Free cookie consent management tool by TermsFeed Policy Generator

source: branches/2845_EnhancedProgress/HeuristicLab.Problems.Orienteering/3.3/Improvers/OrienteeringLocalImprovementOperator.cs @ 16307

Last change on this file since 16307 was 16307, checked in by pfleck, 6 years ago

#2845 Merged trunk changes before source move into branch

File size: 13.5 KB
RevLine 
[11193]1#region License Information
2/* HeuristicLab
[16307]3 * Copyright (C) 2002-2018 Heuristic and Evolutionary Algorithms Laboratory (HEAL)
[11193]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.Collections.Generic;
23using System.Linq;
24using HeuristicLab.Common;
25using HeuristicLab.Core;
26using HeuristicLab.Data;
27using HeuristicLab.Encodings.IntegerVectorEncoding;
28using HeuristicLab.Operators;
29using HeuristicLab.Optimization;
30using HeuristicLab.Parameters;
31using HeuristicLab.Persistence.Default.CompositeSerializers.Storable;
32
33namespace HeuristicLab.Problems.Orienteering {
[12721]34  /// <summary>
35  /// Iterative improvement consists of three basic operators: shortening, vertex insert and vertex
36  /// exchange. The shortening operator tries to rearrange the vertices within a tour in order to
37  /// minimize the cost of the tour. As shortening operator a 2-opt is applied. (Schilde et. al. 2009)
38  /// </summary>
39  [Item("OrienteeringLocalImprovementOperator", @"Implements the iterative improvement procedure described in Schilde M., Doerner K.F., Hartl R.F., Kiechle G. 2009. Metaheuristics for the bi-objective orienteering problem. Swarm Intelligence, Volume 3, Issue 3, pp 179-201.")]
[11193]40  [StorableClass]
[11307]41  public sealed class OrienteeringLocalImprovementOperator : SingleSuccessorOperator, ILocalImprovementOperator {
[11226]42
[11193]43    #region Parameter Properties
44    public ILookupParameter<IntegerVector> IntegerVectorParameter {
[11226]45      get { return (ILookupParameter<IntegerVector>)Parameters["OrienteeringSolution"]; }
[11193]46    }
47    public ILookupParameter<DistanceMatrix> DistanceMatrixParameter {
48      get { return (ILookupParameter<DistanceMatrix>)Parameters["DistanceMatrix"]; }
49    }
50    public ILookupParameter<DoubleArray> ScoresParameter {
51      get { return (ILookupParameter<DoubleArray>)Parameters["Scores"]; }
52    }
53    public ILookupParameter<DoubleValue> MaximumDistanceParameter {
54      get { return (ILookupParameter<DoubleValue>)Parameters["MaximumDistance"]; }
55    }
56    public ILookupParameter<IntValue> StartingPointParameter {
57      get { return (ILookupParameter<IntValue>)Parameters["StartingPoint"]; }
58    }
[11319]59    public ILookupParameter<IntValue> TerminalPointParameter {
60      get { return (ILookupParameter<IntValue>)Parameters["TerminalPoint"]; }
[11193]61    }
[11320]62    public ILookupParameter<DoubleValue> PointVisitingCostsParameter {
63      get { return (ILookupParameter<DoubleValue>)Parameters["PointVisitingCosts"]; }
[11193]64    }
65    #region ILocalImprovementOperator Parameters
[11242]66    public IValueLookupParameter<IntValue> LocalIterationsParameter {
67      get { return (IValueLookupParameter<IntValue>)Parameters["LocalIterations"]; }
68    }
[11193]69    public IValueLookupParameter<IntValue> MaximumIterationsParameter {
70      get { return (IValueLookupParameter<IntValue>)Parameters["MaximumIterations"]; }
71    }
72    public ILookupParameter<IntValue> EvaluatedSolutionsParameter {
73      get { return (ILookupParameter<IntValue>)Parameters["EvaluatedSolutions"]; }
74    }
75    public ILookupParameter<ResultCollection> ResultsParameter {
76      get { return (ILookupParameter<ResultCollection>)Parameters["Results"]; }
77    }
78    #endregion
[11237]79    public ILookupParameter<DoubleValue> QualityParameter {
80      get { return (ILookupParameter<DoubleValue>)Parameters["Quality"]; }
81    }
[11242]82    public IValueParameter<IntValue> MaximumBlockLengthParmeter {
83      get { return (IValueParameter<IntValue>)Parameters["MaximumBlockLength"]; }
84    }
85    public IValueParameter<BoolValue> UseMaximumBlockLengthParmeter {
86      get { return (IValueParameter<BoolValue>)Parameters["UseMaximumBlockLength"]; }
87    }
[11193]88    #endregion
89
90    [StorableConstructor]
91    private OrienteeringLocalImprovementOperator(bool deserializing) : base(deserializing) { }
92    private OrienteeringLocalImprovementOperator(OrienteeringLocalImprovementOperator original, Cloner cloner)
93      : base(original, cloner) {
94    }
95    public OrienteeringLocalImprovementOperator()
96      : base() {
[11226]97      Parameters.Add(new LookupParameter<IntegerVector>("OrienteeringSolution", "The Orienteering Solution given in path representation."));
[11193]98      Parameters.Add(new LookupParameter<DistanceMatrix>("DistanceMatrix", "The matrix which contains the distances between the points."));
99      Parameters.Add(new LookupParameter<DoubleArray>("Scores", "The scores of the points."));
