[7726] | 1 | #region License Information
|
---|
| 2 | /* HeuristicLab
|
---|
[9708] | 3 | * Copyright (C) 2002-2013 Heuristic and Evolutionary Algorithms Laboratory (HEAL)
|
---|
[7726] | 4 | *
|
---|
| 5 | * This file is part of HeuristicLab.
|
---|
| 6 | *
|
---|
| 7 | * HeuristicLab is free software: you can redistribute it and/or modify
|
---|
| 8 | * it under the terms of the GNU General Public License as published by
|
---|
| 9 | * the Free Software Foundation, either version 3 of the License, or
|
---|
| 10 | * (at your option) any later version.
|
---|
| 11 | *
|
---|
| 12 | * HeuristicLab is distributed in the hope that it will be useful,
|
---|
| 13 | * but WITHOUT ANY WARRANTY; without even the implied warranty of
|
---|
| 14 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
---|
| 15 | * GNU General Public License for more details.
|
---|
| 16 | *
|
---|
| 17 | * You should have received a copy of the GNU General Public License
|
---|
| 18 | * along with HeuristicLab. If not, see <http://www.gnu.org/licenses/>.
|
---|
| 19 | */
|
---|
| 20 | #endregion
|
---|
| 21 |
|
---|
[9126] | 22 | using System;
|
---|
[7726] | 23 | using System.Collections.Generic;
|
---|
| 24 | using System.Linq;
|
---|
| 25 | using HeuristicLab.Common;
|
---|
| 26 | using HeuristicLab.Core;
|
---|
| 27 | using HeuristicLab.Data;
|
---|
| 28 | using HeuristicLab.Encodings.SymbolicExpressionTreeEncoding;
|
---|
| 29 | using HeuristicLab.Optimization;
|
---|
| 30 | using HeuristicLab.Parameters;
|
---|
| 31 | using HeuristicLab.Persistence.Default.CompositeSerializers.Storable;
|
---|
| 32 |
|
---|
| 33 | namespace HeuristicLab.Problems.DataAnalysis.Symbolic {
|
---|
| 34 | /// <summary>
|
---|
| 35 | /// An operator that collects the Pareto-best symbolic data analysis solutions for single objective symbolic data analysis problems.
|
---|
| 36 | /// </summary>
|
---|
[7734] | 37 | [Item("SymbolicDataAnalysisSingleObjectiveTrainingParetoBestSolutionAnalyzer", "An operator that analyzes the Pareto-best symbolic data analysis solution for single objective symbolic data analysis problems.")]
|
---|
[7726] | 38 | [StorableClass]
|
---|
[9126] | 39 | public abstract class SymbolicDataAnalysisSingleObjectiveTrainingParetoBestSolutionAnalyzer<S, T> : SymbolicDataAnalysisSingleObjectiveAnalyzer, ISymbolicDataAnalysisInterpreterOperator, ISymbolicDataAnalysisBoundedOperator
|
---|
| 40 | where T : class, ISymbolicDataAnalysisSolution
|
---|
| 41 | where S : class, IDataAnalysisProblemData {
|
---|
| 42 | private const string ProblemDataParameterName = "ProblemData";
|
---|
[7726] | 43 | private const string TrainingBestSolutionsParameterName = "Best training solutions";
|
---|
| 44 | private const string TrainingBestSolutionQualitiesParameterName = "Best training solution qualities";
|
---|
| 45 | private const string ComplexityParameterName = "Complexity";
|
---|
[9126] | 46 | private const string SymbolicDataAnalysisTreeInterpreterParameterName = "SymbolicDataAnalysisTreeInterpreter";
|
---|
| 47 | private const string EstimationLimitsParameterName = "EstimationLimits";
