1 | #region License Information
|
---|
2 | /* HeuristicLab
|
---|
3 | * Copyright (C) 2002-2010 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.Collections.Generic;
|
---|
23 | using HeuristicLab.Common;
|
---|
24 | using HeuristicLab.Core;
|
---|
25 | using HeuristicLab.Persistence.Default.CompositeSerializers.Storable;
|
---|
26 | using HeuristicLab.Data;
|
---|
27 | using System.Linq;
|
---|
28 | using System;
|
---|
29 | using HeuristicLab.Parameters;
|
---|
30 | using HeuristicLab.Encodings.SymbolicExpressionTreeEncoding.Symbols;
|
---|
31 |
|
---|
32 | namespace HeuristicLab.Encodings.SymbolicExpressionTreeEncoding.Crossovers {
|
---|
33 | /// <summary>
|
---|
34 | /// Takes two parent individuals P0 and P1 each. Selects a random node N0 of P0 and a random node N1 of P1.
|
---|
35 | /// And replaces the branch with root0 N0 in P0 with N1 from P1 if the tree-size limits are not violated.
|
---|
36 | /// When recombination with N0 and N1 would create a tree that is too large or invalid the operator randomly selects new N0 and N1
|
---|
37 | /// until a valid configuration is found.
|
---|
38 | /// </summary>
|
---|
39 | [Item("SubtreeCrossover", "An operator which performs subtree swapping crossover.")]
|
---|
40 | [StorableClass]
|
---|
41 | public class SubtreeCrossover : SymbolicExpressionTreeCrossover {
|
---|
42 | public IValueLookupParameter<PercentValue> InternalCrossoverPointProbabilityParameter {
|
---|
43 | get { return (IValueLookupParameter<PercentValue>)Parameters["InternalCrossoverPointProbability"]; }
|
---|
44 | }
|
---|
45 |
|
---|
46 | public SubtreeCrossover()
|
---|
47 | : base() {
|
---|
48 | Parameters.Add(new ValueLookupParameter<PercentValue>("InternalCrossoverPointProbability", "The probability to select an internal crossover point (instead of a leaf node).", new PercentValue(0.9)));
|
---|
49 | }
|
---|
50 |
|
---|
51 | protected override SymbolicExpressionTree Cross(IRandom random,
|
---|
52 | SymbolicExpressionTree parent0, SymbolicExpressionTree parent1,
|
---|
53 | IntValue maxTreeSize, IntValue maxTreeHeight, out bool success) {
|
---|
54 | return Cross(random, parent0, parent1, InternalCrossoverPointProbabilityParameter.ActualValue.Value, maxTreeSize.Value, maxTreeHeight.Value, out success);
|
---|
55 | }
|
---|
56 |
|
---|
57 | public static SymbolicExpressionTree Cross(IRandom random,
|
---|
58 | SymbolicExpressionTree parent0, SymbolicExpressionTree parent1,
|
---|
59 | double internalCrossoverPointProbability, int maxTreeSize, int maxTreeHeight, out bool success) {
|
---|
60 | // select a random crossover point in the first parent
|
---|
61 | SymbolicExpressionTreeNode crossoverPoint0;
|
---|
62 | int replacedSubtreeIndex;
|
---|
63 | SelectCrossoverPoint(random, parent0, internalCrossoverPointProbability, maxTreeSize - 1, maxTreeHeight - 1, out crossoverPoint0, out replacedSubtreeIndex);
|
---|
64 |
|
---|
65 | // calculate the max size and height that the inserted branch can have
|
---|
66 | int maxInsertedBranchSize = maxTreeSize - (parent0.Size - crossoverPoint0.SubTrees[replacedSubtreeIndex].GetSize());
|
---|
67 | int maxInsertedBranchHeight = maxTreeHeight - GetBranchLevel(parent0.Root, crossoverPoint0);
|
---|
68 |
|
---|
69 | var allowedBranches = (from branch in parent1.Root.IterateNodesPrefix()
|
---|
70 | where branch.GetSize() < maxInsertedBranchSize
|
---|
71 | where branch.GetHeight() < maxInsertedBranchHeight
|
---|
72 | where IsMatchingPointType(crossoverPoint0, replacedSubtreeIndex, branch)
|
---|
73 | select branch).ToList();
|
---|
74 |
|
---|
75 | if (allowedBranches.Count() == 0) {
|
---|
76 | success = false;
|
---|
77 | return parent0;
|
---|
78 | } else {
|
---|
79 | var selectedBranch = SelectRandomBranch(random, allowedBranches, internalCrossoverPointProbability);
|
---|
80 |
|
---|
81 | // manipulate the tree of parent0 in place
|
---|
82 | // replace the branch in tree0 with the selected branch from tree1
|
---|
83 | crossoverPoint0.RemoveSubTree(replacedSubtreeIndex);
|
---|
84 | crossoverPoint0.InsertSubTree(replacedSubtreeIndex, selectedBranch);
|
---|
85 | success = true;
|
---|
86 | return parent0;
|
---|
87 | }
|
---|
88 | }
|
---|
89 |
|
---|
90 | private static bool IsMatchingPointType(SymbolicExpressionTreeNode parent, int replacedSubtreeIndex, SymbolicExpressionTreeNode branch) {
|
---|
91 | // check point type for the whole branch
|
---|
92 | foreach (var node in branch.IterateNodesPostfix()) {
|
---|
93 | if (!parent.Grammar.ContainsSymbol(node.Symbol)) return false;
|
---|
94 | else if (node.SubTrees.Count < parent.Grammar.GetMinSubtreeCount(node.Symbol)) return false;
