException if {@link
130     * #optimize(int,UnivariateFunction,GoalType,double,double) optimize}
131     * has not been called.
132     */
133    public UnivariatePointValuePair[] getOptima() {
134        if (optima == null) {
135            throw new MathIllegalStateException(LocalizedFormats.NO_OPTIMUM_COMPUTED_YET);
136        }
137        return optima.clone();
138    }
139
140    /** {@inheritDoc} */
141    public UnivariatePointValuePair optimize(int maxEval, final FUNC f,
142                                                 final GoalType goal,
143                                                 final double min, final double max) {
144        return optimize(maxEval, f, goal, min, max, min + 0.5 * (max - min));
145    }
146
147    /** {@inheritDoc} */
148    public UnivariatePointValuePair optimize(int maxEval, final FUNC f,
149                                                 final GoalType goal,
150                                                 final double min, final double max,
151                                                 final double startValue) {
152        RuntimeException lastException = null;
153        optima = new UnivariatePointValuePair[starts];
154        totalEvaluations = 0;
155
156        // Multi-start loop.
157        for (int i = 0; i < starts; ++i) {
158            // CHECKSTYLE: stop IllegalCatch
159            try {
160                final double s = (i == 0) ? startValue : min + generator.nextDouble() * (max - min);
161                optima[i] = optimizer.optimize(maxEval - totalEvaluations, f, goal, min, max, s);
162            } catch (RuntimeException mue) {
163                lastException = mue;
164                optima[i] = null;
165            }
166            // CHECKSTYLE: resume IllegalCatch
167
168            totalEvaluations += optimizer.getEvaluations();
169        }
170
171        sortPairs(goal);
172
173        if (optima[0] == null) {
174            throw lastException; // cannot be null if starts >=1
175        }
176
177        // Return the point with the best objective function value.
178        return optima[0];
179    }
180
181    /**
182     * Sort the optima from best to worst, followed by {@code null} elements.
183     *
184     * @param goal Goal type.
185     */
186    private void sortPairs(final GoalType goal) {
187        Arrays.sort(optima, new Comparator<UnivariatePointValuePair>() {
188                public int compare(final UnivariatePointValuePair o1,
189                                   final UnivariatePointValuePair o2) {
190                    if (o1 == null) {
191                        return (o2 == null) ? 0 : 1;
192                    } else if (o2 == null) {
193                        return -1;
194                    }
195                    final double v1 = o1.getValue();
196                    final double v2 = o2.getValue();
197                    return (goal == GoalType.MINIMIZE) ?
198                        Double.compare(v1, v2) : Double.compare(v2, v1);
199                }
200            });
201    }
202}