|
---|
[7726] | 48 |
|
---|
[7734] | 49 | public override bool EnabledByDefault {
|
---|
| 50 | get { return false; }
|
---|
| 51 | }
|
---|
| 52 |
|
---|
[7726] | 53 | #region parameter properties
|
---|
| 54 | public ILookupParameter<ItemList<T>> TrainingBestSolutionsParameter {
|
---|
| 55 | get { return (ILookupParameter<ItemList<T>>)Parameters[TrainingBestSolutionsParameterName]; }
|
---|
| 56 | }
|
---|
| 57 | public ILookupParameter<ItemList<DoubleArray>> TrainingBestSolutionQualitiesParameter {
|
---|
| 58 | get { return (ILookupParameter<ItemList<DoubleArray>>)Parameters[TrainingBestSolutionQualitiesParameterName]; }
|
---|
| 59 | }
|
---|
| 60 | public IScopeTreeLookupParameter<DoubleValue> ComplexityParameter {
|
---|
| 61 | get { return (IScopeTreeLookupParameter<DoubleValue>)Parameters[ComplexityParameterName]; }
|
---|
| 62 | }
|
---|
[9126] | 63 | public ILookupParameter<ISymbolicDataAnalysisExpressionTreeInterpreter> SymbolicDataAnalysisTreeInterpreterParameter {
|
---|
| 64 | get { return (ILookupParameter<ISymbolicDataAnalysisExpressionTreeInterpreter>)Parameters[SymbolicDataAnalysisTreeInterpreterParameterName]; }
|
---|
| 65 | }
|
---|
| 66 | public ILookupParameter<S> ProblemDataParameter {
|
---|
| 67 | get { return (ILookupParameter<S>)Parameters[ProblemDataParameterName]; }
|
---|
| 68 | }
|
---|
| 69 | public IValueLookupParameter<DoubleLimit> EstimationLimitsParameter {
|
---|
| 70 | get { return (IValueLookupParameter<DoubleLimit>)Parameters[EstimationLimitsParameterName]; }
|
---|
| 71 | }
|
---|
[7726] | 72 | #endregion
|
---|
| 73 | #region properties
|
---|
| 74 | public ItemList<T> TrainingBestSolutions {
|
---|
| 75 | get { return TrainingBestSolutionsParameter.ActualValue; }
|
---|
| 76 | set { TrainingBestSolutionsParameter.ActualValue = value; }
|
---|
| 77 | }
|
---|
| 78 | public ItemList<DoubleArray> TrainingBestSolutionQualities {
|
---|
| 79 | get { return TrainingBestSolutionQualitiesParameter.ActualValue; }
|
---|
| 80 | set { TrainingBestSolutionQualitiesParameter.ActualValue = value; }
|
---|
| 81 | }
|
---|
| 82 | #endregion
|
---|
| 83 |
|
---|
| 84 | [StorableConstructor]
|
---|
| 85 | protected SymbolicDataAnalysisSingleObjectiveTrainingParetoBestSolutionAnalyzer(bool deserializing) : base(deserializing) { }
|
---|
[9126] | 86 | protected SymbolicDataAnalysisSingleObjectiveTrainingParetoBestSolutionAnalyzer(SymbolicDataAnalysisSingleObjectiveTrainingParetoBestSolutionAnalyzer<S, T> original, Cloner cloner) : base(original, cloner) { }
|
---|
[7726] | 87 | public SymbolicDataAnalysisSingleObjectiveTrainingParetoBestSolutionAnalyzer()
|
---|
| 88 | : base() {
|
---|
[9126] | 89 | Parameters.Add(new LookupParameter<S>(ProblemDataParameterName, "The problem data for the symbolic data analysis solution."));
|
---|
[7726] | 90 | Parameters.Add(new LookupParameter<ItemList<T>>(TrainingBestSolutionsParameterName, "The training best (Pareto-optimal) symbolic data analysis solutions."));
|
---|
| 91 | Parameters.Add(new LookupParameter<ItemList<DoubleArray>>(TrainingBestSolutionQualitiesParameterName, "The qualities of the training best (Pareto-optimal) solutions."));
