É possível usar o solucionador de fluxo de custo mínimo para resolver casos especiais do problema de atribuição.
Na verdade, o fluxo de custo mínimo geralmente pode retornar uma solução mais rapidamente do que o solucionador MIP ou CP-SAT. No entanto, o MIP e o CP-SAT podem resolver uma classe maior de problemas do que o fluxo de custo mínimo. Portanto, na maioria dos casos, o MIP ou o CP-SAT são as melhores escolhas.
As seções a seguir apresentam programas em Python que resolvem os seguintes problemas de atribuição usando o solucionador de fluxo de custo mínimo:
- Um exemplo de atribuição linear mínimo.
- Um problema de atribuição com equipes de funcionários.
Exemplo de atribuição linear
Nesta seção, mostramos como resolver o exemplo descrito na seção Solucionador de atribuição linear como um problema de fluxo de custo mínimo.
Importar as bibliotecas
O código a seguir importa a biblioteca necessária.
Python
from ortools.graph.python import min_cost_flow
C++
#include <cstdint> #include <vector> #include "ortools/graph/min_cost_flow.h"
Java
import com.google.ortools.Loader; import com.google.ortools.graph.MinCostFlow; import com.google.ortools.graph.MinCostFlowBase;
C#
using System; using Google.OrTools.Graph;
Declarar o solucionador
O código a seguir cria o solucionador de fluxo de custo mínimo.
Python
# Instantiate a SimpleMinCostFlow solver. smcf = min_cost_flow.SimpleMinCostFlow()
C++
// Instantiate a SimpleMinCostFlow solver. SimpleMinCostFlow min_cost_flow;
Java
// Instantiate a SimpleMinCostFlow solver. MinCostFlow minCostFlow = new MinCostFlow();
C#
// Instantiate a SimpleMinCostFlow solver. MinCostFlow minCostFlow = new MinCostFlow();
Criar os dados
O diagrama de fluxo do problema consiste no gráfico bipartido da matriz de custo (consulte a visão geral da atribuição para um exemplo um pouco diferente), com uma origem e um coletor adicionados.
Os dados contêm as quatro matrizes a seguir, correspondentes aos nós inicial, nós finais, capacidades e custos do problema. O comprimento de cada matriz é o número de arcos no gráfico.
Python
# Define the directed graph for the flow.
start_nodes = (
[0, 0, 0, 0] + [1, 1, 1, 1, 2, 2, 2, 2, 3, 3, 3, 3, 4, 4, 4, 4] + [5, 6, 7, 8]
)
end_nodes = (
[1, 2, 3, 4] + [5, 6, 7, 8, 5, 6, 7, 8, 5, 6, 7, 8, 5, 6, 7, 8] + [9, 9, 9, 9]
)
capacities = (
[1, 1, 1, 1] + [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1] + [1, 1, 1, 1]
)
costs = (
[0, 0, 0, 0]
+ [90, 76, 75, 70, 35, 85, 55, 65, 125, 95, 90, 105, 45, 110, 95, 115]
+ [0, 0, 0, 0]
)
source = 0
sink = 9
tasks = 4
supplies = [tasks, 0, 0, 0, 0, 0, 0, 0, 0, -tasks]
C++
// Define four parallel arrays: sources, destinations, capacities,
// and unit costs between each pair. For instance, the arc from node 0
// to node 1 has a capacity of 15.
const std::vector<int64_t> start_nodes = {0, 0, 0, 0, 1, 1, 1, 1, 2, 2, 2, 2,
3, 3, 3, 3, 4, 4, 4, 4, 5, 6, 7, 8};
const std::vector<int64_t> end_nodes = {1, 2, 3, 4, 5, 6, 7, 8, 5, 6, 7, 8,
5, 6, 7, 8, 5, 6, 7, 8, 9, 9, 9, 9};
const std::vector<int64_t> capacities = {1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1};
const std::vector<int64_t> unit_costs = {0, 0, 0, 0, 90, 76, 75, 70,
35, 85, 55, 65, 125, 95, 90, 105,
45, 110, 95, 115, 0, 0, 0, 0};
const int64_t source = 0;
const int64_t sink = 9;
const int64_t tasks = 4;
// Define an array of supplies at each node.
const std::vector<int64_t> supplies = {tasks, 0, 0, 0, 0, 0, 0, 0, 0, -tasks};
Java
// Define four parallel arrays: sources, destinations, capacities, and unit costs
// between each pair.
