source: branches/2936_GQAPIntegration/HeuristicLab.Problems.GeneralizedQuadraticAssignment/3.3/Operators/Crossovers/CordeauCrossover.cs @ 16077

Last change on this file since 16077 was 16077, checked in by abeham, 12 months ago

#2936: Added integration branch

File size: 7.9 KB
Line 
1#region License Information
2/* HeuristicLab
3 * Copyright (C) 2002-2018 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
22using System;
23using System.Linq;
24using HeuristicLab.Common;
25using HeuristicLab.Core;
26using HeuristicLab.Data;
27using HeuristicLab.Encodings.IntegerVectorEncoding;
28using HeuristicLab.Parameters;
29using HeuristicLab.Persistence.Default.CompositeSerializers.Storable;
30using HeuristicLab.Random;
31
32namespace HeuristicLab.Problems.GeneralizedQuadraticAssignment {
33  [Item("CordeauCrossover", "The merge procedure that is described in Cordeau, J.-F., Gaudioso, M., Laporte, G., Moccia, L. 2006. A memetic heuristic for the generalized quadratic assignment problem. INFORMS Journal on Computing, 18, pp. 433–443.")]
34  [StorableClass]
35  public class CordeauCrossover : GQAPCrossover,
36    IQualitiesAwareGQAPOperator, IProblemInstanceAwareGQAPOperator {
37   
38    public IScopeTreeLookupParameter<DoubleValue> QualityParameter {
39      get { return (IScopeTreeLookupParameter<DoubleValue>)Parameters["Quality"]; }
40    }
41    public IScopeTreeLookupParameter<Evaluation> EvaluationParameter {
42      get { return (IScopeTreeLookupParameter<Evaluation>)Parameters["Evaluation"]; }
43    }
44    public ILookupParameter<IntValue> EvaluatedSolutionsParameter {
45      get { return (ILookupParameter<IntValue>)Parameters["EvaluatedSolutions"]; }
46    }
47
48    [StorableConstructor]
49    protected CordeauCrossover(bool deserializing) : base(deserializing) { }
50    protected CordeauCrossover(CordeauCrossover original, Cloner cloner)
51      : base(original, cloner) {
52    }
53    public CordeauCrossover()
54      : base() {
55      Parameters.Add(new ScopeTreeLookupParameter<DoubleValue>("Quality", "The quality of the parents", 1));
56      Parameters.Add(new ScopeTreeLookupParameter<Evaluation>("Evaluation", GQAP.EvaluationDescription, 1));
57      Parameters.Add(new LookupParameter<IntValue>("EvaluatedSolutions", "The number of evaluated solutions."));
58    }
59
60    public override IDeepCloneable Clone(Cloner cloner) {
61      return new CordeauCrossover(this, cloner);
62    }
63
64    public static IntegerVector Apply(IRandom random,
65      IntegerVector parent1, DoubleValue quality1,
66      IntegerVector parent2, DoubleValue quality2,
67      GQAPInstance problemInstance, IntValue evaluatedSolutions) {
68      var distances = problemInstance.Distances;
69      var capacities = problemInstance.Capacities;
70      var demands = problemInstance.Demands;
71
72      var medianDistances = Enumerable.Range(0, distances.Rows).Select(x => distances.GetRow(x).Median()).ToArray();
73
74      int m = capacities.Length;
75      int n = demands.Length;
76     
77      bool onefound = false;
78      double fbest, fbestAttempt = double.MaxValue;
79      IntegerVector bestAttempt = null;
80      IntegerVector result = null;
81
82      fbest = quality1.Value;
83      if (quality1.Value > quality2.Value) {
84        var temp = parent1;
85        parent1 = parent2;
86        parent2 = temp;
87        fbest = quality2.Value;
88      }
89      var cap = new double[m];
90      for (var i = 0; i < m; i++) {
91        int unassigned;
92        Array.Clear(cap, 0, m);
93        var child = Merge(parent1, parent2, distances, demands, medianDistances, m, n, i, cap, out unassigned);
94        if (unassigned > 0)
95          TryRandomAssignment(random, demands, capacities, m, n, cap, child, ref unassigned);
96        if (unassigned == 0) {
97          var childFit = problemInstance.ToSingleObjective(problemInstance.Evaluate(child));
98          evaluatedSolutions.Value++;
99          if (childFit <= fbest) {
100            fbest = childFit;
