[14893] | 1 | #region License Information
|
---|
| 2 | /* HeuristicLab
|
---|
| 3 | * Copyright (C) 2002-2016 Heuristic and Evolutionary Algorithms Laboratory (HEAL)
|
---|
| 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 |
|
---|
| 22 | using System;
|
---|
| 23 | using System.Collections.Generic;
|
---|
| 24 | using System.Linq;
|
---|
| 25 | using System.Threading;
|
---|
| 26 | using HeuristicLab.Analysis;
|
---|
| 27 | using HeuristicLab.Common;
|
---|
| 28 | using HeuristicLab.Core;
|
---|
| 29 | using HeuristicLab.Data;
|
---|
| 30 | using HeuristicLab.Encodings.RealVectorEncoding;
|
---|
| 31 | using HeuristicLab.Optimization;
|
---|
| 32 | using HeuristicLab.Persistence.Default.CompositeSerializers.Storable;
|
---|
| 33 | using HeuristicLab.Problems.DataAnalysis;
|
---|
| 34 |
|
---|
| 35 | namespace HeuristicLab.Algorithms.SAPBA.Strategies {
|
---|
| 36 | [StorableClass]
|
---|
| 37 | public abstract class StrategyBase : ParameterizedNamedItem, ISurrogateStrategy {
|
---|
| 38 | #region Properties
|
---|
| 39 | [Storable]
|
---|
[14894] | 40 | protected SurrogateAssistedPopulationBasedAlgorithm Algorithm;
|
---|
[14893] | 41 | [Storable]
|
---|
| 42 | private List<Tuple<RealVector, double>> Samples;
|
---|
| 43 | [Storable]
|
---|
| 44 | protected IRegressionSolution RegressionSolution;
|
---|
| 45 | protected CancellationToken Cancellation;
|
---|
| 46 | private IEnumerable<Tuple<RealVector, double>> TruncatedSamples => Samples.Count > Algorithm.MaximalDatasetSize && Algorithm.MaximalDatasetSize > 0 ? Samples.Skip(Samples.Count - Algorithm.MaximalDatasetSize) : Samples;
|
---|
| 47 | #endregion
|
---|
| 48 |
|
---|
| 49 | #region ResultName
|
---|
| 50 | private const string BestQualityResultName = "Best Quality";
|
---|
| 51 | private const string BestSolutionResultName = "Best Solution";
|
---|
| 52 | private const string QualityTableResultName = "Qualities";
|
---|
| 53 | private const string BestQualityRowName = "Best Quality";
|
---|
| 54 | private const string WorstQualityRowName = "Worst Quality";
|
---|
| 55 | private const string CurrentQualityRowName = "Current Quality";
|
---|
| 56 | private const string MedianQualityRowName = "Median Quality";
|
---|
| 57 | private const string AverageQualityRowName = "Average Quality";
|
---|
| 58 | private const string RegressionSolutionResultName = "Model";
|
---|
| 59 | private const string EvaluatedSoultionsResultName = "EvaluatedSolutions";
|
---|
| 60 | private const string IterationsResultName = "Iterations";
|
---|
| 61 | #endregion
|
---|
| 62 |
|
---|
| 63 | #region constructors
|
---|
| 64 | [StorableConstructor]
|
---|
| 65 | protected StrategyBase(bool deserializing) : base(deserializing) { }
|
---|
| 66 | protected StrategyBase(StrategyBase original, Cloner cloner) : base(original, cloner) {
|
---|
| 67 | if (original.Samples != null) Samples = original.Samples.Select(x => new Tuple<RealVector, double>(cloner.Clone(x.Item1), x.Item2)).ToList();
|
---|
| 68 | RegressionSolution = cloner.Clone(original.RegressionSolution);
|
---|
| 69 | }
|
---|
| 70 | protected StrategyBase() { }
|
---|
| 71 | #endregion
|
---|
| 72 |
|
---|
| 73 | public abstract double Evaluate(RealVector r, IRandom random);
|
---|
| 74 | protected abstract void Analyze(Individual[] individuals, double[] qualities, ResultCollection results, ResultCollection globalResults, IRandom random);
|
---|
| 75 | protected abstract void ProcessPopulation(Individual[] individuals, double[] qualities, IRandom random);
|
---|
| 76 | protected abstract void Initialize();
|
---|
| 77 |
|
---|
| 78 | public void Analyze(Individual[] individuals, double[] qualities, ResultCollection results, IRandom random) {
|
---|
| 79 | Algorithm.Problem.Analyze(individuals, qualities, results, random);
|
---|
| 80 | ProcessPopulation(individuals, qualities, random);
|
---|
| 81 |
|
---|
| 82 | var globalResults = Algorithm.Results;
|
---|
| 83 | if (!globalResults.ContainsKey(EvaluatedSoultionsResultName)) globalResults.Add(new Result(EvaluatedSoultionsResultName, new IntValue(Samples.Count)));
|
---|
| 84 | else ((IntValue)globalResults[EvaluatedSoultionsResultName].Value).Value = Samples.Count;
|
---|
| 85 | if (!globalResults.ContainsKey(IterationsResultName)) globalResults.Add(new Result(IterationsResultName, new IntValue(0)));
|
---|
| 86 | else ((IntValue)globalResults[IterationsResultName].Value).Value++;
|
---|
| 87 |
|
---|
| 88 | if (Samples.Count != 0) {
|
---|
| 89 | var min = Samples.Min(x => x.Item2);
|
---|
| 90 | var max = Samples.Max(x => x.Item2);
|
---|
