Initializing help system before first use

Facility location problem - Data as arrays or collections


Type: Facility location
Rating: 2 (easy-medium)
Description: Solve a facility location problem for which data is given as collections.
File(s): FacilityLocationArray.cpp, FacilityLocationCollection.cpp


FacilityLocationArray.cpp
// (c) 2024-2024 Fair Isaac Corporation

/**
 * This example demonstrates some modeling devices. We model a very simple
 * facility location problem: We have customers and facilities. The constraints
 * are: - each customer must be served from exactly one facility - customers can
 * only be served from open facilities - customer demand must be satisfied -
 * facility capacity must not be exceeded We minimize the sum of transport cost
 * (between customer and facility) and the cost for opening a facility. In this
 * example data is kept in arrays.
 */

#include <iostream>
#include <xpress.hpp>

using namespace xpress;
using namespace xpress::objects;
using xpress::objects::utils::sum;

/** Customer description. */
struct Customer {
  /** Name of customer. */
  std::string name;
  /** Demand for customer. */
  double demand;

  Customer(std::string const &name, double demand)
      : name(name), demand(demand) {}
};

/** Facility descriptor. */
struct Facility {
  /** Name of facility. */
  std::string name;
  /** Capacity of facility. */
  double capacity;
  /** Cost for opening this facility. */
  double cost;

  Facility(std::string const &name, double capacity, double cost)
      : name(name), capacity(capacity), cost(cost) {}
};

/** Customers in this example. */
std::vector<Customer> const customers{Customer{"Customer 1", 80},
                                      Customer{"Customer 2", 270},
                                      Customer{"Customer 3", 250}};
/** Facilities in this example. */
std::vector<Facility> const facilities{Facility{"Facility 1", 500, 1000},
                                       Facility{"Facility 2", 500, 1000},
                                       Facility{"Facility 3", 500, 1000}};
/** Cost for transporting one unit between customer and facility. */
std::vector<std::vector<double>> transportCost{std::vector<double>{4, 5, 6},
                                               std::vector<double>{6, 4, 3},
                                               std::vector<double>{9, 7, 4}};

int main() {
  XpressProblem prob;
  std::vector<Variable> y =
      prob.addVariables(facilities.size())
          .withType(ColumnType::Binary)
          .withName([&](auto f) { return facilities[f].name; })
          .toArray();
  std::vector<std::vector<Variable>> x =
      prob.addVariables(facilities.size(),
                        customers.size())
          .withName([&](auto f, auto c) {
            return xpress::format("x[%s,%s]", facilities[f].name.c_str(),
                                  customers[c].name.c_str());
          })
          .toArray();

  // for each customer c
  // sum(f=1..m) x[f,c] = d
  prob.addConstraints(customers.size(), [&](auto c) {
    return sum(facilities.size(), [&](auto f) { return x[f][c]; }) ==
           customers[c].demand;
  });

  // for each facility f
  // sum(c=1..n) x[f,c] <= capacity[j] * y[f]
  prob.addConstraints(facilities.size(), [&](auto f) {
    return sum(x[f]) <= facilities[f].capacity * y[f];
  });

  // minimize sum(f=1..m) cost[f] * y[f] +
  // sum(c=1..n) sum(f=1..m) cost[f,c] * x[f,c]
  prob.setObjective(sum(facilities.size(),
                        [&](auto f) { return facilities[f].cost * y[f]; }) +
                    sum(customers.size(), [&](auto c) {
                      return sum(facilities.size(),
                                 [&](auto f) {
                        return transportCost[f][c] * x[f][c];
                      });
                    }));

  prob.writeProb("facilitylocationarray.lp", "l");

  prob.optimize();
  if (prob.attributes.getSolStatus() != SolStatus::Optimal)
    throw std::runtime_error("failed to optimize with status " +
                             to_string(prob.attributes.getSolStatus()));
  auto sol = prob.getSolution();
  for (unsigned f = 0; f < facilities.size(); ++f) {
    if (y[f].getValue(sol) > 0.5) {
      std::cout << "Facility " << facilities[f].name << " is open, serves"
                << std::endl;
      for (unsigned c = 0; c < customers.size(); ++c) {
        if (x[f][c].getValue(sol) > 0.0)
          std::cout << "  " << customers[c].name << ": "
                    << x[f][c].getValue(sol) << std::endl;
      }
    }
  }
  return 0;
}

FacilityLocationCollection.cpp
// (c) 2024-2024 Fair Isaac Corporation

/**
 * This example demonstrates some modeling devices. We model a very simple
 * facility location problem: We have customers and facilities. The constraints
 * are: - each customer must be served from exactly one facility - customers can
 * only be served from open facilities - customer demand must be satisfied -
 * facility capacity must not be exceeded We minimize the sum of transport cost
 * (between customer and facility) and the cost for opening a facility. In this
 * example data is kept in collections.
 */

#include <iostream>
#include <list>
#include <xpress.hpp>

using namespace xpress;
using namespace xpress::objects;
using namespace xpress::maps;
using xpress::objects::utils::sum;

/** Customer description. */
struct Customer {
  /** Name of customer. */
  std::string name;
  /** Demand for customer. */
  double demand;