|
---|
| 92 | Parameters.Add(new ScopeTreeLookupParameter<DoubleValue>(ComplexityParameterName, "The complexity of each tree."));
|
---|
[9126] | 93 | Parameters.Add(new LookupParameter<ISymbolicDataAnalysisExpressionTreeInterpreter>(SymbolicDataAnalysisTreeInterpreterParameterName, "The symbolic data analysis tree interpreter for the symbolic expression tree."));
|
---|
| 94 | Parameters.Add(new ValueLookupParameter<DoubleLimit>(EstimationLimitsParameterName, "The lower and upper limit for the estimated values produced by the symbolic classification model."));
|
---|
[7726] | 95 | }
|
---|
| 96 |
|
---|
| 97 | public override IOperation Apply() {
|
---|
| 98 | var results = ResultCollection;
|
---|
| 99 | // create empty parameter and result values
|
---|
| 100 | if (TrainingBestSolutions == null) {
|
---|
| 101 | TrainingBestSolutions = new ItemList<T>();
|
---|
| 102 | TrainingBestSolutionQualities = new ItemList<DoubleArray>();
|
---|
| 103 | results.Add(new Result(TrainingBestSolutionQualitiesParameter.Name, TrainingBestSolutionQualitiesParameter.Description, TrainingBestSolutionQualities));
|
---|
| 104 | results.Add(new Result(TrainingBestSolutionsParameter.Name, TrainingBestSolutionsParameter.Description, TrainingBestSolutions));
|
---|
| 105 | }
|
---|
| 106 |
|
---|
| 107 | IList<Tuple<double, double>> trainingBestQualities = TrainingBestSolutionQualities
|
---|
| 108 | .Select(x => Tuple.Create(x[0], x[1]))
|
---|
| 109 | .ToList();
|
---|
| 110 |
|
---|
| 111 | #region find best trees
|
---|
| 112 | IList<int> nonDominatedIndexes = new List<int>();
|
---|
| 113 | ISymbolicExpressionTree[] tree = SymbolicExpressionTree.ToArray();
|
---|
| 114 | List<double> qualities = Quality.Select(x => x.Value).ToList();
|
---|
| 115 |
|
---|
| 116 | List<double> complexities;
|
---|
[9126] | 117 | if (ComplexityParameter.ActualValue != null && ComplexityParameter.ActualValue.Length == qualities.Count) {
|
---|
[7726] | 118 | complexities = ComplexityParameter.ActualValue.Select(x => x.Value).ToList();
|
---|
| 119 | } else {
|
---|
| 120 | complexities = tree.Select(t => (double)t.Length).ToList();
|
---|
| 121 | }
|
---|
| 122 | List<Tuple<double, double>> fitness = new List<Tuple<double, double>>();
|
---|
| 123 | for (int i = 0; i < qualities.Count; i++)
|
---|
| 124 | fitness.Add(Tuple.Create(qualities[i], complexities[i]));
|
---|
| 125 | var maximization = Tuple.Create(Maximization.Value, false);// complexity must be minimized
|
---|
| 126 | List<Tuple<double, double>> newNonDominatedQualities = new List<Tuple<double, double>>();
|
---|
| 127 | for (int i = 0; i < tree.Length; i++) {
|
---|
| 128 | if (IsNonDominated(fitness[i], trainingBestQualities, maximization) &&
|
---|
| 129 | IsNonDominated(fitness[i], newNonDominatedQualities, maximization) &&
|
---|
| 130 | IsNonDominated(fitness[i], fitness.Skip(i + 1), maximization)) {
|
---|
| 131 | if (!newNonDominatedQualities.Contains(fitness[i])) {
|
---|
| 132 | newNonDominatedQualities.Add(fitness[i]);
|
---|
| 133 | nonDominatedIndexes.Add(i);
