source: branches/PTSP/HeuristicLab.Problems.PTSP/3.3/MoveEvaluators/TwoOpt/PTSPAnalyticalInversionMovePathEvaluator.cs @ 12219

#region License Information
/* HeuristicLab
 * Copyright (C) 2002-2015 Heuristic and Evolutionary Algorithms Laboratory (HEAL)
 *
 * This file is part of HeuristicLab.
 *
 * HeuristicLab is free software: you can redistribute it and/or modify
 * it under the terms of the GNU General Public License as published by
 * the Free Software Foundation, either version 3 of the License, or
 * (at your option) any later version.
 *
 * HeuristicLab is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU General Public License for more details.
 *
 * You should have received a copy of the GNU General Public License
 * along with HeuristicLab. If not, see <http://www.gnu.org/licenses/>.
 */
#endregion

using HeuristicLab.Common;
using HeuristicLab.Core;
using HeuristicLab.Data;
using HeuristicLab.Encodings.PermutationEncoding;
using HeuristicLab.Parameters;
using HeuristicLab.Persistence.Default.CompositeSerializers.Storable;
using System;
namespace HeuristicLab.Problems.PTSP {
  /// <summary>
  /// An operator to evaluate inversion moves (2-opt).
  /// </summary>
  [Item("PTSPAnalyticalInversionMovePathEvaluator", "Evaluates an inversion move (2-opt) of the PTSP in with the closed form expression")]
  [StorableClass]
  public class PTSPAnalyticalInversionMovePathEvaluator : PTSPPathMoveEvaluator, IPermutationInversionMoveOperator {

    private static ItemList<DoubleValue> probabilities;
    private static DoubleMatrix A;
    private static DoubleMatrix B;

    public override Type EvaluatorType {
      get { return typeof(PTSPAnalyticalInversionMovePathEvaluator); }
    }
    public ILookupParameter<InversionMove> InversionMoveParameter {
      get { return (ILookupParameter<InversionMove>)Parameters["InversionMove"]; }
    }

    public IValueParameter<ItemList<DoubleValue>> ProbabilitiesParameter {
      get { return (IValueParameter<ItemList<DoubleValue>>)Parameters["Probabilities"]; }
    }

    public IValueParameter<DoubleMatrix> AParameter {
      get { return (IValueParameter<DoubleMatrix>)Parameters["A"]; }
    }

    public IValueParameter<DoubleMatrix> BParameter {
      get { return (IValueParameter<DoubleMatrix>)Parameters["B"]; }
    }

    [StorableConstructor]
    protected PTSPAnalyticalInversionMovePathEvaluator(bool deserializing) : base(deserializing) { }
    protected PTSPAnalyticalInversionMovePathEvaluator(PTSPAnalyticalInversionMovePathEvaluator original, Cloner cloner) : base(original, cloner) { }
    public PTSPAnalyticalInversionMovePathEvaluator()
      : base() {
      Parameters.Add(new LookupParameter<InversionMove>("InversionMove", "The move to evaluate."));
      Parameters.Add(new ValueParameter<ItemList<DoubleValue>>("Probabilities", "The probabilities of the current instance"));
      Parameters.Add(new ValueParameter<DoubleMatrix>("A", "The matrix A for delta evaluation"));
      Parameters.Add(new ValueParameter<DoubleMatrix>("B", "The matrix B for delta evaluation"));
    }

    public override IDeepCloneable Clone(Cloner cloner) {
      return new PTSPAnalyticalInversionMovePathEvaluator(this, cloner);
    }

