source: branches/VRP/HeuristicLab.Problems.VehicleRouting/3.4/Encodings/Alba/Creators/PushForwardInsertionCreator.cs @ 4365

#region License Information
/* HeuristicLab
 * Copyright (C) 2002-2010 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 System;
using System.Collections.Generic;
using HeuristicLab.Core;
using HeuristicLab.Data;
using HeuristicLab.Encodings.PermutationEncoding;
using HeuristicLab.Optimization;
using HeuristicLab.Parameters;
using HeuristicLab.Persistence.Default.CompositeSerializers.Storable;
using HeuristicLab.Problems.VehicleRouting.Variants;

namespace HeuristicLab.Problems.VehicleRouting.Encodings.Alba {
  [Item("PushForwardCreator", "The push forward insertion heuristic.  It is implemented as described in Sam, and Thangiah, R. (1999). A Hybrid Genetic Algorithms, Simulated Annealing and Tabu Search Heuristic for Vehicle Routing Problems with Time Windows. Practical Handbook of Genetic Algorithms, Volume III, pp 347–381.")]
  [StorableClass]
  public sealed class PushForwardCreator : DefaultRepresentationCreator, IStochasticOperator, ITimeWindowedOperator {
    #region IStochasticOperator Members
    public ILookupParameter<IRandom> RandomParameter {
      get { return (LookupParameter<IRandom>)Parameters["Random"]; }
    }
    #endregion

    public IValueParameter<DoubleValue> Alpha {
      get { return (IValueParameter<DoubleValue>)Parameters["Alpha"]; }
    }
    public IValueParameter<DoubleValue> AlphaVariance {
      get { return (IValueParameter<DoubleValue>)Parameters["AlphaVariance"]; }
    }
    public IValueParameter<DoubleValue> Beta {
      get { return (IValueParameter<DoubleValue>)Parameters["Beta"]; }
    }
    public IValueParameter<DoubleValue> BetaVariance {
      get { return (IValueParameter<DoubleValue>)Parameters["BetaVariance"]; }
    }
    public IValueParameter<DoubleValue> Gamma {
      get { return (IValueParameter<DoubleValue>)Parameters["Gamma"]; }
    }
    public IValueParameter<DoubleValue> GammaVariance {
      get { return (IValueParameter<DoubleValue>)Parameters["GammaVariance"]; }
    }

    [StorableConstructor]
    private PushForwardCreator(bool deserializing) : base(deserializing) { }

    public PushForwardCreator()
      : base() {
      Parameters.Add(new LookupParameter<IRandom>("Random", "The pseudo random number generator."));
      Parameters.Add(new ValueParameter<DoubleValue>("Alpha", "The alpha value.", new DoubleValue(0.7)));
      Parameters.Add(new ValueParameter<DoubleValue>("AlphaVariance", "The alpha variance.", new DoubleValue(0.5)));
      Parameters.Add(new ValueParameter<DoubleValue>("Beta", "The beta value.", new DoubleValue(0.1)));
      Parameters.Add(new ValueParameter<DoubleValue>("BetaVariance", "The beta variance.", new DoubleValue(0.07)));
      Parameters.Add(new ValueParameter<DoubleValue>("Gamma", "The gamma value.", new DoubleValue(0.2)));
      Parameters.Add(new ValueParameter<DoubleValue>("GammaVariance", "The gamma variance.", new DoubleValue(0.14)));
    }

    // use the Box-Mueller transform in the polar form to generate a N(0,1) random variable out of two uniformly distributed random variables
    private double Gauss(IRandom random) {
      double u = 0.0, v = 0.0, s = 0.0;
      do {
        u = (random.NextDouble() * 2) - 1;
        v = (random.NextDouble() * 2) - 1;
        s = Math.Sqrt(u * u + v * v);
      } while (s < Double.Epsilon || s > 1);
      return u * Math.Sqrt((-2.0 * Math.Log(s)) / s);
    }

    private double N(double mu, double sigma, IRandom random) {
      return mu + (sigma * Gauss(random)); // transform the random variable sampled from N(0,1) to N(mu,sigma)
    }

    private double Distance(int start, int end) {
      return ProblemInstance.GetDistance(start, end);
    }