|
---|
| 134 | }
|
---|
| 135 | }
|
---|
| 136 | }
|
---|
| 137 | #endregion
|
---|
| 138 |
|
---|
| 139 | #region update Pareto-optimal solution archive
|
---|
| 140 | if (nonDominatedIndexes.Count > 0) {
|
---|
| 141 | ItemList<DoubleArray> nonDominatedQualities = new ItemList<DoubleArray>();
|
---|
| 142 | ItemList<T> nonDominatedSolutions = new ItemList<T>();
|
---|
| 143 | // add all new non-dominated solutions to the archive
|
---|
| 144 | foreach (var index in nonDominatedIndexes) {
|
---|
| 145 | T solution = CreateSolution(tree[index]);
|
---|
| 146 | nonDominatedSolutions.Add(solution);
|
---|
| 147 | nonDominatedQualities.Add(new DoubleArray(new double[] { fitness[index].Item1, fitness[index].Item2 }));
|
---|
| 148 | }
|
---|
| 149 | // add old non-dominated solutions only if they are not dominated by one of the new solutions
|
---|
| 150 | for (int i = 0; i < trainingBestQualities.Count; i++) {
|
---|
| 151 | if (IsNonDominated(trainingBestQualities[i], newNonDominatedQualities, maximization)) {
|
---|
| 152 | if (!newNonDominatedQualities.Contains(trainingBestQualities[i])) {
|
---|
| 153 | nonDominatedSolutions.Add(TrainingBestSolutions[i]);
|
---|
| 154 | nonDominatedQualities.Add(TrainingBestSolutionQualities[i]);
|
---|
| 155 | }
|
---|
| 156 | }
|
---|
| 157 | }
|
---|
| 158 |
|
---|
| 159 | // make sure solutions and qualities are ordered in the results
|
---|
| 160 | var orderedIndexes =
|
---|
| 161 | nonDominatedSolutions.Select((s, i) => i).OrderBy(i => nonDominatedQualities[i][0]).ToArray();
|
---|
| 162 |
|
---|
| 163 | var orderedNonDominatedSolutions = new ItemList<T>();
|
---|
| 164 | var orderedNonDominatedQualities = new ItemList<DoubleArray>();
|
---|
| 165 | foreach (var i in orderedIndexes) {
|
---|
| 166 | orderedNonDominatedQualities.Add(nonDominatedQualities[i]);
|
---|
| 167 | orderedNonDominatedSolutions.Add(nonDominatedSolutions[i]);
|
---|
| 168 | }
|
---|
| 169 |
|
---|
| 170 | TrainingBestSolutions = orderedNonDominatedSolutions;
|
---|
| 171 | TrainingBestSolutionQualities = orderedNonDominatedQualities;
|
---|
| 172 |
|
---|
| 173 | results[TrainingBestSolutionsParameter.Name].Value = orderedNonDominatedSolutions;
|
---|
| 174 | results[TrainingBestSolutionQualitiesParameter.Name].Value = orderedNonDominatedQualities;
|
---|
| 175 | }
|
---|
| 176 | #endregion
|
---|
| 177 | return base.Apply();
|
---|
| 178 | }
|
---|
| 179 |
|
---|
| 180 | protected abstract T CreateSolution(ISymbolicExpressionTree bestTree);
|
---|
| 181 |
|
---|
| 182 | private bool IsNonDominated(Tuple<double, double> point, IEnumerable<Tuple<double, double>> points, Tuple<bool, bool> maximization) {
|
---|
| 183 | return !points.Any(p => IsBetterOrEqual(p.Item1, point.Item1, maximization.Item1) &&
|
---|
| 184 | IsBetterOrEqual(p.Item2, point.Item2, maximization.Item2));
|
---|
| 185 | }
|
---|
| 186 | private bool IsBetterOrEqual(double lhs, double rhs, bool maximization) {
|
---|
| 187 | if (maximization) return lhs >= rhs;
|
---|
| 188 | else return lhs <= rhs;
|
---|
| 189 | }
|
---|
| 190 | }
|
---|
| 191 | }
|
---|