int[] startNodes =
new int[] {0, 0, 0, 0, 1, 1, 1, 1, 2, 2, 2, 2, 3, 3, 3, 3, 4, 4, 4, 4, 5, 6, 7, 8};
int[] endNodes =
new int[] {1, 2, 3, 4, 5, 6, 7, 8, 5, 6, 7, 8, 5, 6, 7, 8, 5, 6, 7, 8, 9, 9, 9, 9};
int[] capacities =
new int[] {1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1};
int[] unitCosts = new int[] {
0, 0, 0, 0, 90, 76, 75, 70, 35, 85, 55, 65, 125, 95, 90, 105, 45, 110, 95, 115, 0, 0, 0, 0};
int source = 0;
int sink = 9;
int tasks = 4;
// Define an array of supplies at each node.
int[] supplies = new int[] {tasks, 0, 0, 0, 0, 0, 0, 0, 0, -tasks};
C#
// Define four parallel arrays: sources, destinations, capacities, and unit costs
// between each pair.
int[] startNodes = { 0, 0, 0, 0, 1, 1, 1, 1, 2, 2, 2, 2, 3, 3, 3, 3, 4, 4, 4, 4, 5, 6, 7, 8 };
int[] endNodes = { 1, 2, 3, 4, 5, 6, 7, 8, 5, 6, 7, 8, 5, 6, 7, 8, 5, 6, 7, 8, 9, 9, 9, 9 };
int[] capacities = { 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1 };
int[] unitCosts = { 0, 0, 0, 0, 90, 76, 75, 70, 35, 85, 55, 65,
125, 95, 90, 105, 45, 110, 95, 115, 0, 0, 0, 0 };
int source = 0;
int sink = 9;
int tasks = 4;
// Define an array of supplies at each node.
int[] supplies = { tasks, 0, 0, 0, 0, 0, 0, 0, 0, -tasks };
Para deixar claro como os dados são configurados, cada matriz é dividida em três submatrizes:
- A primeira matriz corresponde aos arcos que saem da origem.
- A segunda matriz corresponde aos arcos entre workers e tarefas.
Para o
costs, é apenas a matriz de custo (usada pelo solucionador de atribuição linear), nivelada em um vetor. - A terceira matriz corresponde aos arcos que levam ao coletor.
Os dados também incluem o vetor supplies, que fornece o fornecimento em cada
nó.
Como um problema de fluxo de custo mínimo representa um problema de atribuição
Como o problema do fluxo de custo mínimo acima representa um problema de atribuição? Primeiro, como a capacidade de cada arco é de 1, o fornecimento de 4 na fonte força cada um dos quatro arcos que levam aos trabalhadores a ter um fluxo de 1.
Em seguida, a condição fluxo-in-igual a fluxo-saída força o fluxo para fora de cada worker a ser 1. Se possível, o solucionador direcionaria esse fluxo pelo arco de custo mínimo que sai de cada worker. No entanto, o solucionador não pode direcionar os fluxos de dois workers diferentes para uma única tarefa. Se isso acontecesse, haveria um fluxo combinado de 2 para essa tarefa, que não poderia ser enviado pelo arco único com capacidade 1 da tarefa para o coletor. Isso significa que o solucionador só pode atribuir uma tarefa a um único worker, conforme exigido pelo problema de atribuição.
Por fim, a condição fluxo-in-igual a fluxo-saída força cada tarefa a ter um fluxo de saída de 1. Portanto, cada tarefa é executada por algum worker.
Criar o gráfico e as restrições
O código a seguir cria o gráfico e as restrições.
Python
# Add each arc.
for i in range(len(start_nodes)):
smcf.add_arc_with_capacity_and_unit_cost(
start_nodes[i], end_nodes[i], capacities[i], costs[i]
)
# Add node supplies.
for i in range(len(supplies)):
smcf.set_node_supply(i, supplies[i])
C++
// Add each arc.
for (int i = 0; i < start_nodes.size(); ++i) {
int arc = min_cost_flow.AddArcWithCapacityAndUnitCost(
start_nodes[i], end_nodes[i], capacities[i], unit_costs[i]);
if (arc != i) LOG(FATAL) << "Internal error";
}
// Add node supplies.
for (int i = 0; i < supplies.size(); ++i) {
min_cost_flow.SetNodeSupply(i, supplies[i]);
}
Java
// Add each arc.
for (int i = 0; i < startNodes.length; ++i) {
int arc = minCostFlow.addArcWithCapacityAndUnitCost(
startNodes[i], endNodes[i], capacities[i], unitCosts[i]);
if (arc != i) {
throw new Exception("Internal error");
}
}
// Add node supplies.
for (int i = 0; i < supplies.length; ++i) {
minCostFlow.setNodeSupply(i, supplies[i]);
}
C#
// Add each arc.
for (int i = 0; i < startNodes.Length; ++i)
{
int arc =
minCostFlow.AddArcWithCapacityAndUnitCost(startNodes[i], endNodes[i], capacities[i], unitCosts[i]);
if (arc != i)
throw new Exception("Internal error");
}
// Add node supplies.
for (int i = 0; i < supplies.Length; ++i)
{
minCostFlow.SetNodeSupply(i, supplies[i]);
}
Invocar o solucionador
O código a seguir invoca o solucionador e exibe a solução.