    private double TravelDistance(List<int> route, int begin) {
      double distance = 0;
      for (int i = begin; i < route.Count - 1 && (i == begin || route[i] != 0); i++) {
        distance += Distance(route[i], route[i + 1]);
      }
      return distance;
    }

    private bool SubrouteConstraintsOK(List<int> route, int begin) {     
      Tour subroute = new Tour();
      subroute.Stops.AddRange(route.GetRange(begin,
          route.Count - begin));

      return ProblemInstance.Feasible(subroute);
    }

    protected override List<int> CreateSolution() {
      double alpha, beta, gamma;
      alpha = N(Alpha.Value.Value, Math.Sqrt(AlphaVariance.Value.Value), RandomParameter.ActualValue);
      beta = N(Beta.Value.Value, Math.Sqrt(BetaVariance.Value.Value), RandomParameter.ActualValue);
      gamma = N(Gamma.Value.Value, Math.Sqrt(GammaVariance.Value.Value), RandomParameter.ActualValue);

      double x0 = ProblemInstance.Coordinates[0, 0];
      double y0 = ProblemInstance.Coordinates[0, 1];
      double distance = 0;
      double cost = 0;
      double minimumCost = double.MaxValue;
      List<int> unroutedList = new List<int>();
      List<double> costList = new List<double>();
      int index;
      int indexOfMinimumCost = -1;
      int indexOfCustomer = -1;

      /*-----------------------------------------------------------------------------
       * generate cost list
       *-----------------------------------------------------------------------------
       */
      for (int i = 1; i <= ProblemInstance.Cities.Value; i++) {
        distance = Distance(i, 0);
        if (ProblemInstance.Coordinates[i, 0] < x0) distance = -distance;

        cost = -alpha * distance + // distance 0 <-> City[i]
               beta * (ProblemInstance as ITimeWindowedProblemInstance).DueTime[i] + // latest arrival time
               gamma * (Math.Asin((ProblemInstance.Coordinates[i, 1] - y0) / distance) / 360 * distance); // polar angle

        index = 0;
        while (index < costList.Count && costList[index] < cost) index++;
        costList.Insert(index, cost);
        unroutedList.Insert(index, i);
      }

      /*------------------------------------------------------------------------------
       * route customers according to cost list
       *------------------------------------------------------------------------------
       */
      int routeIndex = 0;
      int currentRoute = 0;
      int c;
      int customer = -1;
      int subTourCount = 1;
      List<int> route = new List<int>(ProblemInstance.Cities.Value + ProblemInstance.Vehicles.Value - 1);
      minimumCost = double.MaxValue;
      indexOfMinimumCost = -1;
      route.Add(0);
      route.Add(0);
      route.Insert(1, unroutedList[0]);
      unroutedList.RemoveAt(0);
      currentRoute = routeIndex;
      routeIndex++;

      do {
        for (c = 0; c < unroutedList.Count; c++) {
          for (int i = currentRoute + 1; i < route.Count; i++) {
            route.Insert(i, (int)unroutedList[c]);
            if (route[currentRoute] != 0) { throw new Exception("currentRoute not depot"); }
            cost = TravelDistance(route, currentRoute);
            if (cost < minimumCost && SubrouteConstraintsOK(route, currentRoute)) {
              minimumCost = cost;
              indexOfMinimumCost = i;
              customer = (int)unroutedList[c];
              indexOfCustomer = c;
            }
            route.RemoveAt(i);
          }
        }
        // insert customer if found
        if (indexOfMinimumCost != -1) {
          route.Insert(indexOfMinimumCost, customer);
          routeIndex++;
          unroutedList.RemoveAt(indexOfCustomer);
          costList.RemoveAt(indexOfCustomer);
        } else { // no feasible customer found
          routeIndex++;
          route.Insert(routeIndex, 0);
          currentRoute = routeIndex;
          route.Insert(route.Count - 1, (int)unroutedList[0]);
          unroutedList.RemoveAt(0);
          routeIndex++;
          subTourCount++;
        }
        // reset minimum    
        minimumCost = double.MaxValue;
        indexOfMinimumCost = -1;
        indexOfCustomer = -1;
        customer = -1;
      } while (unroutedList.Count > 0);
      while (route.Count < ProblemInstance.Cities.Value + ProblemInstance.Vehicles.Value)
        route.Add(0);

      return route;
    }
  }
}