100      Parameters.Add(new LookupParameter<DoubleValue>("MaximumDistance", "The maximum distance constraint for a Orienteering solution."));
101      Parameters.Add(new LookupParameter<IntValue>("StartingPoint", "Index of the starting point."));
[11319]102      Parameters.Add(new LookupParameter<IntValue>("TerminalPoint", "Index of the ending point."));
[11320]103      Parameters.Add(new LookupParameter<DoubleValue>("PointVisitingCosts", "The costs for visiting a point."));
[11228]104
[11242]105      Parameters.Add(new ValueLookupParameter<IntValue>("LocalIterations", "The number of iterations that have already been performed.", new IntValue(0)));
[11193]106      Parameters.Add(new ValueLookupParameter<IntValue>("MaximumIterations", "The maximum number of generations which should be processed.", new IntValue(150)));
107      Parameters.Add(new LookupParameter<IntValue>("EvaluatedSolutions", "The number of evaluated moves."));
108      Parameters.Add(new LookupParameter<ResultCollection>("Results", "The name of the collection where the results are stored."));
[11237]109      Parameters.Add(new LookupParameter<DoubleValue>("Quality", "The quality value of the solution."));
[11242]110
111      Parameters.Add(new ValueParameter<IntValue>("MaximumBlockLength", "The maximum length of the 2-opt shortening.", new IntValue(30)));
[11245]112      Parameters.Add(new ValueParameter<BoolValue>("UseMaximumBlockLength", "Use a limitation of the length for the 2-opt shortening.", new BoolValue(false)));
[11193]113    }
114
115    public override IDeepCloneable Clone(Cloner cloner) {
116      return new OrienteeringLocalImprovementOperator(this, cloner);
117    }
118
119    public override IOperation Apply() {
120      int numPoints = ScoresParameter.ActualValue.Length;
121      var distances = DistanceMatrixParameter.ActualValue;
122      var scores = ScoresParameter.ActualValue;
[11320]123      double pointVisitingCosts = PointVisitingCostsParameter.ActualValue.Value;
[11228]124      double maxLength = MaximumDistanceParameter.ActualValue.Value;
[11242]125      int maxIterations = MaximumIterationsParameter.ActualValue.Value;
126      int maxBlockLength = MaximumBlockLengthParmeter.Value.Value;
127      bool useMaxBlockLength = UseMaximumBlockLengthParmeter.Value.Value;
[11193]128
[12721]129      bool solutionChanged = true;
[11193]130
131      var tour = IntegerVectorParameter.ActualValue.ToList();
132
[11323]133      double tourLength = 0;
[11228]134      double tourScore = tour.Sum(point => scores[point]);
[11193]135
[11242]136      var localIterations = LocalIterationsParameter.ActualValue;
137      var evaluatedSolutions = EvaluatedSolutionsParameter.ActualValue;
138      int evaluations = 0;
139
[11193]140      // Check if the tour can be improved by adding or replacing points
[12721]141      while (solutionChanged && localIterations.Value < maxIterations) {
142        solutionChanged = false;
[11193]143
[11323]144        if (localIterations.Value == 0)
145          tourLength = distances.CalculateTourLength(tour, pointVisitingCosts);
146
[11193]147        // Try to shorten the path
[11242]148        ShortenPath(tour, distances, maxBlockLength, useMaxBlockLength, ref tourLength, ref evaluations);
[11193]149
150        // Determine all points that have not yet been visited by this tour
151        var visitablePoints = Enumerable.Range(0, numPoints).Except(tour).ToList();
152
153        // Determine if any of the visitable points can be included at any position within the tour
154        IncludeNewPoints(tour, visitablePoints,
[11320]155          distances, pointVisitingCosts, maxLength, scores,
[12721]156          ref tourLength, ref tourScore, ref evaluations, ref solutionChanged);
[11193]157
158        // Determine if any of the visitable points can take the place of an already visited point in the tour to improve the scores
159        ReplacePoints(tour, visitablePoints,
160          distances, maxLength, scores,
[12721]161          ref tourLength, ref tourScore, ref evaluations, ref solutionChanged);
[11242]162
163        localIterations.Value++;
[11193]164      }
165
[11242]166      localIterations.Value = 0;
167      evaluatedSolutions.Value += evaluations;
168
[11228]169      // Set new tour
[11193]170      IntegerVectorParameter.ActualValue = new IntegerVector(tour.ToArray());
[11237]171      QualityParameter.ActualValue.Value = tourScore;
[11193]172
173      return base.Apply();
174    }
175
[11242]176    private void ShortenPath(List<int> tour, DistanceMatrix distances, int maxBlockLength, bool useMaxBlockLength, ref double tourLength, ref int evaluations) {
[12721]177      bool solutionChanged;
[11193]178      int pathSize = tour.Count;
[11242]179      maxBlockLength = (useMaxBlockLength && (pathSize > maxBlockLength + 1)) ? maxBlockLength : (pathSize - 2);
[11193]180
181      // Perform a 2-opt
[12721]182      do {
