[645] | 1 | #region License Information
|
---|
| 2 | /* HeuristicLab
|
---|
[3237] | 3 | * Copyright (C) 2002-2010 Heuristic and Evolutionary Algorithms Laboratory (HEAL)
|
---|
[645] | 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 |
|
---|
[4068] | 22 | using System;
|
---|
[645] | 23 | using System.Collections.Generic;
|
---|
[4068] | 24 | using System.Linq;
|
---|
[4674] | 25 | using HeuristicLab.Common;
|
---|
[645] | 26 | using HeuristicLab.Core;
|
---|
[3237] | 27 | using HeuristicLab.Data;
|
---|
| 28 | using HeuristicLab.Parameters;
|
---|
[4068] | 29 | using HeuristicLab.Persistence.Default.CompositeSerializers.Storable;
|
---|
[645] | 30 |
|
---|
[3462] | 31 | namespace HeuristicLab.Encodings.SymbolicExpressionTreeEncoding.Crossovers {
|
---|
[3237] | 32 | /// <summary>
|
---|
| 33 | /// Takes two parent individuals P0 and P1 each. Selects a random node N0 of P0 and a random node N1 of P1.
|
---|
| 34 | /// And replaces the branch with root0 N0 in P0 with N1 from P1 if the tree-size limits are not violated.
|
---|
| 35 | /// When recombination with N0 and N1 would create a tree that is too large or invalid the operator randomly selects new N0 and N1
|
---|
| 36 | /// until a valid configuration is found.
|
---|
| 37 | /// </summary>
|
---|
| 38 | [Item("SubtreeCrossover", "An operator which performs subtree swapping crossover.")]
|
---|
| 39 | [StorableClass]
|
---|
[4674] | 40 | public sealed class SubtreeCrossover : SymbolicExpressionTreeCrossover {
|
---|
[3237] | 41 | public IValueLookupParameter<PercentValue> InternalCrossoverPointProbabilityParameter {
|
---|
| 42 | get { return (IValueLookupParameter<PercentValue>)Parameters["InternalCrossoverPointProbability"]; }
|
---|
[645] | 43 | }
|
---|
[4674] | 44 | [StorableConstructor]
|
---|
| 45 | private SubtreeCrossover(bool deserializing) : base(deserializing) { }
|
---|
| 46 | private SubtreeCrossover(SubtreeCrossover original, Cloner cloner) : base(original, cloner) { }
|
---|
[3237] | 47 | public SubtreeCrossover()
|
---|
| 48 | : base() {
|
---|
| 49 | Parameters.Add(new ValueLookupParameter<PercentValue>("InternalCrossoverPointProbability", "The probability to select an internal crossover point (instead of a leaf node).", new PercentValue(0.9)));
|
---|
| 50 | }
|
---|
| 51 |
|
---|
[4674] | 52 | public override IDeepCloneable Clone(Cloner cloner) {
|
---|
| 53 | return new SubtreeCrossover(this, cloner);
|
---|
| 54 | }
|
---|
| 55 |
|
---|
[3338] | 56 | protected override SymbolicExpressionTree Cross(IRandom random,
|
---|
[3237] | 57 | SymbolicExpressionTree parent0, SymbolicExpressionTree parent1,
|
---|
[3294] | 58 | IntValue maxTreeSize, IntValue maxTreeHeight, out bool success) {
|
---|
[3338] | 59 | return Cross(random, parent0, parent1, InternalCrossoverPointProbabilityParameter.ActualValue.Value, maxTreeSize.Value, maxTreeHeight.Value, out success);
|
---|
[3237] | 60 | }
|
---|
| 61 |
|
---|
[3338] | 62 | public static SymbolicExpressionTree Cross(IRandom random,
|
---|
[3237] | 63 | SymbolicExpressionTree parent0, SymbolicExpressionTree parent1,
|
---|
[3294] | 64 | double internalCrossoverPointProbability, int maxTreeSize, int maxTreeHeight, out bool success) {
|
---|
| 65 | // select a random crossover point in the first parent
|
---|
| 66 | SymbolicExpressionTreeNode crossoverPoint0;
|
---|
| 67 | int replacedSubtreeIndex;
|
---|
[3369] | 68 | SelectCrossoverPoint(random, parent0, internalCrossoverPointProbability, maxTreeSize - 1, maxTreeHeight - 1, out crossoverPoint0, out replacedSubtreeIndex);
|
---|
[645] | 69 |
|
---|
[3294] | 70 | // calculate the max size and height that the inserted branch can have
|
---|