    public static double EvaluateByCoordinates(Permutation permutation, InversionMove move, DoubleMatrix coordinates, PTSPAnalyticalInversionMovePathEvaluator evaluator) {
      int edge1source = permutation.GetCircular(move.Index1 - 1);
      int edge1target = permutation[move.Index1];
      int edge2source = permutation[move.Index2];
      int edge2target = permutation.GetCircular(move.Index2 + 1);
      if (move.Index2 - move.Index1 >= permutation.Length - 2) return 0;
      double moveQuality = 0;
      // remove two edges
      moveQuality -= evaluator.CalculateDistance(coordinates[edge1source, 0], coordinates[edge1source, 1],
            coordinates[edge1target, 0], coordinates[edge1target, 1]);
      moveQuality -= evaluator.CalculateDistance(coordinates[edge2source, 0], coordinates[edge2source, 1],
        coordinates[edge2target, 0], coordinates[edge2target, 1]);
      // add two edges
      moveQuality += evaluator.CalculateDistance(coordinates[edge1source, 0], coordinates[edge1source, 1],
        coordinates[edge2source, 0], coordinates[edge2source, 1]);
      moveQuality += evaluator.CalculateDistance(coordinates[edge1target, 0], coordinates[edge1target, 1],
        coordinates[edge2target, 0], coordinates[edge2target, 1]);
      return moveQuality;
    }

    public static double EvaluateByDistanceMatrix(Permutation permutation, InversionMove move, DistanceMatrix distanceMatrix) {
      int i = move.Index1;
      int j = move.Index2;
      return RecursiveExpectedCost(1, i, j) + RecursiveExpectedCost(2, i, j) - RecursiveExpectedCost(3, i, j);
    }

    protected static double RecursiveExpectedCost(int s, int i, int j) {
      switch (s) {
        case 1:
          if (j == i + 1) {
            return (1/(1-probabilities[i+1].Value))*A[i,2]+(1 - probabilities[i].Value)*(A[i+1,1]-A[i+1,probabilities.Capacity-1]);
          } else if (i == j) {
            return A[i,1];
          } else {
            // Equation 25
            return ((1 - probabilities[i].Value) / (1 - probabilities[j].Value)) * RecursiveExpectedCost(1, i + 1, j - 1) + (1 - probabilities[i].Value) * (1);
          }
          
        case 2:
          if (j == i + 1) {
            return (1 - probabilities[i + 1].Value) * (B[i, 1] - B[i, probabilities.Capacity - 1]) + (1 / (1 - probabilities[i].Value)) * (B[i + 1, 2]);
          } else if (i == j) {
            return B[i,1];
          } else {
            return 0;
            // Equation 26
          }

        case 3:
          if (j == i + 1) {
            return A[i, 2] + A[i + 1, 1] - A[i + 1, probabilities.Capacity - 1] + B[i, 1] - B[i, probabilities.Capacity - 1] + B[i + 1, 2];
          } else if (i == j) {
            return A[i,1]+B[i,1];
          } else {
            return 0;
            // Equation 27
          }
        default:
          return 0;
      }
    }

    private double Q(int i, int j, DoubleArray probabilities) {
      double prod = 1;
      for (int k = i; k <= j; k++) {
        prod *= (1 - probabilities[k]);
      }
      return prod;
    }

    private double Q_hat(int i, int j, DoubleArray probabilities) {
      double prod = 1;
      for (int k = i; k <= j; k++) {
        prod *= (1-probabilities[k]);
      }
      return prod;
    }

    protected override double EvaluateByCoordinates(Permutation permutation, DoubleMatrix coordinates) {
      return EvaluateByCoordinates(permutation, InversionMoveParameter.ActualValue, coordinates, this);
    }

    protected override double EvaluateByDistanceMatrix(Permutation permutation, DistanceMatrix distanceMatrix) {
      A = AParameter.Value;
      B = BParameter.Value;
      probabilities = ProbabilitiesParameter.Value;
      return EvaluateByDistanceMatrix(permutation, InversionMoveParameter.ActualValue, distanceMatrix);
    }

    protected override double CalculateDistance(double x1, double y1, double x2, double y2) {
      return Math.Round(Math.Sqrt((x1 - x2) * (x1 - x2) + (y1 - y2) * (y1 - y2)));
    }
  }
}