CapacitatedVehicleRoutingProblemWithTimeWindows.java

//
// Copyright 2012 Google
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
//     http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
import com.google.ortools.constraintsolver.Assignment;
import com.google.ortools.constraintsolver.FirstSolutionStrategy;
import com.google.ortools.constraintsolver.IntVar;
import com.google.ortools.constraintsolver.RoutingDimension;
import com.google.ortools.constraintsolver.RoutingIndexManager;
import com.google.ortools.constraintsolver.RoutingModel;
import com.google.ortools.constraintsolver.RoutingSearchParameters;
import com.google.ortools.constraintsolver.main;
import java.util.ArrayList;
import java.util.List;
import java.util.Random;
import java.util.function.LongBinaryOperator;
import java.util.function.LongUnaryOperator;
import java.util.logging.Logger;

// A pair class

class Pair<K, V> {
  final K first;
  final V second;

  public static <K, V> Pair<K, V> of(K element0, V element1) {
    return new Pair<K, V>(element0, element1);
  }

  public Pair(K element0, V element1) {
    this.first = element0;
    this.second = element1;
  }
}

/**
 * Sample showing how to model and solve a capacitated vehicle routing problem with time windows
 * using the swig-wrapped version of the vehicle routing library in src/constraint_solver.
 */
public class CapacitatedVehicleRoutingProblemWithTimeWindows {

  static {
    System.loadLibrary("jniortools");
  }

  private static Logger logger =
      Logger.getLogger(CapacitatedVehicleRoutingProblemWithTimeWindows.class.getName());

  // Locations representing either an order location or a vehicle route
  // start/end.
  private List<Pair<Integer, Integer>> locations = new ArrayList();

  // Quantity to be picked up for each order.
  private List<Integer> orderDemands = new ArrayList();
  // Time window in which each order must be performed.
  private List<Pair<Integer, Integer>> orderTimeWindows = new ArrayList();
  // Penalty cost "paid" for dropping an order.
  private List<Integer> orderPenalties = new ArrayList();

  // Capacity of the vehicles.
  private int vehicleCapacity = 0;
  // Latest time at which each vehicle must end its tour.
  private List<Integer> vehicleEndTime = new ArrayList();
  // Cost per unit of distance of each vehicle.
  private List<Integer> vehicleCostCoefficients = new ArrayList();
  // Vehicle start and end indices. They have to be implemented as int[] due
  // to the available SWIG-ed interface.
  private int vehicleStarts[];
  private int vehicleEnds[];

  // Random number generator to produce data.
  private final Random randomGenerator = new Random(0xBEEF);

  /**
   * Creates a Manhattan Distance evaluator with 'costCoefficient'.
   *
   * @param manager Node Index Manager.
   * @param costCoefficient The coefficient to apply to the evaluator.
   */
  private LongBinaryOperator buildManhattanCallback(RoutingIndexManager manager, int costCoefficient) {
    return new LongBinaryOperator() {
      public long applyAsLong(long firstIndex, long secondIndex) {
        try {
          int firstNode = manager.indexToNode(firstIndex);
          int secondNode = manager.indexToNode(secondIndex);
          Pair<Integer, Integer> firstLocation = locations.get(firstNode);
          Pair<Integer, Integer> secondLocation = locations.get(secondNode);
          return (long) costCoefficient
              * (Math.abs(firstLocation.first - secondLocation.first)
                  + Math.abs(firstLocation.second - secondLocation.second));
        } catch (Throwable throwed) {
          logger.warning(throwed.getMessage());
          return 0;
        }
      }
    };
  }