| 71 | int maxInsertedBranchSize = maxTreeSize - (parent0.Size - crossoverPoint0.SubTrees[replacedSubtreeIndex].GetSize());
|
---|
| 72 | int maxInsertedBranchHeight = maxTreeHeight - GetBranchLevel(parent0.Root, crossoverPoint0);
|
---|
[645] | 73 |
|
---|
[3997] | 74 | List<SymbolicExpressionTreeNode> allowedBranches = new List<SymbolicExpressionTreeNode>();
|
---|
| 75 | parent1.Root.ForEachNodePostfix((n) => {
|
---|
| 76 | if (n.GetSize() < maxInsertedBranchSize &&
|
---|
| 77 | n.GetHeight() < maxInsertedBranchHeight &&
|
---|
| 78 | IsMatchingPointType(crossoverPoint0, replacedSubtreeIndex, n))
|
---|
| 79 | allowedBranches.Add(n);
|
---|
| 80 | });
|
---|
[645] | 81 |
|
---|
[3997] | 82 | if (allowedBranches.Count == 0) {
|
---|
[3297] | 83 | success = false;
|
---|
| 84 | return parent0;
|
---|
| 85 | } else {
|
---|
[3294] | 86 | var selectedBranch = SelectRandomBranch(random, allowedBranches, internalCrossoverPointProbability);
|
---|
[645] | 87 |
|
---|
[3294] | 88 | // manipulate the tree of parent0 in place
|
---|
| 89 | // replace the branch in tree0 with the selected branch from tree1
|
---|
| 90 | crossoverPoint0.RemoveSubTree(replacedSubtreeIndex);
|
---|
| 91 | crossoverPoint0.InsertSubTree(replacedSubtreeIndex, selectedBranch);
|
---|
| 92 | success = true;
|
---|
| 93 | return parent0;
|
---|
[645] | 94 | }
|
---|
| 95 | }
|
---|
| 96 |
|
---|
[3338] | 97 | private static bool IsMatchingPointType(SymbolicExpressionTreeNode parent, int replacedSubtreeIndex, SymbolicExpressionTreeNode branch) {
|
---|
| 98 | // check syntax constraints of direct parent - child relation
|
---|
[4106] | 99 | if (!parent.Grammar.ContainsSymbol(branch.Symbol) ||
|
---|
| 100 | !parent.Grammar.IsAllowedChild(parent.Symbol, branch.Symbol, replacedSubtreeIndex)) return false;
|
---|
[3338] | 101 |
|
---|
[3997] | 102 | bool result = true;
|
---|
| 103 | // check point type for the whole branch
|
---|
| 104 | branch.ForEachNodePostfix((n) => {
|
---|
[3998] | 105 | result =
|
---|
| 106 | result &&
|
---|
[4106] | 107 | parent.Grammar.ContainsSymbol(n.Symbol) &&
|
---|
[3998] | 108 | n.SubTrees.Count >= parent.Grammar.GetMinSubtreeCount(n.Symbol) &&
|
---|
[4106] | 109 | n.SubTrees.Count <= parent.Grammar.GetMaxSubtreeCount(n.Symbol);
|
---|
[3997] | 110 | });
|
---|
| 111 | return result;
|
---|
[3294] | 112 | }
|
---|
| 113 |
|
---|
[3369] | 114 | private static void SelectCrossoverPoint(IRandom random, SymbolicExpressionTree parent0, double internalNodeProbability, int maxBranchSize, int maxBranchHeight, out SymbolicExpressionTreeNode crossoverPoint, out int subtreeIndex) {
|
---|
[3997] | 115 | if (internalNodeProbability < 0.0 || internalNodeProbability > 1.0) throw new ArgumentException("internalNodeProbability");
|
---|
| 116 | List<CrossoverPoint> internalCrossoverPoints = new List<CrossoverPoint>();
|
---|
| 117 | List<CrossoverPoint> leafCrossoverPoints = new List<CrossoverPoint>();
|
---|
| 118 | parent0.Root.ForEachNodePostfix((n) => {
|
---|
| 119 | if (n.SubTrees.Count > 0 &&
|
---|
| 120 | n.GetSize() < maxBranchSize &&
|
---|
| 121 | n.GetHeight() < maxBranchHeight &&
|
---|
| 122 | n != parent0.Root
|
---|
| 123 | ) {
|
---|
| 124 | foreach (var child in n.SubTrees) {
|
---|
| 125 | if (child.SubTrees.Count > 0)
|
---|
| 126 | internalCrossoverPoints.Add(new CrossoverPoint(n, child));
|
---|
| 127 | else
|
---|
| 128 | leafCrossoverPoints.Add(new CrossoverPoint(n, child));
|
---|
| 129 | }
|
---|
| 130 | }
|
---|
| 131 | });
|
---|
| 132 | if (random.NextDouble() < internalNodeProbability) {
|
---|
| 133 | // select from internal node if possible
|
---|
| 134 | if (internalCrossoverPoints.Count > 0) {
|
---|
| 135 | // select internal crossover point or leaf