Python
# Find the minimum cost flow between node 0 and node 10. status = smcf.solve()
C++
// Find the min cost flow. int status = min_cost_flow.Solve();
Java
// Find the min cost flow. MinCostFlowBase.Status status = minCostFlow.solve();
C#
// Find the min cost flow. MinCostFlow.Status status = minCostFlow.Solve();
A solução consiste nos arcos entre workers e tarefas atribuídos a um fluxo de 1 pelo solucionador. Os arcos conectados à origem ou ao coletor não fazem parte da solução.
O programa verifica cada arco para conferir se eles têm o fluxo 1 e, em caso afirmativo, mostra o
Tail (nó inicial) e o Head (nó final) do arco, que correspondem a um
worker e uma tarefa na atribuição.
Saída do programa
Python
if status == smcf.OPTIMAL:
print("Total cost = ", smcf.optimal_cost())
print()
for arc in range(smcf.num_arcs()):
# Can ignore arcs leading out of source or into sink.
if smcf.tail(arc) != source and smcf.head(arc) != sink:
# Arcs in the solution have a flow value of 1. Their start and end nodes
# give an assignment of worker to task.
if smcf.flow(arc) > 0:
print(
"Worker %d assigned to task %d. Cost = %d"
% (smcf.tail(arc), smcf.head(arc), smcf.unit_cost(arc))
)
else:
print("There was an issue with the min cost flow input.")
print(f"Status: {status}")
C++
if (status == MinCostFlow::OPTIMAL) {
LOG(INFO) << "Total cost: " << min_cost_flow.OptimalCost();
LOG(INFO) << "";
for (std::size_t i = 0; i < min_cost_flow.NumArcs(); ++i) {
// Can ignore arcs leading out of source or into sink.
if (min_cost_flow.Tail(i) != source && min_cost_flow.Head(i) != sink) {
// Arcs in the solution have a flow value of 1. Their start and end
// nodes give an assignment of worker to task.
if (min_cost_flow.Flow(i) > 0) {
LOG(INFO) << "Worker " << min_cost_flow.Tail(i)
<< " assigned to task " << min_cost_flow.Head(i)
<< " Cost: " << min_cost_flow.UnitCost(i);
}
}
}
} else {
LOG(INFO) << "Solving the min cost flow problem failed.";
LOG(INFO) << "Solver status: " << status;
}
Java
if (status == MinCostFlow.Status.OPTIMAL) {
System.out.println("Total cost: " + minCostFlow.getOptimalCost());
System.out.println();
for (int i = 0; i < minCostFlow.getNumArcs(); ++i) {
// Can ignore arcs leading out of source or into sink.
if (minCostFlow.getTail(i) != source && minCostFlow.getHead(i) != sink) {
// Arcs in the solution have a flow value of 1. Their start and end nodes
// give an assignment of worker to task.
if (minCostFlow.getFlow(i) > 0) {
System.out.println("Worker " + minCostFlow.getTail(i) + " assigned to task "
+ minCostFlow.getHead(i) + " Cost: " + minCostFlow.getUnitCost(i));
}
}
}
} else {
System.out.println("Solving the min cost flow problem failed.");
System.out.println("Solver status: " + status);
}
C#
if (status == MinCostFlow.Status.OPTIMAL)
{
Console.WriteLine("Total cost: " + minCostFlow.OptimalCost());
Console.WriteLine("");
for (int i = 0; i < minCostFlow.NumArcs(); ++i)
{
// Can ignore arcs leading out of source or into sink.
if (minCostFlow.Tail(i) != source && minCostFlow.Head(i) != sink)
{
// Arcs in the solution have a flow value of 1. Their start and end nodes
// give an assignment of worker to task.
if (minCostFlow.Flow(i) > 0)
{
Console.WriteLine("Worker " + minCostFlow.Tail(i) + " assigned to task " + minCostFlow.Head(i) +
" Cost: " + minCostFlow.UnitCost(i));
}
}
}
}
else
{
Console.WriteLine("Solving the min cost flow problem failed.");
Console.WriteLine("Solver status: " + status);
}
Este é o resultado do programa.
Total cost = 265 Worker 1 assigned to task 8. Cost = 70 Worker 2 assigned to task 7. Cost = 55 Worker 3 assigned to task 6. Cost = 95 Worker 4 assigned to task 5. Cost = 45 Time = 0.000245 seconds
O resultado é o mesmo que o do solucionador de atribuição linear, exceto pela numeração diferente de workers e custos. O solucionador de atribuição linear é um pouco mais rápido que o fluxo de custo mínimo: 0,000147 segundos x 0,000458 segundos.
