1 | #region License Information
|
---|
2 | /* HeuristicLab
|
---|
3 | * Copyright (C) 2002-2008 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 System.Linq;
|
---|
24 | using HeuristicLab.Core;
|
---|
25 | using HeuristicLab.Random;
|
---|
26 | using System.Diagnostics;
|
---|
27 | using HeuristicLab.GP.Interfaces;
|
---|
28 |
|
---|
29 | namespace HeuristicLab.Encodings.SymbolicExpressionTree {
|
---|
30 | public class CutOutNodeManipulation : GPManipulatorBase {
|
---|
31 | public override string Description {
|
---|
32 | get {
|
---|
33 | return @"Takes a tree, selects a random node of the tree and then tries to replace a random sub-tree
|
---|
34 | of that node with one of the childs of the selected child.
|
---|
35 |
|
---|
36 | O O
|
---|
37 | / \ / \
|
---|
38 | O X O 2
|
---|
39 | / \ 2 is selected => / \
|
---|
40 | 1 2 4 5
|
---|
41 | / / \
|
---|
42 | 3 4 5
|
---|
43 | ";
|
---|
44 | }
|
---|
45 | }
|
---|
46 |
|
---|
47 | public CutOutNodeManipulation()
|
---|
48 | : base() {
|
---|
49 | }
|
---|
50 |
|
---|
51 |
|
---|
52 | internal override IOperation Manipulate(MersenneTwister random, IGeneticProgrammingModel gpModel, FunctionLibrary library, int maxTreeSize, int maxTreeHeight, IScope scope) {
|
---|
53 | TreeGardener gardener = new TreeGardener(random, library);
|
---|
54 | IFunctionTree parent = gardener.GetRandomParentNode(gpModel.FunctionTree);
|
---|
55 | // parent == null means we should cut out the root node
|
---|
56 | // => return a random sub-tree of the root
|
---|
57 | if (parent == null) {
|
---|
58 | // when there are sub-trees then replace the old tree with a random sub-tree
|
---|
59 | if (gpModel.FunctionTree.SubTrees.Count > 0) {
|
---|
60 | gpModel.FunctionTree = gpModel.FunctionTree.SubTrees[random.Next(gpModel.FunctionTree.SubTrees.Count)];
|
---|
61 | Debug.Assert(gardener.IsValidTree(gpModel.FunctionTree));
|
---|
62 | // we reused a sub-tree so we don't have to schedule initialization operations
|
---|
63 | return null;
|
---|
64 | } else {
|
---|
65 | // we want to cut the root node and there are no sub-trees => create a new random terminal
|
---|
66 | gpModel.FunctionTree = gardener.CreateRandomTree(gardener.Terminals, 1, 1);
|
---|
67 | Debug.Assert(gardener.IsValidTree(gpModel.FunctionTree));
|
---|
68 |
|
---|
69 | // schedule an operation to initialize the whole tree
|
---|
70 | return Util.CreateInitializationOperation(TreeGardener.GetAllSubTrees(gpModel.FunctionTree), scope);
|
---|
71 | }
|
---|
72 | }
|
---|
73 |
|
---|
74 | // select a child to cut away
|
---|
75 | int childIndex = random.Next(parent.SubTrees.Count);
|
---|
76 | IFunctionTree child = parent.SubTrees[childIndex];
|
---|
77 | // match the sub-trees of the child with the allowed sub-trees of the parent
|
---|
78 | ICollection<IFunction> allowedFunctions = gardener.GetAllowedSubFunctions(parent.Function, childIndex);
|
---|
79 | IFunctionTree[] possibleChilds = child.SubTrees.Where(t => allowedFunctions.Contains(t.Function)).ToArray();
|
---|
80 | if (possibleChilds.Length > 0) {
|
---|
81 | // replace child with a random child of that child
|
---|
82 | IFunctionTree selectedChild = possibleChilds[random.Next(possibleChilds.Length)];
|
---|
83 | parent.RemoveSubTree(childIndex);
|
---|
84 | parent.InsertSubTree(childIndex, selectedChild);
|
---|
85 | Debug.Assert(gardener.IsValidTree(gpModel.FunctionTree));
|
---|
86 | // recalculate the size and height of our tree
|
---|
87 | gpModel.Size = gpModel.FunctionTree.GetSize();
|
---|
88 | gpModel.Height = gpModel.FunctionTree.GetHeight();
|
---|
89 | // don't need to schedule initialization operations
|
---|
90 | return null;
|
---|
91 | } else {
|
---|
92 | // can't reuse an existing branch => create a new tree
|
---|
93 | // determine the level of the parent
|
---|
94 | int parentLevel = gardener.GetBranchLevel(gpModel.FunctionTree, parent);
|
---|
95 | // first remove the old child (first step essential!)
|
---|
96 | parent.RemoveSubTree(childIndex);
|
---|
97 | // then determine the number of nodes left over after the child has been removed!
|
---|
98 | int remainingNodes = gpModel.FunctionTree.GetSize();
|
---|
99 | allowedFunctions = gardener.GetAllowedSubFunctions(parent.Function, childIndex);
|
---|
100 | IFunctionTree newFunctionTree = gardener.CreateRandomTree(allowedFunctions, maxTreeSize - remainingNodes, maxTreeHeight - parentLevel);
|
---|
101 | parent.InsertSubTree(childIndex, newFunctionTree);
|
---|
102 | Debug.Assert(gardener.IsValidTree(gpModel.FunctionTree));
|
---|
103 | // recalculate size and height
|
---|
104 | gpModel.Size = gpModel.FunctionTree.GetSize();
|
---|
105 | gpModel.Height = gpModel.FunctionTree.GetHeight();
|
---|
106 | // schedule an initialization operation for the new function-tree
|
---|
107 | return Util.CreateInitializationOperation(TreeGardener.GetAllSubTrees(newFunctionTree), scope);
|
---|
108 | }
|
---|
109 | }
|
---|
110 | }
|
---|
111 | }
|
---|