101            result = child;
102            onefound = true;
103          }
104          if (!onefound && fbestAttempt > childFit) {
105            bestAttempt = child;
106            fbestAttempt = childFit;
107          }
108        }
109      }
110
111      if (!onefound) {
112        var i = random.Next(m);
113        int unassigned;
114        Array.Clear(cap, 0, m);
115        var child = Merge(parent1, parent2, distances, demands, medianDistances, m, n, i, cap, out unassigned);
116        RandomAssignment(random, demands, capacities, m, n, cap, child, ref unassigned);
117
118        var childFit = problemInstance.ToSingleObjective(problemInstance.Evaluate(child));
119        evaluatedSolutions.Value++;
120        if (childFit < fbest) {
121          fbest = childFit;
122          result = child;
123          onefound = true;
124        }
125
126        if (!onefound && fbestAttempt > childFit) {
127          bestAttempt = child;
128          fbestAttempt = childFit;
129        }
130      }
131        /*if (tabufix(&son, 0.5 * sqrt(n * m), round(n * m * log10(n)), &tabufix_it)) {
132          solution_cost(&son);
133          if (son.cost < fbest) {
134            fbest = son.cost;
135            *sptr = son;
136            onefound = TRUE;
137            merge_fixed++;
138          }*/
139      return result ?? bestAttempt;
140    }
141
142    private static IntegerVector Merge(IntegerVector p1, IntegerVector p2,
143      DoubleMatrix distances, DoubleArray demands, double[] mediana,
144      int m, int n, int i, double[] cap, out int unassigned) {
145      unassigned = n;
146      var child = new IntegerVector(n);
147      for (var k = 0; k < n; k++) {
148        child[k] = -1;
149        var ik1 = p1[k];
150        var ik2 = p2[k];
151        if (distances[i, ik1] < mediana[i]) {
152          child[k] = ik1;
153          cap[ik1] += demands[k];
154          unassigned--;
155        } else if (distances[i, ik2] > mediana[i]) {
156          child[k] = ik2;
157          cap[ik2] += demands[k];
158          unassigned--;
159        };
160      }
161      return child;
162    }
163
164    private static bool TryRandomAssignment(IRandom random, DoubleArray demands, DoubleArray capacities, int m, int n, double[] cap, IntegerVector child, ref int unassigned) {
165      var unbiasedOrder = Enumerable.Range(0, n).Shuffle(random).ToList();
166      for (var idx = 0; idx < n; idx++) {
167        var k = unbiasedOrder[idx];
168        if (child[k] < 0) {
169          var feasibleInserts = Enumerable.Range(0, m)
170            .Select((v, i) => new { Pos = i, Slack = capacities[i] - cap[i] })
171            .Where(x => x.Slack >= demands[k]).ToList();
172          if (feasibleInserts.Count == 0) return false;
173          var j = feasibleInserts.SampleRandom(random).Pos;
174          child[k] = j;
175          cap[j] += demands[k];
176          unassigned--;
177        }
178      }
179      return true;
180    }
181
182    private static void RandomAssignment(IRandom random, DoubleArray demands, DoubleArray capacities, int m, int n, double[] cap, IntegerVector child, ref int unassigned) {
183      for (var k = 0; k < n; k++) {
184        if (child[k] < 0) {
185          var j = random.Next(m);
186          child[k] = j;
187          cap[j] += demands[k];
188          unassigned--;
189        }
190      }
191    }
192
193    protected override IntegerVector Cross(IRandom random, ItemArray<IntegerVector> parents,
194      GQAPInstance problemInstance) {
195      if (parents == null) throw new ArgumentNullException("parents");
196      if (parents.Length != 2) throw new ArgumentException(Name + " works only with exactly two parents.");
197
198      var qualities = QualityParameter.ActualValue;
199      return Apply(random,
200        parents[0], qualities[0],
201        parents[1], qualities[1],
202        problemInstance,
203        EvaluatedSolutionsParameter.ActualValue);
204    }
205  }
206}
Note: See TracBrowser for help on using the repository browser.