Todo o programa
Confira o programa inteiro abaixo.
Python
"""Linear assignment example."""
from ortools.graph.python import min_cost_flow
def main():
"""Solving an Assignment Problem with MinCostFlow."""
# Instantiate a SimpleMinCostFlow solver.
smcf = min_cost_flow.SimpleMinCostFlow()
# Define the directed graph for the flow.
start_nodes = (
[0, 0, 0, 0] + [1, 1, 1, 1, 2, 2, 2, 2, 3, 3, 3, 3, 4, 4, 4, 4] + [5, 6, 7, 8]
)
end_nodes = (
[1, 2, 3, 4] + [5, 6, 7, 8, 5, 6, 7, 8, 5, 6, 7, 8, 5, 6, 7, 8] + [9, 9, 9, 9]
)
capacities = (
[1, 1, 1, 1] + [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1] + [1, 1, 1, 1]
)
costs = (
[0, 0, 0, 0]
+ [90, 76, 75, 70, 35, 85, 55, 65, 125, 95, 90, 105, 45, 110, 95, 115]
+ [0, 0, 0, 0]
)
source = 0
sink = 9
tasks = 4
supplies = [tasks, 0, 0, 0, 0, 0, 0, 0, 0, -tasks]
# Add each arc.
for i in range(len(start_nodes)):
smcf.add_arc_with_capacity_and_unit_cost(
start_nodes[i], end_nodes[i], capacities[i], costs[i]
)
# Add node supplies.
for i in range(len(supplies)):
smcf.set_node_supply(i, supplies[i])
# Find the minimum cost flow between node 0 and node 10.
status = smcf.solve()
if status == smcf.OPTIMAL:
print("Total cost = ", smcf.optimal_cost())
print()
for arc in range(smcf.num_arcs()):
# Can ignore arcs leading out of source or into sink.
if smcf.tail(arc) != source and smcf.head(arc) != sink:
# Arcs in the solution have a flow value of 1. Their start and end nodes
# give an assignment of worker to task.
if smcf.flow(arc) > 0:
print(
"Worker %d assigned to task %d. Cost = %d"
% (smcf.tail(arc), smcf.head(arc), smcf.unit_cost(arc))
)
else:
print("There was an issue with the min cost flow input.")
print(f"Status: {status}")
if __name__ == "__main__":
main()
C++
#include <cstdint>
#include <vector>
#include "ortools/graph/min_cost_flow.h"
namespace operations_research {
// MinCostFlow simple interface example.
void AssignmentMinFlow() {
// Instantiate a SimpleMinCostFlow solver.
SimpleMinCostFlow min_cost_flow;
// Define four parallel arrays: sources, destinations, capacities,
// and unit costs between each pair. For instance, the arc from node 0
// to node 1 has a capacity of 15.
const std::vector<int64_t> start_nodes = {0, 0, 0, 0, 1, 1, 1, 1, 2, 2, 2, 2,
3, 3, 3, 3, 4, 4, 4, 4, 5, 6, 7, 8};
const std::vector<int64_t> end_nodes = {1, 2, 3, 4, 5, 6, 7, 8, 5, 6, 7, 8,
5, 6, 7, 8, 5, 6, 7, 8, 9, 9, 9, 9};
const std::vector<int64_t> capacities = {1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1};
const std::vector<int64_t> unit_costs = {0, 0, 0, 0, 90, 76, 75, 70,
35, 85, 55, 65, 125, 95, 90, 105,
45, 110, 95, 115, 0, 0, 0, 0};
const int64_t source = 0;
const int64_t sink = 9;
const int64_t tasks = 4;
// Define an array of supplies at each node.
const std::vector<int64_t> supplies = {tasks, 0, 0, 0, 0, 0, 0, 0, 0, -tasks};
// Add each arc.
for (int i = 0; i < start_nodes.size(); ++i) {
int arc = min_cost_flow.AddArcWithCapacityAndUnitCost(
start_nodes[i], end_nodes[i], capacities[i], unit_costs[i]);
if (arc != i) LOG(FATAL) << "Internal error";
}
// Add node supplies.
for (int i = 0; i < supplies.size(); ++i) {
min_cost_flow.SetNodeSupply(i, supplies[i]);
}
// Find the min cost flow.
int status = min_cost_flow.Solve();
if (status == MinCostFlow::OPTIMAL) {
LOG(INFO) << "Total cost: " << min_cost_flow.OptimalCost();
LOG(INFO) << "";
for (std::size_t i = 0; i < min_cost_flow.NumArcs(); ++i) {
// Can ignore arcs leading out of source or into sink.
if (min_cost_flow.Tail(i) != source && min_cost_flow.Head(i) != sink) {
// Arcs in the solution have a flow value of 1. Their start and end
// nodes give an assignment of worker to task.