  Customer(std::string const &name = "", double demand = 0.0)
      : name(name), demand(demand) {}

  bool operator==(Customer const &other) const { return name == other.name; }
};

/** Facility descriptor. */
struct Facility {
  /** Name of facility. */
  std::string name;
  /** Capacity of facility. */
  double capacity;
  /** Cost for opening this facility. */
  double cost;

  Facility(std::string const &name = "", double capacity = 0.0,
           double cost = 0.0)
      : name(name), capacity(capacity), cost(cost) {}

  bool operator==(Facility const &other) const { return name == other.name; }
};

// Specialize std::hash for Customer and Facility so that we can use the
// classes as keys in hash maps.
namespace std {
template <> struct hash<Customer> {
  size_t operator()(Customer const &c) const {
    return hash<std::string>()(c.name);
  }
};
template <> struct hash<Facility> {
  size_t operator()(Facility const &f) const {
    return hash<std::string>()(f.name);
  }
};
} // namespace std

/** Customers in this example. */
std::list<Customer> customers{Customer("Customer 1", 80),
                              Customer("Customer 2", 270),
                              Customer("Customer 3", 250)};
/** Facilities in this example. */
std::list<Facility> facilities{Facility("Facility 1", 500, 1000),
                               Facility("Facility 2", 500, 1000),
                               Facility("Facility 3", 500, 1000)};

/** Cost for transporting one unit between customer and facility. */
HashMap2<Facility, Customer, double> transportCost;
void initTransportCost() {
  std::vector<Customer> c(customers.begin(), customers.end());
  std::vector<Facility> f(facilities.begin(), facilities.end());
  transportCost.put(f[0], c[0], 4.0);
  transportCost.put(f[0], c[1], 5.0);
  transportCost.put(f[0], c[2], 6.0);
  transportCost.put(f[1], c[0], 6.0);
  transportCost.put(f[1], c[1], 4.0);
  transportCost.put(f[1], c[2], 3.0);
  transportCost.put(f[2], c[0], 9.0);
  transportCost.put(f[2], c[1], 7.0);
  transportCost.put(f[2], c[2], 4.0);
}

int main() {
  initTransportCost();
  XpressProblem prob;
  // Variables that indicate whether a facility is open or not.
  // This is a one-dimensional hash map that is indexed with Facility instances.
  // For indexing we can use operator[] if we know the key exists.
  std::unordered_map<Facility, Variable> y =
      prob.addVariables(facilities)
          .withType(ColumnType::Binary)
          .withName([](Facility f) { return f.name; })
          .toMap();
  // Variables that indicate how much of a customer is served from a facility.
  // This is a two-dimensional hash map that is indexed with (Facility,Customer)
  // pairs.
  // For indexing we can either use operator[](MapKey2<Facility,Customer>) or
  // the custom lookup operator()(Facility,Customer).
  // In this example we use the latter.
  HashMap2<Facility, Customer, Variable> x =
      prob.addVariables(facilities, customers)
          .withName([](auto f, auto c) {
            return xpress::format("x[%s,%s]", f.name.c_str(), c.name.c_str());
          })
          .toMap();

  // for each customer c
  // sum(f=1..m) x[f,c] = d
  prob.addConstraints(customers, [&](Customer const &c) {
    return sum(facilities, [&](Facility const &f) { return x(f, c); }) ==
           c.demand;
  });

  // for each facility f
  // sum(c=1..n) x[f,c] <= capacity[j] * y[f]
  prob.addConstraints(facilities, [&](Facility const &f) {
    return sum(customers, [&](Customer const &c) { return x(f, c); }) <=
           f.capacity * y[f];
  });

  // minimize sum(j=1..m) cost[j] * y[j] +
  // sum(i=1..n) sum(j=1..m) cost[f,c] * x[f,c]
  prob.setObjective(
      sum(facilities, [&](Facility const &f) { return f.cost * y[f]; }) +
      sum(customers, [&](Customer const &c) {
        return sum(facilities, [&](Facility const &f) {
          return transportCost(f, c) * x(f, c);
        });
      }));

  prob.writeProb("facilitylocationcollection.lp", "l");

  prob.optimize();
  if (prob.attributes.getSolStatus() != SolStatus::Optimal)
    throw std::runtime_error("failed to optimize with status " +
                             to_string(prob.attributes.getSolStatus()));
  auto sol = prob.getSolution();
  for (Facility const &f : facilities) {
    if (y[f].getValue(sol) > 0.5) {
      std::cout << "Facility " << f.name << " is open, serves" << std::endl;
      for (Customer const &c : customers) {
        if (x(f, c).getValue(sol) > 0.0)
          std::cout << "  " << c.name << ": " << x(f, c).getValue(sol)
                    << std::endl;
      }
    }
  }
  return 0;
}

© 2001-2025 Fair Isaac Corporation. All rights reserved. This documentation is the property of Fair Isaac Corporation (“FICO”). Receipt or possession of this documentation does not convey rights to disclose, reproduce, make derivative works, use, or allow others to use it except solely for internal evaluation purposes to determine whether to purchase a license to the software described in this documentation, or as otherwise set forth in a written software license agreement between you and FICO (or a FICO affiliate). Use of this documentation and the software described in it must conform strictly to the foregoing permitted uses, and no other use is permitted.