Free cookie consent management tool by TermsFeed Policy Generator

source: trunk/HeuristicLab.Algorithms.MOCMAEvolutionStrategy/3.3/Individual.cs @ 15769

Last change on this file since 15769 was 15583, checked in by swagner, 7 years ago

#2640: Updated year of copyrights in license headers

File size: 7.9 KB
Line 
1#region License Information
2/* HeuristicLab
3 * Copyright (C) 2002-2018 Heuristic and Evolutionary Algorithms Laboratory (HEAL)
4 * and the BEACON Center for the Study of Evolution in Action.
5 *
6 * This file is part of HeuristicLab.
7 *
8 * HeuristicLab is free software: you can redistribute it and/or modify
9 * it under the terms of the GNU General Public License as published by
10 * the Free Software Foundation, either version 3 of the License, or
11 * (at your option) any later version.
12 *
13 * HeuristicLab is distributed in the hope that it will be useful,
14 * but WITHOUT ANY WARRANTY; without even the implied warranty of
15 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
16 * GNU General Public License for more details.
17 *
18 * You should have received a copy of the GNU General Public License
19 * along with HeuristicLab. If not, see <http://www.gnu.org/licenses/>.
20 */
21#endregion
22
23using System;
24using System.Linq;
25using HeuristicLab.Common;
26using HeuristicLab.Encodings.RealVectorEncoding;
27using HeuristicLab.Persistence.Default.CompositeSerializers.Storable;
28using HeuristicLab.Random;
29
30namespace HeuristicLab.Algorithms.MOCMAEvolutionStrategy {
31  [StorableClass]
32  public class Individual : IDeepCloneable {
33
34    #region Properties
35    [Storable]
36    private MOCMAEvolutionStrategy strategy;
37
38    //Chromosome
39    [Storable]
40    public RealVector Mean { get; private set; }
41    [Storable]
42    private double sigma;//stepsize
43    [Storable]
44    private RealVector evolutionPath; // pc
45    [Storable]
46    private RealVector lastStep;
47    [Storable]
48    private RealVector lastZ;
49    [Storable]
50    private double[,] lowerCholesky;
51
52    //Phenotype
53    [Storable]
54    public double[] Fitness { get; set; }
55    [Storable]
56    public double[] PenalizedFitness { get; set; }
57    [Storable]
58    public bool Selected { get; set; }
59    [Storable]
60    public double SuccessProbability { get; set; }
61    #endregion
62
63    #region Constructors and Cloning
64    [StorableConstructor]
65    protected Individual(bool deserializing) { }
66
67    public Individual(RealVector mean, double pSucc, double sigma, RealVector pc, double[,] c, MOCMAEvolutionStrategy strategy) {
68      Mean = mean;
69      lastStep = new RealVector(mean.Length);
70      SuccessProbability = pSucc;
71      this.sigma = sigma;
72      evolutionPath = pc;
73      CholeskyDecomposition(c);
74      Selected = true;
75      this.strategy = strategy;
76    }
77
78    public Individual(Individual other) {
79      SuccessProbability = other.SuccessProbability;
80      sigma = other.sigma;
81      evolutionPath = (RealVector)other.evolutionPath.Clone();
82      Mean = (RealVector)other.Mean.Clone();
83      lowerCholesky = (double[,])other.lowerCholesky.Clone();
84      Selected = true;
85      strategy = other.strategy;
86    }
87
88    public Individual(Individual other, Cloner cloner) {
89      strategy = cloner.Clone(other.strategy);
90      Mean = cloner.Clone(other.Mean);
91      sigma = other.sigma;
92      evolutionPath = cloner.Clone(other.evolutionPath);
93      lastStep = cloner.Clone(other.evolutionPath);
94      lastZ = cloner.Clone(other.lastZ);
95      lowerCholesky = other.lowerCholesky != null ? other.lowerCholesky.Clone() as double[,] : null;
96      Fitness = other.Fitness != null ? other.Fitness.Select(x => x).ToArray() : null;
97      PenalizedFitness = other.PenalizedFitness != null ? other.PenalizedFitness.Select(x => x).ToArray() : null;
98      Selected = other.Selected;
99      SuccessProbability = other.SuccessProbability;
100    }
101
102    public object Clone() {
103      return new Cloner().Clone(this);
104    }
105
106    public IDeepCloneable Clone(Cloner cloner) {
107      return new Individual(this, cloner);
108    }
109    #endregion
110
111    public void Mutate(NormalDistributedRandom gauss) {
112      //sampling a random z from N(0,I) where I is the Identity matrix;
113      lastZ = new RealVector(Mean.Length);
