1 | #region License Information
|
---|
2 | /* HeuristicLab
|
---|
3 | * Copyright (C) 2002-2011 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 HeuristicLab.Persistence.Default.CompositeSerializers.Storable;
|
---|
26 | using HeuristicLab.Common;
|
---|
27 | using HeuristicLab.Core;
|
---|
28 | using HeuristicLab.Encodings.SymbolicExpressionTreeEncoding;
|
---|
29 |
|
---|
30 | namespace HeuristicLab.Problems.DataAnalysis.Symbolic {
|
---|
31 |
|
---|
32 | [Item("ProbabilisticFunctionalCrossover", "An operator which performs subtree swapping based on the behavioral similarity between subtrees.")]
|
---|
33 | public sealed class SymbolicDataAnalysisExpressionProbabilisticFunctionalCrossover<T> : SymbolicDataAnalysisExpressionCrossover<T> where T : class, IDataAnalysisProblemData {
|
---|
34 | [StorableConstructor]
|
---|
35 | private SymbolicDataAnalysisExpressionProbabilisticFunctionalCrossover(bool deserializing) : base(deserializing) { }
|
---|
36 | private SymbolicDataAnalysisExpressionProbabilisticFunctionalCrossover(SymbolicDataAnalysisExpressionCrossover<T> original, Cloner cloner)
|
---|
37 | : base(original, cloner) { }
|
---|
38 | public SymbolicDataAnalysisExpressionProbabilisticFunctionalCrossover()
|
---|
39 | : base() {
|
---|
40 | Name = "ProbabilisticFunctionalCrossover";
|
---|
41 | }
|
---|
42 | public override IDeepCloneable Clone(Cloner cloner) { return new SymbolicDataAnalysisExpressionProbabilisticFunctionalCrossover<T>(this, cloner); }
|
---|
43 |
|
---|
44 | protected override ISymbolicExpressionTree Cross(IRandom random, ISymbolicExpressionTree parent0, ISymbolicExpressionTree parent1) {
|
---|
45 | ISymbolicDataAnalysisExpressionTreeInterpreter interpreter = SymbolicDataAnalysisTreeInterpreterParameter.ActualValue;
|
---|
46 | List<int> rows = GenerateRowsToEvaluate().ToList();
|
---|
47 | T problemData = ProblemDataParameter.ActualValue;
|
---|
48 | return Cross(random, parent0, parent1, interpreter, problemData,
|
---|
49 | rows, MaximumSymbolicExpressionTreeDepth.Value, MaximumSymbolicExpressionTreeLength.Value);
|
---|
50 | }
|
---|
51 |
|
---|
52 | public override ISymbolicExpressionTree Crossover(IRandom random, ISymbolicExpressionTree parent0, ISymbolicExpressionTree parent1) {
|
---|
53 | return Cross(random, parent0, parent1);
|
---|
54 | }
|
---|
55 |
|
---|
56 | /// <summary>
|
---|
57 | /// Takes two parent individuals P0 and P1.
|
---|
58 | /// Randomly choose a node i from the first parent, then for each matching node j from the second parent, calculate the behavioral distance based on the range:
|
---|
59 | /// d(i,j) = 0.5 * ( abs(max(i) - max(j)) + abs(min(i) - min(j)) ).
|
---|
60 | /// Next, assign probabilities for the selection of the second cross point based on the inversed and normalized behavioral distance and
|
---|
61 | /// choose the second crosspoint via a random weighted selection procedure.
|
---|
62 | /// </summary>
|
---|
63 | public static ISymbolicExpressionTree Cross(IRandom random, ISymbolicExpressionTree parent0, ISymbolicExpressionTree parent1,
|
---|
64 | ISymbolicDataAnalysisExpressionTreeInterpreter interpreter, T problemData, IList<int> rows, int maxDepth, int maxLength) {
|
---|
65 | var crossoverPoints0 = new List<CutPoint>();
|
---|
66 | parent0.Root.ForEachNodePostfix((n) => {
|
---|
67 | if (n.Subtrees.Any() && n != parent0.Root)
|
---|
68 | foreach (var child in n.Subtrees)
|
---|
69 | crossoverPoints0.Add(new CutPoint(n, child));
|
---|
70 | });
|
---|
71 | var crossoverPoint0 = crossoverPoints0.SelectRandom(random);
|
---|
72 | int level = parent0.Root.GetBranchLevel(crossoverPoint0.Child);
|
---|
73 | int length = parent0.Root.GetLength() - crossoverPoint0.Child.GetLength();
|
---|
74 |
|
---|
75 | var allowedBranches = new List<ISymbolicExpressionTreeNode>();
|
---|
76 | parent1.Root.ForEachNodePostfix((n) => {
|
---|
77 | if (n.Subtrees.Any() && n != parent1.Root)
|
---|
78 | allowedBranches.AddRange(n.Subtrees.Where(s => crossoverPoint0.IsMatchingPointType(s) && s.GetDepth() + level <= maxDepth && s.GetLength() + length <= maxLength));
|
---|
79 | });
|
---|
80 |
|
---|
81 | if (allowedBranches.Count == 0)