if (min_cost_flow.Flow(i) > 0) {
LOG(INFO) << "Worker " << min_cost_flow.Tail(i)
<< " assigned to task " << min_cost_flow.Head(i)
<< " Cost: " << min_cost_flow.UnitCost(i);
}
}
}
} else {
LOG(INFO) << "Solving the min cost flow problem failed.";
LOG(INFO) << "Solver status: " << status;
}
}
} // namespace operations_research
int main() {
operations_research::AssignmentMinFlow();
return EXIT_SUCCESS;
}
Java
package com.google.ortools.graph.samples;
import com.google.ortools.Loader;
import com.google.ortools.graph.MinCostFlow;
import com.google.ortools.graph.MinCostFlowBase;
/** Minimal Assignment Min Flow. */
public class AssignmentMinFlow {
public static void main(String[] args) throws Exception {
Loader.loadNativeLibraries();
// Instantiate a SimpleMinCostFlow solver.
MinCostFlow minCostFlow = new MinCostFlow();
// Define four parallel arrays: sources, destinations, capacities, and unit costs
// between each pair.
int[] startNodes =
new int[] {0, 0, 0, 0, 1, 1, 1, 1, 2, 2, 2, 2, 3, 3, 3, 3, 4, 4, 4, 4, 5, 6, 7, 8};
int[] endNodes =
new int[] {1, 2, 3, 4, 5, 6, 7, 8, 5, 6, 7, 8, 5, 6, 7, 8, 5, 6, 7, 8, 9, 9, 9, 9};
int[] capacities =
new int[] {1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1};
int[] unitCosts = new int[] {
0, 0, 0, 0, 90, 76, 75, 70, 35, 85, 55, 65, 125, 95, 90, 105, 45, 110, 95, 115, 0, 0, 0, 0};
int source = 0;
int sink = 9;
int tasks = 4;
// Define an array of supplies at each node.
int[] supplies = new int[] {tasks, 0, 0, 0, 0, 0, 0, 0, 0, -tasks};
// Add each arc.
for (int i = 0; i < startNodes.length; ++i) {
int arc = minCostFlow.addArcWithCapacityAndUnitCost(
startNodes[i], endNodes[i], capacities[i], unitCosts[i]);
if (arc != i) {
throw new Exception("Internal error");
}
}
// Add node supplies.
for (int i = 0; i < supplies.length; ++i) {
minCostFlow.setNodeSupply(i, supplies[i]);
}
// Find the min cost flow.
MinCostFlowBase.Status status = minCostFlow.solve();
if (status == MinCostFlow.Status.OPTIMAL) {
System.out.println("Total cost: " + minCostFlow.getOptimalCost());
System.out.println();
for (int i = 0; i < minCostFlow.getNumArcs(); ++i) {
// Can ignore arcs leading out of source or into sink.
if (minCostFlow.getTail(i) != source && minCostFlow.getHead(i) != sink) {
// Arcs in the solution have a flow value of 1. Their start and end nodes
// give an assignment of worker to task.
if (minCostFlow.getFlow(i) > 0) {
System.out.println("Worker " + minCostFlow.getTail(i) + " assigned to task "
+ minCostFlow.getHead(i) + " Cost: " + minCostFlow.getUnitCost(i));
}
}
}
} else {
System.out.println("Solving the min cost flow problem failed.");
System.out.println("Solver status: " + status);
}
}
private AssignmentMinFlow() {}
}
C#
using System;
using Google.OrTools.Graph;
public class AssignmentMinFlow
{
static void Main()
{
// Instantiate a SimpleMinCostFlow solver.
MinCostFlow minCostFlow = new MinCostFlow();
// Define four parallel arrays: sources, destinations, capacities, and unit costs
// between each pair.
int[] startNodes = { 0, 0, 0, 0, 1, 1, 1, 1, 2, 2, 2, 2, 3, 3, 3, 3, 4, 4, 4, 4, 5, 6, 7, 8 };
int[] endNodes = { 1, 2, 3, 4, 5, 6, 7, 8, 5, 6, 7, 8, 5, 6, 7, 8, 5, 6, 7, 8, 9, 9, 9, 9 };
int[] capacities = { 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1 };
int[] unitCosts = { 0, 0, 0, 0, 90, 76, 75, 70, 35, 85, 55, 65,
125, 95, 90, 105, 45, 110, 95, 115, 0, 0, 0, 0 };
int source = 0;
int sink = 9;
int tasks = 4;
// Define an array of supplies at each node.
int[] supplies = { tasks, 0, 0, 0, 0, 0, 0, 0, 0, -tasks };
// Add each arc.
for (int i = 0; i < startNodes.Length; ++i)
{
int arc =
minCostFlow.AddArcWithCapacityAndUnitCost(startNodes[i], endNodes[i], capacities[i], unitCosts[i]);
if (arc != i)
throw new Exception("Internal error");
}
// Add node supplies.
for (int i = 0; i < supplies.Length; ++i)
{
minCostFlow.SetNodeSupply(i, supplies[i]);
}
// Find the min cost flow.