| 91 | var bestIdx = Algorithm.Problem.Maximization ? Samples.ArgMax(x => x.Item2) : Samples.ArgMin(x => x.Item2);
|
---|
| 92 |
|
---|
| 93 | if (!globalResults.ContainsKey(BestQualityResultName)) globalResults.Add(new Result(BestQualityResultName, new DoubleValue(0.0)));
|
---|
[14894] | 94 | ((DoubleValue)globalResults[BestQualityResultName].Value).Value = Samples[bestIdx].Item2;
|
---|
| 95 | if (!globalResults.ContainsKey(BestSolutionResultName)) globalResults.Add(new Result(BestSolutionResultName, new RealVector()));
|
---|
| 96 | globalResults[BestSolutionResultName].Value = Samples[bestIdx].Item1;
|
---|
[14893] | 97 |
|
---|
| 98 | DataTable table;
|
---|
| 99 | if (!globalResults.ContainsKey(QualityTableResultName)) {
|
---|
| 100 | table = new DataTable("Qualites", "Qualites over iteration");
|
---|
| 101 | globalResults.Add(new Result(QualityTableResultName, table));
|
---|
| 102 | table.Rows.Add(new DataRow(BestQualityRowName, "Best Quality"));
|
---|
| 103 | table.Rows.Add(new DataRow(WorstQualityRowName, "Worst Quality"));
|
---|
| 104 | table.Rows.Add(new DataRow(CurrentQualityRowName, "Current Quality"));
|
---|
| 105 | table.Rows.Add(new DataRow(MedianQualityRowName, "Median Quality"));
|
---|
| 106 | table.Rows.Add(new DataRow(AverageQualityRowName, "Average Quality"));
|
---|
| 107 | }
|
---|
| 108 | table = (DataTable)globalResults[QualityTableResultName].Value;
|
---|
| 109 | table.Rows[BestQualityResultName].Values.Add(Algorithm.Problem.Maximization ? max : min);
|
---|
| 110 | table.Rows[WorstQualityRowName].Values.Add(Algorithm.Problem.Maximization ? min : max);
|
---|
| 111 | table.Rows[CurrentQualityRowName].Values.Add(Samples[Samples.Count - 1].Item2);
|
---|
| 112 | table.Rows[AverageQualityRowName].Values.Add(Samples.Average(x => x.Item2));
|
---|
| 113 | table.Rows[MedianQualityRowName].Values.Add(Samples.Select(x => x.Item2).Median());
|
---|
| 114 | }
|
---|
| 115 |
|
---|
| 116 | if (RegressionSolution != null) {
|
---|
| 117 | if (!globalResults.ContainsKey(RegressionSolutionResultName))
|
---|
| 118 | globalResults.Add(new Result(RegressionSolutionResultName, RegressionSolution));
|
---|
| 119 | else
|
---|
| 120 | globalResults[RegressionSolutionResultName].Value = RegressionSolution;
|
---|
| 121 | }
|
---|
| 122 |
|
---|
| 123 | Analyze(individuals, qualities, results, globalResults, random);
|
---|
| 124 | }
|
---|
| 125 | public void Initialize(SurrogateAssistedPopulationBasedAlgorithm algorithm) {
|
---|
| 126 | Algorithm = algorithm;
|
---|
| 127 | Samples = algorithm.InitialSamples?.ToList() ?? new List<Tuple<RealVector, double>>();
|
---|
| 128 | RegressionSolution = null;
|
---|
| 129 | Initialize();
|
---|
| 130 | }
|
---|
| 131 |
|
---|
| 132 | #region Helpers for Subclasses
|
---|
| 133 | protected void BuildRegressionSolution(IRandom random) {
|
---|
| 134 | RegressionSolution = EgoUtilities.BuildModel(Cancellation, TruncatedSamples, Algorithm.RegressionAlgorithm, random, Algorithm.RemoveDuplicates, RegressionSolution);
|
---|
| 135 | }
|
---|
| 136 | protected Tuple<RealVector, double> EvaluateSample(RealVector point, IRandom random) {
|
---|
[14894] | 137 | Cancellation.ThrowIfCancellationRequested();
|
---|
| 138 | if (Samples.Count >= Algorithm.MaximumEvaluations) { Algorithm.OptimizationAlgorithm.Stop(); return new Tuple<RealVector, double>(point, 0.0); }
|
---|
[14893] | 139 | var p = new Tuple<RealVector, double>(point, Algorithm.Problem.Evaluate(GetIndividual(point), random));
|
---|
| 140 | Samples.Add(p);
|
---|
| 141 | return p;
|
---|
| 142 | }
|
---|
| 143 | protected Tuple<RealVector, double> EstimateSample(RealVector point, IRandom random) {
|
---|
[14894] | 144 | if (Samples.Count == Algorithm.InitialEvaluations && RegressionSolution == null) BuildRegressionSolution(random);
|
---|
[14893] | 145 | return Samples.Count < Algorithm.InitialEvaluations ? EvaluateSample(point, random) : new Tuple<RealVector, double>(point, RegressionSolution.Model.GetEstimation(point));
|
---|
| 146 | }
|
---|
| 147 | #endregion
|
---|
| 148 |
|
---|
| 149 | #region Helpers
|
---|
| 150 | private Individual GetIndividual(RealVector r) {
|
---|
| 151 | var scope = new Scope();
|
---|
| 152 | scope.Variables.Add(new Variable(Algorithm.Problem.Encoding.Name, r));
|
---|
| 153 | return new SingleEncodingIndividual(Algorithm.Problem.Encoding, scope);
|
---|
| 154 | }
|
---|
| 155 |
|
---|
| 156 | public void UpdateCancellation(CancellationToken cancellationToken) {
|
---|
| 157 | Cancellation = cancellationToken;
|
---|
| 158 | }
|
---|
| 159 | #endregion
|
---|
| 160 | }
|
---|
| 161 | } |
---|