|
---|
82 | return parent0;
|
---|
83 |
|
---|
84 | var dataset = problemData.Dataset;
|
---|
85 |
|
---|
86 | // create symbols in order to improvize an ad-hoc tree so that the child can be evaluated
|
---|
87 | var rootSymbol = new ProgramRootSymbol();
|
---|
88 | var startSymbol = new StartSymbol();
|
---|
89 | var tree0 = CreateTreeFromNode(random, crossoverPoint0.Child, rootSymbol, startSymbol); // this will change crossoverPoint0.Child.Parent
|
---|
90 | List<double> estimatedValues0 = interpreter.GetSymbolicExpressionTreeValues(tree0, dataset, rows).ToList();
|
---|
91 | double min0 = estimatedValues0.Min();
|
---|
92 | double max0 = estimatedValues0.Max();
|
---|
93 | crossoverPoint0.Child.Parent = crossoverPoint0.Parent; // restore correct parent
|
---|
94 |
|
---|
95 | var weights = new List<double>();
|
---|
96 | foreach (var node in allowedBranches) {
|
---|
97 | var parent = node.Parent;
|
---|
98 | var tree1 = CreateTreeFromNode(random, node, rootSymbol, startSymbol);
|
---|
99 | List<double> estimatedValues1 = interpreter.GetSymbolicExpressionTreeValues(tree1, dataset, rows).ToList();
|
---|
100 | double min1 = estimatedValues1.Min();
|
---|
101 | double max1 = estimatedValues1.Max();
|
---|
102 | double behavioralDistance = (Math.Abs(min0 - min1) + Math.Abs(max0 - max1)) / 2; // this can be NaN of Infinity because some trees are crazy like exp(exp(exp(...))), we correct that below
|
---|
103 | weights.Add(behavioralDistance);
|
---|
104 | node.Parent = parent; // restore correct node parent
|
---|
105 | }
|
---|
106 |
|
---|
107 | // remove branches with an infinite or NaN behavioral distance
|
---|
108 | int count = weights.Count, idx = 0;
|
---|
109 | while (idx < count) {
|
---|
110 | if (Double.IsNaN(weights[idx]) || Double.IsInfinity(weights[idx])) {
|
---|
111 | weights.RemoveAt(idx);
|
---|
112 | allowedBranches.RemoveAt(idx);
|
---|
113 | --count;
|
---|
114 | } else {
|
---|
115 | ++idx;
|
---|
116 | }
|
---|
117 | }
|
---|
118 |
|
---|
119 | // check if there are any allowed branches left
|
---|
120 | if (allowedBranches.Count == 0)
|
---|
121 | return parent0;
|
---|
122 |
|
---|
123 | ISymbolicExpressionTreeNode selectedBranch;
|
---|
124 | double sum = weights.Sum();
|
---|
125 | if (sum == 0.0)
|
---|
126 | selectedBranch = allowedBranches[0]; // just return the first, since we don't care (all weights are zero)
|
---|
127 | else {
|
---|
128 | // transform similarity distances into probabilities by normalizing and inverting the values
|
---|
129 | for (int i = 0; i != weights.Count; ++i)
|
---|
130 | weights[i] = (1 - weights[i] / sum);
|
---|
131 |
|
---|
132 | //selectedBranch = allowedBranches.SelectRandom(weights, random);
|
---|
133 | selectedBranch = SelectRandomBranch(random, allowedBranches, weights);
|
---|
134 | }
|
---|
135 | swap(crossoverPoint0, selectedBranch);
|
---|
136 | return parent0;
|
---|
137 | }
|
---|
138 |
|
---|
139 | private static void swap(CutPoint crossoverPoint, ISymbolicExpressionTreeNode selectedBranch) {
|
---|
140 | if (crossoverPoint.Child != null) {
|
---|
141 | // manipulate the tree of parent0 in place
|
---|
142 | // replace the branch in tree0 with the selected branch from tree1
|
---|
143 | crossoverPoint.Parent.RemoveSubtree(crossoverPoint.ChildIndex);
|
---|
144 | if (selectedBranch != null) {
|
---|
145 | crossoverPoint.Parent.InsertSubtree(crossoverPoint.ChildIndex, selectedBranch);
|
---|
146 | }
|
---|
147 | } else {
|
---|
148 | // child is null (additional child should be added under the parent)
|
---|
149 | if (selectedBranch != null) {
|
---|
150 | crossoverPoint.Parent.AddSubtree(selectedBranch);
|
---|
151 | }
|
---|
152 | }
|
---|
153 | }
|
---|
154 |
|
---|
155 | private static ISymbolicExpressionTreeNode SelectRandomBranch(IRandom random, IList<ISymbolicExpressionTreeNode> nodes, IList<double> weights) {
|
---|
156 | double r = weights.Sum() * random.NextDouble();
|
---|
157 | for (int i = 0; i != nodes.Count; ++i) {
|
---|
158 | if (r < weights[i])
|
---|
159 | return nodes[i];
|
---|
160 | r -= weights[i];
|
---|
161 | }
|
---|
162 | return nodes.Last();
|
---|
163 | }
|
---|
164 | }
|
---|
165 | }
|
---|