MinCostFlow.Status status = minCostFlow.Solve();
if (status == MinCostFlow.Status.OPTIMAL)
{
Console.WriteLine("Total cost: " + minCostFlow.OptimalCost());
Console.WriteLine("");
for (int i = 0; i < minCostFlow.NumArcs(); ++i)
{
// Can ignore arcs leading out of source or into sink.
if (minCostFlow.Tail(i) != source && minCostFlow.Head(i) != sink)
{
// Arcs in the solution have a flow value of 1. Their start and end nodes
// give an assignment of worker to task.
if (minCostFlow.Flow(i) > 0)
{
Console.WriteLine("Worker " + minCostFlow.Tail(i) + " assigned to task " + minCostFlow.Head(i) +
" Cost: " + minCostFlow.UnitCost(i));
}
}
}
}
else
{
Console.WriteLine("Solving the min cost flow problem failed.");
Console.WriteLine("Solver status: " + status);
}
}
}
Atribuição com equipes de funcionários
Esta seção apresenta um problema de atribuição mais geral. Nesse problema, seis trabalhadores são divididos em duas equipes. O problema é atribuir quatro tarefas aos trabalhadores para que a carga de trabalho seja igualmente equilibrada entre as equipes, ou seja, cada equipe executa duas delas.
Para uma solução do solucionador MIP para esse problema, consulte Atribuição com equipes de funcionários.
As seções a seguir descrevem um programa que resolve o problema usando o solucionador de fluxo de custo mínimo.
Importar as bibliotecas
O código a seguir importa a biblioteca necessária.
Python
from ortools.graph.python import min_cost_flow
C++
#include <cstdint> #include <vector> #include "ortools/graph/min_cost_flow.h"
Java
import com.google.ortools.Loader; import com.google.ortools.graph.MinCostFlow; import com.google.ortools.graph.MinCostFlowBase;
C#
using System; using Google.OrTools.Graph;
Declarar o solucionador
O código a seguir cria o solucionador de fluxo de custo mínimo.
Python
smcf = min_cost_flow.SimpleMinCostFlow()
C++
// Instantiate a SimpleMinCostFlow solver. SimpleMinCostFlow min_cost_flow;
Java
// Instantiate a SimpleMinCostFlow solver. MinCostFlow minCostFlow = new MinCostFlow();
C#
// Instantiate a SimpleMinCostFlow solver. MinCostFlow minCostFlow = new MinCostFlow();
Criar os dados
O código a seguir cria os dados para o programa.
Python
# Define the directed graph for the flow.
team_a = [1, 3, 5]
team_b = [2, 4, 6]
start_nodes = (
# fmt: off
[0, 0]
+ [11, 11, 11]
+ [12, 12, 12]
+ [1, 1, 1, 1, 2, 2, 2, 2, 3, 3, 3, 3, 4, 4, 4, 4, 5, 5, 5, 5, 6, 6, 6, 6]
+ [7, 8, 9, 10]
# fmt: on
)
end_nodes = (
# fmt: off
[11, 12]
+ team_a
+ team_b
+ [7, 8, 9, 10, 7, 8, 9, 10, 7, 8, 9, 10, 7, 8, 9, 10, 7, 8, 9, 10, 7, 8, 9, 10]
+ [13, 13, 13, 13]
# fmt: on
)
capacities = (
# fmt: off
[2, 2]
+ [1, 1, 1]
+ [1, 1, 1]
+ [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]
+ [1, 1, 1, 1]
# fmt: on
)
costs = (
# fmt: off
[0, 0]
+ [0, 0, 0]
+ [0, 0, 0]
+ [90, 76, 75, 70, 35, 85, 55, 65, 125, 95, 90, 105, 45, 110, 95, 115, 60, 105, 80, 75, 45, 65, 110, 95]
+ [0, 0, 0, 0]
# fmt: on
)
source = 0
sink = 13
tasks = 4
# Define an array of supplies at each node.
supplies = [tasks, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, -tasks]
C++
// Define the directed graph for the flow.
