[7476] | 1 | #region License Information
|
---|
| 2 | /* HeuristicLab
|
---|
[12130] | 3 | * Copyright (C) 2002-2015 Heuristic and Evolutionary Algorithms Laboratory (HEAL)
|
---|
[7476] | 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.Collections.Generic;
|
---|
| 23 | using System.Linq;
|
---|
| 24 | using HeuristicLab.Common;
|
---|
| 25 | using HeuristicLab.Core;
|
---|
[7481] | 26 | using HeuristicLab.Data;
|
---|
[7476] | 27 | using HeuristicLab.Encodings.SymbolicExpressionTreeEncoding;
|
---|
[7481] | 28 | using HeuristicLab.Parameters;
|
---|
| 29 | using HeuristicLab.Persistence.Default.CompositeSerializers.Storable;
|
---|
[12547] | 30 | using HeuristicLab.Random;
|
---|
[7476] | 31 |
|
---|
| 32 | namespace HeuristicLab.Problems.DataAnalysis.Symbolic {
|
---|
[7496] | 33 | [Item("SemanticSimilarityCrossover", "An operator which performs subtree swapping based on the notion semantic similarity between subtrees\n" +
|
---|
| 34 | "(criteria: mean of the absolute differences between the estimated output values of the two subtrees, falling into a user-defined range)\n" +
|
---|
| 35 | "- Take two parent individuals P0 and P1\n" +
|
---|
| 36 | "- Randomly choose a node N from the P0\n" +
|
---|
| 37 | "- Find the first node M that satisfies the semantic similarity criteria\n" +
|
---|
| 38 | "- Swap N for M and return P0")]
|
---|
[7476] | 39 | public sealed class SymbolicDataAnalysisExpressionSemanticSimilarityCrossover<T> : SymbolicDataAnalysisExpressionCrossover<T> where T : class, IDataAnalysisProblemData {
|
---|
| 40 | private const string SemanticSimilarityRangeParameterName = "SemanticSimilarityRange";
|
---|
| 41 |
|
---|
| 42 | #region Parameter properties
|
---|
| 43 | public IValueParameter<DoubleRange> SemanticSimilarityRangeParameter {
|
---|
| 44 | get { return (IValueParameter<DoubleRange>)Parameters[SemanticSimilarityRangeParameterName]; }
|
---|
| 45 | }
|
---|
| 46 | #endregion
|
---|
| 47 |
|
---|
| 48 | #region Properties
|
---|
| 49 | public DoubleRange SemanticSimilarityRange {
|
---|
| 50 | get { return SemanticSimilarityRangeParameter.Value; }
|
---|
| 51 | }
|
---|
| 52 | #endregion
|
---|
| 53 |
|
---|
| 54 | [StorableConstructor]
|
---|
| 55 | private SymbolicDataAnalysisExpressionSemanticSimilarityCrossover(bool deserializing) : base(deserializing) { }
|
---|
| 56 | private SymbolicDataAnalysisExpressionSemanticSimilarityCrossover(SymbolicDataAnalysisExpressionCrossover<T> original, Cloner cloner) : base(original, cloner) { }
|
---|
| 57 | public SymbolicDataAnalysisExpressionSemanticSimilarityCrossover()
|
---|
| 58 | : base() {
|
---|
| 59 | Parameters.Add(new ValueLookupParameter<DoubleRange>(SemanticSimilarityRangeParameterName, "Semantic similarity interval.", new DoubleRange(0.0001, 10)));
|
---|
[7521] | 60 | name = "SemanticSimilarityCrossover";
|
---|
[7476] | 61 | }
|
---|
[7488] | 62 | public override IDeepCloneable Clone(Cloner cloner) {
|
---|
| 63 | return new SymbolicDataAnalysisExpressionSemanticSimilarityCrossover<T>(this, cloner);
|
---|
| 64 | }
|
---|
[7476] | 65 |
|
---|
[7481] | 66 | public override ISymbolicExpressionTree Crossover(IRandom random, ISymbolicExpressionTree parent0, ISymbolicExpressionTree parent1) {
|
---|
[7476] | 67 | ISymbolicDataAnalysisExpressionTreeInterpreter interpreter = SymbolicDataAnalysisTreeInterpreterParameter.ActualValue;
|
---|
| 68 | List<int> rows = GenerateRowsToEvaluate().ToList();
|
---|
| 69 | T problemData = ProblemDataParameter.ActualValue;
|
---|
| 70 | return Cross(random, parent0, parent1, interpreter, problemData, rows, MaximumSymbolicExpressionTreeDepth.Value, MaximumSymbolicExpressionTreeLength.Value, SemanticSimilarityRange);
|
---|
| 71 | }
|
---|
| 72 |
|
---|
| 73 | /// <summary>
|
---|
| 74 | /// Takes two parent individuals P0 and P1.