114      var n = lastZ.Length;
115      for (var i = 0; i < n; i++) lastZ[i] = gauss.NextDouble();
116      //Matrixmultiplication: lastStep = lowerCholesky * lastZ;
117      lastStep = new RealVector(Mean.Length);
118      for (var i = 0; i < n; i++) {
119        double sum = 0;
120        for (var j = 0; j <= i; j++) sum += lowerCholesky[i, j] * lastZ[j];
121        lastStep[i] = sum;
122      }
123      //add the step to x weighted by stepsize;
124      for (var i = 0; i < Mean.Length; i++) Mean[i] += sigma * lastStep[i];
125    }
126
127    public void UpdateAsParent(bool offspringSuccessful) {
128      SuccessProbability = (1 - strategy.StepSizeLearningRate) * SuccessProbability + strategy.StepSizeLearningRate * (offspringSuccessful ? 1 : 0);
129      sigma *= Math.Exp(1 / strategy.StepSizeDampeningFactor * (SuccessProbability - strategy.TargetSuccessProbability) / (1 - strategy.TargetSuccessProbability));
130      if (!offspringSuccessful) return;
131      if (SuccessProbability < strategy.SuccessThreshold && lastZ != null) {
132        var stepNormSqr = lastZ.Sum(d => d * d);
133        var rate = strategy.CovarianceMatrixUnlearningRate;
134        if (stepNormSqr > 1 && 1 < strategy.CovarianceMatrixUnlearningRate * (2 * stepNormSqr - 1)) rate = 1 / (2 * stepNormSqr - 1);
135        CholeskyUpdate(lastStep, 1 + rate, -rate);
136      } else RoundUpdate();
137    }
138
139    public void UpdateAsOffspring() {
140      SuccessProbability = (1 - strategy.StepSizeLearningRate) * SuccessProbability + strategy.StepSizeLearningRate;
141      sigma *= Math.Exp(1 / strategy.StepSizeDampeningFactor * (SuccessProbability - strategy.TargetSuccessProbability) / (1 - strategy.TargetSuccessProbability));
142      var evolutionpathUpdateWeight = strategy.EvolutionPathLearningRate * (2.0 - strategy.EvolutionPathLearningRate);
143      if (SuccessProbability < strategy.SuccessThreshold) {
144        UpdateEvolutionPath(1 - strategy.EvolutionPathLearningRate, evolutionpathUpdateWeight);
145        CholeskyUpdate(evolutionPath, 1 - strategy.CovarianceMatrixLearningRate, strategy.CovarianceMatrixLearningRate);
146      } else RoundUpdate();
147    }
148
149    public void UpdateEvolutionPath(double learningRate, double updateWeight) {
150      updateWeight = Math.Sqrt(updateWeight);
151      for (var i = 0; i < evolutionPath.Length; i++) {
152        evolutionPath[i] *= learningRate;
153        evolutionPath[i] += updateWeight * lastStep[i];
154      }
155    }
156
157    #region Helpers
158    private void CholeskyDecomposition(double[,] c) {
159      if (!alglib.spdmatrixcholesky(ref c, c.GetLength(0), false))
160        throw new ArgumentException("Covariancematrix is not symmetric positiv definit");
161      lowerCholesky = (double[,])c.Clone();
162    }
163
164    private void CholeskyUpdate(RealVector v, double alpha, double beta) {
165      var n = v.Length;
166      var temp = new double[n];
167      for (var i = 0; i < n; i++) temp[i] = v[i];
168      double betaPrime = 1;
169      var a = Math.Sqrt(alpha);
170      for (var j = 0; j < n; j++) {
171        var ljj = a * lowerCholesky[j, j];
172        var dj = ljj * ljj;
173        var wj = temp[j];
174        var swj2 = beta * wj * wj;
175        var gamma = dj * betaPrime + swj2;
176        var x1 = dj + swj2 / betaPrime;
177        if (x1 < 0.0) return;
178        var nLjj = Math.Sqrt(x1);
179        lowerCholesky[j, j] = nLjj;
180        betaPrime += swj2 / dj;
181        if (j + 1 >= n) continue;
182        for (var i = j + 1; i < n; i++) lowerCholesky[i, j] *= a;
183        for (var i = j + 1; i < n; i++) temp[i] = wj / ljj * lowerCholesky[i, j];
184        if (gamma.IsAlmost(0)) continue;
185        for (var i = j + 1; i < n; i++) lowerCholesky[i, j] *= nLjj / ljj;
186        for (var i = j + 1; i < n; i++) lowerCholesky[i, j] += nLjj * beta * wj / gamma * temp[i];
187      }
188    }
189
190    private void RoundUpdate() {
191      var evolutionPathUpdateWeight = strategy.EvolutionPathLearningRate * (2.0 - strategy.EvolutionPathLearningRate);
192      UpdateEvolutionPath(1 - strategy.EvolutionPathLearningRate, 0);
193      CholeskyUpdate(evolutionPath, 1 - strategy.CovarianceMatrixLearningRate + evolutionPathUpdateWeight, strategy.CovarianceMatrixLearningRate);
194    }
195    #endregion
196  }
197}
Note: See TracBrowser for help on using the repository browser.