const std::vector<int64_t> team_A = {1, 3, 5};
const std::vector<int64_t> team_B = {2, 4, 6};
const std::vector<int64_t> start_nodes = {
0, 0, 11, 11, 11, 12, 12, 12, 1, 1, 1, 1, 2, 2, 2, 2, 3, 3,
3, 3, 4, 4, 4, 4, 5, 5, 5, 5, 6, 6, 6, 6, 7, 8, 9, 10};
const std::vector<int64_t> end_nodes = {
11, 12, 1, 3, 5, 2, 4, 6, 7, 8, 9, 10, 7, 8, 9, 10, 7, 8,
9, 10, 7, 8, 9, 10, 7, 8, 9, 10, 7, 8, 9, 10, 13, 13, 13, 13};
const std::vector<int64_t> capacities = {2, 2, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1};
const std::vector<int64_t> unit_costs = {
0, 0, 0, 0, 0, 0, 0, 0, 90, 76, 75, 70,
35, 85, 55, 65, 125, 95, 90, 105, 45, 110, 95, 115,
60, 105, 80, 75, 45, 65, 110, 95, 0, 0, 0, 0};
const int64_t source = 0;
const int64_t sink = 13;
const int64_t tasks = 4;
// Define an array of supplies at each node.
const std::vector<int64_t> supplies = {tasks, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, -tasks};
Java
// Define the directed graph for the flow.
// int[] teamA = new int[] {1, 3, 5};
// int[] teamB = new int[] {2, 4, 6};
int[] startNodes = new int[] {0, 0, 11, 11, 11, 12, 12, 12, 1, 1, 1, 1, 2, 2, 2, 2, 3, 3, 3, 3,
4, 4, 4, 4, 5, 5, 5, 5, 6, 6, 6, 6, 7, 8, 9, 10};
int[] endNodes = new int[] {11, 12, 1, 3, 5, 2, 4, 6, 7, 8, 9, 10, 7, 8, 9, 10, 7, 8, 9, 10, 7,
8, 9, 10, 7, 8, 9, 10, 7, 8, 9, 10, 13, 13, 13, 13};
int[] capacities = new int[] {2, 2, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1};
int[] unitCosts = new int[] {0, 0, 0, 0, 0, 0, 0, 0, 90, 76, 75, 70, 35, 85, 55, 65, 125, 95,
90, 105, 45, 110, 95, 115, 60, 105, 80, 75, 45, 65, 110, 95, 0, 0, 0, 0};
int source = 0;
int sink = 13;
int tasks = 4;
// Define an array of supplies at each node.
int[] supplies = new int[] {tasks, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, -tasks};
C#
// Define the directed graph for the flow.
int[] teamA = { 1, 3, 5 };
int[] teamB = { 2, 4, 6 };
// Define four parallel arrays: sources, destinations, capacities, and unit costs
// between each pair.
int[] startNodes = { 0, 0, 11, 11, 11, 12, 12, 12, 1, 1, 1, 1, 2, 2, 2, 2, 3, 3,
3, 3, 4, 4, 4, 4, 5, 5, 5, 5, 6, 6, 6, 6, 7, 8, 9, 10 };
int[] endNodes = { 11, 12, 1, 3, 5, 2, 4, 6, 7, 8, 9, 10, 7, 8, 9, 10, 7, 8,
9, 10, 7, 8, 9, 10, 7, 8, 9, 10, 7, 8, 9, 10, 13, 13, 13, 13 };
int[] capacities = { 2, 2, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1 };
int[] unitCosts = { 0, 0, 0, 0, 0, 0, 0, 0, 90, 76, 75, 70, 35, 85, 55, 65, 125, 95,
90, 105, 45, 110, 95, 115, 60, 105, 80, 75, 45, 65, 110, 95, 0, 0, 0, 0 };
int source = 0;
int sink = 13;
int tasks = 4;
// Define an array of supplies at each node.
int[] supplies = { tasks, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, -tasks };
Os workers correspondem aos nós 1 a 6. A equipe A consiste nos workers 1, 3 e 5, e a equipe B consiste nos workers 2, 4 e 6. As tarefas são numeradas de 7 a 10.
Há dois novos nós, 11 e 12, entre a origem e os workers. O nó 11 está conectado aos nós da equipe A, e o nó 12 está conectado aos nós da equipe B, com arcos de capacidade 1. O gráfico abaixo mostra apenas os nós e arcos da origem até os workers.
A chave para equilibrar a carga de trabalho é que a origem 0 está conectada aos nós 11 e 12 por arcos de capacidade 2. Isso significa que os nós 11 e 12 (e, portanto, as equipes A e B) podem ter um fluxo máximo de 2. Como resultado, cada equipe pode executar no máximo duas das tarefas.