  /**
   * Creates order data. Location of the order is random, as well as its demand (quantity), time
   * window and penalty.
   *
   * @param numberOfOrders number of orders to build.
   * @param xMax maximum x coordinate in which orders are located.
   * @param yMax maximum y coordinate in which orders are located.
   * @param demandMax maximum quantity of a demand.
   * @param timeWindowMax maximum starting time of the order time window.
   * @param timeWindowWidth duration of the order time window.
   * @param penaltyMin minimum pernalty cost if order is dropped.
   * @param penaltyMax maximum pernalty cost if order is dropped.
   */
  private void buildOrders(
      int numberOfOrders,
      int xMax,
      int yMax,
      int demandMax,
      int timeWindowMax,
      int timeWindowWidth,
      int penaltyMin,
      int penaltyMax) {
    logger.info("Building orders.");
    for (int order = 0; order < numberOfOrders; ++order) {
      locations.add(Pair.of(randomGenerator.nextInt(xMax + 1), randomGenerator.nextInt(yMax + 1)));
      orderDemands.add(randomGenerator.nextInt(demandMax + 1));
      int timeWindowStart = randomGenerator.nextInt(timeWindowMax + 1);
      orderTimeWindows.add(Pair.of(timeWindowStart, timeWindowStart + timeWindowWidth));
      orderPenalties.add(randomGenerator.nextInt(penaltyMax - penaltyMin + 1) + penaltyMin);
    }
  }

  /**
   * Creates fleet data. Vehicle starting and ending locations are random, as well as vehicle costs
   * per distance unit.
   *
   * @param numberOfVehicles
   * @param xMax maximum x coordinate in which orders are located.
   * @param yMax maximum y coordinate in which orders are located.
   * @param endTime latest end time of a tour of a vehicle.
   * @param capacity capacity of a vehicle.
   * @param costCoefficientMax maximum cost per distance unit of a vehicle (mimimum is 1),
   */
  private void buildFleet(
      int numberOfVehicles, int xMax, int yMax, int endTime, int capacity, int costCoefficientMax) {
    logger.info("Building fleet.");
    vehicleCapacity = capacity;
    vehicleStarts = new int[numberOfVehicles];
    vehicleEnds = new int[numberOfVehicles];
    for (int vehicle = 0; vehicle < numberOfVehicles; ++vehicle) {
      vehicleStarts[vehicle] = locations.size();
      locations.add(Pair.of(randomGenerator.nextInt(xMax + 1), randomGenerator.nextInt(yMax + 1)));
      vehicleEnds[vehicle] = locations.size();
      locations.add(Pair.of(randomGenerator.nextInt(xMax + 1), randomGenerator.nextInt(yMax + 1)));
      vehicleEndTime.add(endTime);
      vehicleCostCoefficients.add(randomGenerator.nextInt(costCoefficientMax) + 1);
    }
  }

  /** Solves the current routing problem. */
  private void solve(final int numberOfOrders, final int numberOfVehicles) {
    logger.info(
        "Creating model with " + numberOfOrders + " orders and " + numberOfVehicles + " vehicles.");
    // Finalizing model
    final int numberOfLocations = locations.size();

    RoutingIndexManager manager =
        new RoutingIndexManager(numberOfLocations, numberOfVehicles, vehicleStarts, vehicleEnds);
    RoutingModel model = new RoutingModel(manager);

    // Setting up dimensions
    final int bigNumber = 100000;
    final LongBinaryOperator callback = buildManhattanCallback(manager, 1);
    final String timeStr = "time";
    model.addDimension(
        model.registerTransitCallback(callback), bigNumber, bigNumber, false, timeStr);
    RoutingDimension timeDimension = model.getMutableDimension(timeStr);

    LongUnaryOperator demandCallback =
        new LongUnaryOperator() {
          public long applyAsLong(long index) {
            try {
              int node = manager.indexToNode(index);
              if (node < numberOfOrders) {
                return orderDemands.get(node);
              }
              return 0;
            } catch (Throwable throwed) {
              logger.warning(throwed.getMessage());
              return 0;
            }
          }
        };
    final String capacityStr = "capacity";
    model.addDimension(
        model.registerUnaryTransitCallback(demandCallback), 0, vehicleCapacity, true, capacityStr);
    RoutingDimension capacityDimension = model.getMutableDimension(capacityStr);