183        solutionChanged = false;
[11193]184
185        for (int blockLength = 2; blockLength < maxBlockLength; blockLength++) {
186          // If an optimization has been done, start from the beginning
[12721]187          if (solutionChanged) break;
[11193]188
189          for (int position = 1; position < (pathSize - blockLength); position++) {
190            // If an optimization has been done, start from the beginning
[12721]191            if (solutionChanged) break;
[11193]192
[11242]193            evaluations++;
194
[11193]195            double newLength = tourLength;
[11228]196            // Recalculate length of whole swapped part, in case distances are not symmetric
197            for (int index = position - 1; index < position + blockLength; index++) newLength -= distances[tour[index], tour[index + 1]];
198            for (int index = position + blockLength - 1; index > position; index--) newLength += distances[tour[index], tour[index - 1]];
[11193]199            newLength += distances[tour[position - 1], tour[position + blockLength - 1]];
200            newLength += distances[tour[position], tour[position + blockLength]];
201
[12721]202            if (newLength < tourLength - 0.00001) {
203              // Avoid cycling caused by precision
[11228]204              var reversePart = tour.GetRange(position, blockLength).AsEnumerable().Reverse();
[11193]205
206              tour.RemoveRange(position, blockLength);
[11228]207              tour.InsertRange(position, reversePart);
[11193]208
209              tourLength = newLength;
210
211              // Re-run the optimization
[12721]212              solutionChanged = true;
[11193]213            }
214          }
215        }
[12721]216      } while (solutionChanged);
[11193]217    }
[12721]218
[11193]219    private void IncludeNewPoints(List<int> tour, List<int> visitablePoints,
[11320]220      DistanceMatrix distances, double pointVisitingCosts, double maxLength, DoubleArray scores,
[12721]221      ref double tourLength, ref double tourScore, ref int evaluations, ref bool solutionChanged) {
[11193]222
223      for (int tourPosition = 1; tourPosition < tour.Count; tourPosition++) {
224        // If an optimization has been done, start from the beginning
[12721]225        if (solutionChanged) break;
[11193]226
227        for (int i = 0; i < visitablePoints.Count; i++) {
228          // If an optimization has been done, start from the beginning
[12721]229          if (solutionChanged) break;
[11193]230
[11242]231          evaluations++;
232
[11320]233          double detour = distances.CalculateInsertionCosts(tour, tourPosition, visitablePoints[i], pointVisitingCosts);
[11193]234
235          // Determine if including the point does not violate any constraint
236          if (tourLength + detour <= maxLength) {
237            // Insert the new point at this position
238            tour.Insert(tourPosition, visitablePoints[i]);
239
[11237]240            // Update the overall tour tourLength and score
[11193]241            tourLength += detour;
[11237]242            tourScore += scores[visitablePoints[i]];
[11193]243
244            // Re-run this optimization
[12721]245            solutionChanged = true;
[11193]246          }
247        }
248      }
249    }
[12721]250
[11193]251    private void ReplacePoints(List<int> tour, List<int> visitablePoints,
252      DistanceMatrix distances, double maxLength, DoubleArray scores,
[12721]253      ref double tourLength, ref double tourScore, ref int evaluations, ref bool solutionChanged) {
[11193]254
255      for (int tourPosition = 1; tourPosition < tour.Count - 1; tourPosition++) {
256        // If an optimization has been done, start from the beginning
[12721]257        if (solutionChanged) break;
[11193]258
259        for (int i = 0; i < visitablePoints.Count; i++) {
260          // If an optimization has been done, start from the beginning
[12721]261          if (solutionChanged) break;
[11193]262
[11242]263          evaluations++;
264
[11194]265          double detour = distances.CalculateReplacementCosts(tour, tourPosition, visitablePoints[i]);
[11193]266
267          double oldPointScore = scores[tour[tourPosition]];
268          double newPointScore = scores[visitablePoints[i]];
269
[11228]270          if ((tourLength + detour <= maxLength) && (newPointScore > oldPointScore)) {
[11193]271            // Replace the old point by the new one
272            tour[tourPosition] = visitablePoints[i];
273
274            // Update the overall tour tourLength
275            tourLength += detour;
276
277            // Update the scores achieved by visiting this point
278            tourScore += newPointScore - oldPointScore;
279
280            // Re-run this optimization
[12721]281            solutionChanged = true;
[11193]282          }
283        }
284      }
285    }
286  }
287}
Note: See TracBrowser for help on using the repository browser.