|
---|
| 75 | /// Randomly choose a node i from the first parent, then get a node j from the second parent that matches the semantic similarity criteria.
|
---|
| 76 | /// </summary>
|
---|
| 77 | public static ISymbolicExpressionTree Cross(IRandom random, ISymbolicExpressionTree parent0, ISymbolicExpressionTree parent1, ISymbolicDataAnalysisExpressionTreeInterpreter interpreter,
|
---|
| 78 | T problemData, List<int> rows, int maxDepth, int maxLength, DoubleRange range) {
|
---|
| 79 | var crossoverPoints0 = new List<CutPoint>();
|
---|
| 80 | parent0.Root.ForEachNodePostfix((n) => {
|
---|
[7496] | 81 | if (n.Parent != null && n.Parent != parent0.Root)
|
---|
| 82 | crossoverPoints0.Add(new CutPoint(n.Parent, n));
|
---|
[7476] | 83 | });
|
---|
[12547] | 84 |
|
---|
| 85 | var crossoverPoint0 = crossoverPoints0.SampleRandom(random);
|
---|
[7476] | 86 | int level = parent0.Root.GetBranchLevel(crossoverPoint0.Child);
|
---|
| 87 | int length = parent0.Root.GetLength() - crossoverPoint0.Child.GetLength();
|
---|
| 88 |
|
---|
| 89 | var allowedBranches = new List<ISymbolicExpressionTreeNode>();
|
---|
| 90 | parent1.Root.ForEachNodePostfix((n) => {
|
---|
[7496] | 91 | if (n.Parent != null && n.Parent != parent1.Root) {
|
---|
| 92 | if (n.GetDepth() + level <= maxDepth && n.GetLength() + length <= maxLength && crossoverPoint0.IsMatchingPointType(n))
|
---|
| 93 | allowedBranches.Add(n);
|
---|
| 94 | }
|
---|
[7476] | 95 | });
|
---|
| 96 |
|
---|
| 97 | if (allowedBranches.Count == 0)
|
---|
| 98 | return parent0;
|
---|
| 99 |
|
---|
| 100 | var dataset = problemData.Dataset;
|
---|
| 101 |
|
---|
| 102 | // create symbols in order to improvize an ad-hoc tree so that the child can be evaluated
|
---|
| 103 | var rootSymbol = new ProgramRootSymbol();
|
---|
| 104 | var startSymbol = new StartSymbol();
|
---|
| 105 | var tree0 = CreateTreeFromNode(random, crossoverPoint0.Child, rootSymbol, startSymbol);
|
---|
| 106 | List<double> estimatedValues0 = interpreter.GetSymbolicExpressionTreeValues(tree0, dataset, rows).ToList();
|
---|
| 107 | crossoverPoint0.Child.Parent = crossoverPoint0.Parent; // restore parent
|
---|
| 108 | ISymbolicExpressionTreeNode selectedBranch = null;
|
---|
| 109 |
|
---|
| 110 | // pick the first node that fulfills the semantic similarity conditions
|
---|
| 111 | foreach (var node in allowedBranches) {
|
---|
| 112 | var parent = node.Parent;
|
---|
| 113 | var tree1 = CreateTreeFromNode(random, node, startSymbol, rootSymbol); // this will affect node.Parent
|
---|
| 114 | List<double> estimatedValues1 = interpreter.GetSymbolicExpressionTreeValues(tree1, dataset, rows).ToList();
|
---|
| 115 | node.Parent = parent; // restore parent
|
---|
| 116 |
|
---|
| 117 | OnlineCalculatorError errorState;
|
---|
| 118 | double ssd = OnlineMeanAbsoluteErrorCalculator.Calculate(estimatedValues0, estimatedValues1, out errorState);
|
---|
| 119 |
|
---|
[7488] | 120 | if (range.Start <= ssd && ssd <= range.End) {
|
---|
| 121 | selectedBranch = node;
|
---|
| 122 | break;
|
---|
| 123 | }
|
---|
[7476] | 124 | }
|
---|
| 125 |
|
---|
| 126 | // perform the actual swap
|
---|
| 127 | if (selectedBranch != null)
|
---|
[7496] | 128 | Swap(crossoverPoint0, selectedBranch);
|
---|
[7476] | 129 | return parent0;
|
---|
| 130 | }
|
---|
| 131 | }
|
---|
| 132 | }
|
---|