    // Setting up vehicles
    LongBinaryOperator[] callbacks = new LongBinaryOperator[numberOfVehicles];
    for (int vehicle = 0; vehicle < numberOfVehicles; ++vehicle) {
      final int costCoefficient = vehicleCostCoefficients.get(vehicle);
      callbacks[vehicle] = buildManhattanCallback(manager, costCoefficient);
      final int vehicleCost = model.registerTransitCallback(callbacks[vehicle]);
      model.setArcCostEvaluatorOfVehicle(vehicleCost, vehicle);
      timeDimension.cumulVar(model.end(vehicle)).setMax(vehicleEndTime.get(vehicle));
    }

    // Setting up orders
    for (int order = 0; order < numberOfOrders; ++order) {
      timeDimension
          .cumulVar(order)
          .setRange(orderTimeWindows.get(order).first, orderTimeWindows.get(order).second);
      long[] orderIndices = {manager.nodeToIndex(order)};
      model.addDisjunction(orderIndices, orderPenalties.get(order));
    }

    // Solving
    RoutingSearchParameters parameters =
        main.defaultRoutingSearchParameters()
            .toBuilder()
            .setFirstSolutionStrategy(FirstSolutionStrategy.Value.ALL_UNPERFORMED)
            .build();

    logger.info("Search");
    Assignment solution = model.solveWithParameters(parameters);

    if (solution != null) {
      String output = "Total cost: " + solution.objectiveValue() + "\n";
      // Dropped orders
      String dropped = "";
      for (int order = 0; order < numberOfOrders; ++order) {
        if (solution.value(model.nextVar(order)) == order) {
          dropped += " " + order;
        }
      }
      if (dropped.length() > 0) {
        output += "Dropped orders:" + dropped + "\n";
      }
      // Routes
      for (int vehicle = 0; vehicle < numberOfVehicles; ++vehicle) {
        String route = "Vehicle " + vehicle + ": ";
        long order = model.start(vehicle);
        // Empty route has a minimum of two nodes: Start => End
        if (model.isEnd(solution.value(model.nextVar(order)))) {
          route += "Empty";
        } else {
          for (; !model.isEnd(order); order = solution.value(model.nextVar(order))) {
            IntVar load = capacityDimension.cumulVar(order);
            IntVar time = timeDimension.cumulVar(order);
            route +=
                order
                    + " Load("
                    + solution.value(load)
                    + ") "
                    + "Time("
                    + solution.min(time)
                    + ", "
                    + solution.max(time)
                    + ") -> ";
          }
          IntVar load = capacityDimension.cumulVar(order);
          IntVar time = timeDimension.cumulVar(order);
          route +=
              order
                  + " Load("
                  + solution.value(load)
                  + ") "
                  + "Time("
                  + solution.min(time)
                  + ", "
                  + solution.max(time)
                  + ")";
        }
        output += route + "\n";
      }
      logger.info(output);
    }
  }

  public static void main(String[] args) throws Exception {
    CapacitatedVehicleRoutingProblemWithTimeWindows problem =
        new CapacitatedVehicleRoutingProblemWithTimeWindows();
    final int xMax = 20;
    final int yMax = 20;
    final int demandMax = 3;
    final int timeWindowMax = 24 * 60;
    final int timeWindowWidth = 4 * 60;
    final int penaltyMin = 50;
    final int penaltyMax = 100;
    final int endTime = 24 * 60;
    final int costCoefficientMax = 3;

    final int orders = 100;
    final int vehicles = 20;
    final int capacity = 50;

    problem.buildOrders(
        orders, xMax, yMax, demandMax, timeWindowMax, timeWindowWidth, penaltyMin, penaltyMax);
    problem.buildFleet(vehicles, xMax, yMax, endTime, capacity, costCoefficientMax);
    problem.solve(orders, vehicles);
  }
}