|
---|
| 136 | var selectedCrossoverPoint = internalCrossoverPoints[random.Next(internalCrossoverPoints.Count)];
|
---|
| 137 | crossoverPoint = selectedCrossoverPoint.Parent;
|
---|
| 138 | subtreeIndex = selectedCrossoverPoint.SubtreeIndex;
|
---|
| 139 | } else {
|
---|
| 140 | // otherwise select external node
|
---|
| 141 | var selectedCrossoverPoint = leafCrossoverPoints[random.Next(leafCrossoverPoints.Count)];
|
---|
| 142 | crossoverPoint = selectedCrossoverPoint.Parent;
|
---|
| 143 | subtreeIndex = selectedCrossoverPoint.SubtreeIndex;
|
---|
| 144 | }
|
---|
| 145 | } else if (leafCrossoverPoints.Count > 0) {
|
---|
| 146 | // select from leaf crossover point if possible
|
---|
[3294] | 147 | var selectedCrossoverPoint = leafCrossoverPoints[random.Next(leafCrossoverPoints.Count)];
|
---|
[3997] | 148 | crossoverPoint = selectedCrossoverPoint.Parent;
|
---|
[3294] | 149 | subtreeIndex = selectedCrossoverPoint.SubtreeIndex;
|
---|
[3997] | 150 | } else {
|
---|
| 151 | // otherwise select internal crossover point
|
---|
[3237] | 152 | var selectedCrossoverPoint = internalCrossoverPoints[random.Next(internalCrossoverPoints.Count)];
|
---|
[3997] | 153 | crossoverPoint = selectedCrossoverPoint.Parent;
|
---|
[3237] | 154 | subtreeIndex = selectedCrossoverPoint.SubtreeIndex;
|
---|
[645] | 155 | }
|
---|
| 156 | }
|
---|
[3237] | 157 |
|
---|
| 158 | private static SymbolicExpressionTreeNode SelectRandomBranch(IRandom random, IEnumerable<SymbolicExpressionTreeNode> branches, double internalNodeProbability) {
|
---|
| 159 | if (internalNodeProbability < 0.0 || internalNodeProbability > 1.0) throw new ArgumentException("internalNodeProbability");
|
---|
[3997] | 160 | List<SymbolicExpressionTreeNode> allowedInternalBranches;
|
---|
| 161 | List<SymbolicExpressionTreeNode> allowedLeafBranches;
|
---|
| 162 | if (random.NextDouble() < internalNodeProbability) {
|
---|
| 163 | // select internal node if possible
|
---|
| 164 | allowedInternalBranches = (from branch in branches
|
---|
| 165 | where branch.SubTrees.Count > 0
|
---|
| 166 | select branch).ToList();
|
---|
| 167 | if (allowedInternalBranches.Count > 0) {
|
---|
| 168 | return allowedInternalBranches.SelectRandom(random);
|
---|
| 169 | } else {
|
---|
| 170 | // no internal nodes allowed => select leaf nodes
|
---|
| 171 | allowedLeafBranches = (from branch in branches
|
---|
[3989] | 172 | where branch.SubTrees.Count == 0
|
---|
| 173 | select branch).ToList();
|
---|
[3997] | 174 | return allowedLeafBranches.SelectRandom(random);
|
---|
| 175 | }
|
---|
[3237] | 176 | } else {
|
---|
[3997] | 177 | // select leaf node if possible
|
---|
| 178 | allowedLeafBranches = (from branch in branches
|
---|
| 179 | where branch.SubTrees.Count == 0
|
---|
| 180 | select branch).ToList();
|
---|
| 181 | if (allowedLeafBranches.Count > 0) {
|
---|
| 182 | return allowedLeafBranches.SelectRandom(random);
|
---|
| 183 | } else {
|
---|
| 184 | allowedInternalBranches = (from branch in branches
|
---|
| 185 | where branch.SubTrees.Count > 0
|
---|
| 186 | select branch).ToList();
|
---|
| 187 | return allowedInternalBranches.SelectRandom(random);
|
---|
| 188 | }
|
---|
[3237] | 189 | }
|
---|
| 190 | }
|
---|
| 191 |
|
---|
| 192 | private static int GetBranchLevel(SymbolicExpressionTreeNode root, SymbolicExpressionTreeNode point) {
|
---|
| 193 | if (root == point) return 0;
|
---|
| 194 | foreach (var subtree in root.SubTrees) {
|
---|
| 195 | int branchLevel = GetBranchLevel(subtree, point);
|
---|
| 196 | if (branchLevel < int.MaxValue) return 1 + branchLevel;
|
---|
| 197 | }
|
---|
| 198 | return int.MaxValue;
|
---|
| 199 | }
|
---|
[645] | 200 | }
|
---|
| 201 | }
|
---|