|
---|
95 | else if (node.SubTrees.Count > parent.Grammar.GetMaxSubtreeCount(node.Symbol)) return false;
|
---|
96 | }
|
---|
97 |
|
---|
98 | // check syntax constraints of direct parent - child relation
|
---|
99 | if (!parent.Grammar.IsAllowedChild(parent.Symbol, branch.Symbol, replacedSubtreeIndex)) return false;
|
---|
100 |
|
---|
101 | return true;
|
---|
102 | }
|
---|
103 |
|
---|
104 | private static void SelectCrossoverPoint(IRandom random, SymbolicExpressionTree parent0, double internalNodeProbability, int maxBranchSize, int maxBranchHeight, out SymbolicExpressionTreeNode crossoverPoint, out int subtreeIndex) {
|
---|
105 | var crossoverPoints = from branch in parent0.Root.IterateNodesPrefix()
|
---|
106 | where branch.SubTrees.Count > 0
|
---|
107 | where branch != parent0.Root
|
---|
108 | where branch.GetSize() < maxBranchSize
|
---|
109 | where branch.GetHeight() < maxBranchHeight
|
---|
110 | from index in Enumerable.Range(0, branch.SubTrees.Count)
|
---|
111 | let p = new { CrossoverPoint = branch, SubtreeIndex = index, IsLeaf = branch.SubTrees[index].SubTrees.Count == 0 }
|
---|
112 | select p;
|
---|
113 | var internalCrossoverPoints = (from p in crossoverPoints
|
---|
114 | where !p.IsLeaf
|
---|
115 | select p).ToList();
|
---|
116 | var leafCrossoverPoints = (from p in crossoverPoints
|
---|
117 | where p.IsLeaf
|
---|
118 | select p).ToList();
|
---|
119 | if (internalCrossoverPoints.Count == 0) {
|
---|
120 | var selectedCrossoverPoint = leafCrossoverPoints[random.Next(leafCrossoverPoints.Count)];
|
---|
121 | crossoverPoint = selectedCrossoverPoint.CrossoverPoint;
|
---|
122 | subtreeIndex = selectedCrossoverPoint.SubtreeIndex;
|
---|
123 | } else if (leafCrossoverPoints.Count == 0) {
|
---|
124 | var selectedCrossoverPoint = internalCrossoverPoints[random.Next(internalCrossoverPoints.Count)];
|
---|
125 | crossoverPoint = selectedCrossoverPoint.CrossoverPoint;
|
---|
126 | subtreeIndex = selectedCrossoverPoint.SubtreeIndex;
|
---|
127 | } else if (random.NextDouble() < internalNodeProbability && internalCrossoverPoints.Count > 0) {
|
---|
128 | // select internal crossover point or leaf
|
---|
129 | var selectedCrossoverPoint = internalCrossoverPoints[random.Next(internalCrossoverPoints.Count)];
|
---|
130 | crossoverPoint = selectedCrossoverPoint.CrossoverPoint;
|
---|
131 | subtreeIndex = selectedCrossoverPoint.SubtreeIndex;
|
---|
132 | } else {
|
---|
133 | var selectedCrossoverPoint = leafCrossoverPoints[random.Next(leafCrossoverPoints.Count)];
|
---|
134 | crossoverPoint = selectedCrossoverPoint.CrossoverPoint;
|
---|
135 | subtreeIndex = selectedCrossoverPoint.SubtreeIndex;
|
---|
136 | }
|
---|
137 | }
|
---|
138 |
|
---|
139 | private static SymbolicExpressionTreeNode SelectRandomBranch(IRandom random, IEnumerable<SymbolicExpressionTreeNode> branches, double internalNodeProbability) {
|
---|
140 | if (internalNodeProbability < 0.0 || internalNodeProbability > 1.0) throw new ArgumentException("internalNodeProbability");
|
---|
141 | var groupedBranches = (from branch in branches
|
---|
142 | group branch by branch.SubTrees.Count into g
|
---|
143 | select g).ToList();
|
---|
144 | var allowedInternalBranches = (from g in groupedBranches
|
---|
145 | where g.Key > 0
|
---|
146 | from branch in g
|
---|
147 | select branch).ToList();
|
---|
148 | var allowedLeafBranches = (from g in groupedBranches
|
---|
149 | where g.Key == 0
|
---|
150 | from leaf in g
|
---|
151 | select leaf).ToList();
|
---|
152 | if (allowedInternalBranches.Count == 0) {
|
---|
153 | return allowedLeafBranches.SelectRandom(random);
|
---|
154 | } else if (allowedLeafBranches.Count == 0) {
|
---|
155 | return allowedInternalBranches.SelectRandom(random);
|
---|
156 | } else if (random.NextDouble() < internalNodeProbability) {
|
---|
157 | // when leaf and internal nodes are possible then choose either a leaf or internal node with internalNodeProbability
|
---|
158 | return allowedInternalBranches.SelectRandom(random);
|
---|
159 | } else {
|
---|
160 | return allowedLeafBranches.SelectRandom(random);
|
---|
161 | }
|
---|
162 | }
|
---|
163 |
|
---|
164 | private static int GetBranchLevel(SymbolicExpressionTreeNode root, SymbolicExpressionTreeNode point) {
|
---|
165 | if (root == point) return 0;
|
---|
166 | foreach (var subtree in root.SubTrees) {
|
---|
167 | int branchLevel = GetBranchLevel(subtree, point);
|
---|
168 | if (branchLevel < int.MaxValue) return 1 + branchLevel;
|
---|
169 | }
|
---|
170 | return int.MaxValue;
|
---|
171 | }
|
---|
172 | }
|
---|
173 | }
|
---|