Criar as restrições
Python
# Add each arc.
for i in range(0, len(start_nodes)):
smcf.add_arc_with_capacity_and_unit_cost(
start_nodes[i], end_nodes[i], capacities[i], costs[i]
)
# Add node supplies.
for i in range(0, len(supplies)):
smcf.set_node_supply(i, supplies[i])
C++
// Add each arc.
for (int i = 0; i < start_nodes.size(); ++i) {
int arc = min_cost_flow.AddArcWithCapacityAndUnitCost(
start_nodes[i], end_nodes[i], capacities[i], unit_costs[i]);
if (arc != i) LOG(FATAL) << "Internal error";
}
// Add node supplies.
for (int i = 0; i < supplies.size(); ++i) {
min_cost_flow.SetNodeSupply(i, supplies[i]);
}
Java
// Add each arc.
for (int i = 0; i < startNodes.length; ++i) {
int arc = minCostFlow.addArcWithCapacityAndUnitCost(
startNodes[i], endNodes[i], capacities[i], unitCosts[i]);
if (arc != i) {
throw new Exception("Internal error");
}
}
// Add node supplies.
for (int i = 0; i < supplies.length; ++i) {
minCostFlow.setNodeSupply(i, supplies[i]);
}
C#
// Add each arc.
for (int i = 0; i < startNodes.Length; ++i)
{
int arc =
minCostFlow.AddArcWithCapacityAndUnitCost(startNodes[i], endNodes[i], capacities[i], unitCosts[i]);
if (arc != i)
throw new Exception("Internal error");
}
// Add node supplies.
for (int i = 0; i < supplies.Length; ++i)
{
minCostFlow.SetNodeSupply(i, supplies[i]);
}
Invocar o solucionador
Python
# Find the minimum cost flow between node 0 and node 10. status = smcf.solve()
C++
// Find the min cost flow. int status = min_cost_flow.Solve();
Java
// Find the min cost flow. MinCostFlowBase.Status status = minCostFlow.solve();
C#
// Find the min cost flow. MinCostFlow.Status status = minCostFlow.Solve();
Saída do programa
Python
if status == smcf.OPTIMAL:
print("Total cost = ", smcf.optimal_cost())
print()
for arc in range(smcf.num_arcs()):
# Can ignore arcs leading out of source or intermediate, or into sink.
if (
smcf.tail(arc) != source
and smcf.tail(arc) != 11
and smcf.tail(arc) != 12
and smcf.head(arc) != sink
):
# Arcs in the solution will have a flow value of 1.
# There start and end nodes give an assignment of worker to task.
if smcf.flow(arc) > 0:
print(
"Worker %d assigned to task %d. Cost = %d"
% (smcf.tail(arc), smcf.head(arc), smcf.unit_cost(arc))
)
else:
print("There was an issue with the min cost flow input.")
print(f"Status: {status}")
C++
if (status == MinCostFlow::OPTIMAL) {
LOG(INFO) << "Total cost: " << min_cost_flow.OptimalCost();
LOG(INFO) << "";
for (std::size_t i = 0; i < min_cost_flow.NumArcs(); ++i) {
// Can ignore arcs leading out of source or intermediate nodes, or into
// sink.
if (min_cost_flow.Tail(i) != source && min_cost_flow.Tail(i) != 11 &&
min_cost_flow.Tail(i) != 12 && min_cost_flow.Head(i) != sink) {
// Arcs in the solution have a flow value of 1. Their start and end
// nodes give an assignment of worker to task.
if (min_cost_flow.Flow(i) > 0) {
LOG(INFO) << "Worker " << min_cost_flow.Tail(i)
<< " assigned to task " << min_cost_flow.Head(i)
<< " Cost: " << min_cost_flow.UnitCost(i);
}
}
}
} else {
LOG(INFO) << "Solving the min cost flow problem failed.";
LOG(INFO) << "Solver status: " << status;
}
Java
if (status == MinCostFlow.Status.OPTIMAL) {
System.out.println("Total cost: " + minCostFlow.getOptimalCost());
System.out.println();
for (int i = 0; i < minCostFlow.getNumArcs(); ++i) {
// Can ignore arcs leading out of source or intermediate nodes, or into sink.
if (minCostFlow.getTail(i) != source && minCostFlow.getTail(i) != 11
&& minCostFlow.getTail(i) != 12 && minCostFlow.getHead(i) != sink) {
// Arcs in the solution have a flow value of 1. Their start and end nodes
// give an assignment of worker to task.
if (minCostFlow.getFlow(i) > 0) {
System.out.println("Worker " + minCostFlow.getTail(i) + " assigned to task "
+ minCostFlow.getHead(i) + " Cost: " + minCostFlow.getUnitCost(i));
}
}
}
} else {
System.out.println("Solving the min cost flow problem failed.");
System.out.println("Solver status: " + status);
}
C#
if (status == MinCostFlow.Status.OPTIMAL)
{
Console.WriteLine("Total cost: " + minCostFlow.OptimalCost());
Console.WriteLine("");
for (int i = 0; i < minCostFlow.NumArcs(); ++i)
{
// Can ignore arcs leading out of source or into sink.
if (minCostFlow.Tail(i) != source && minCostFlow.Tail(i) != 11 && minCostFlow.Tail(i) != 12 &&
minCostFlow.Head(i) != sink)
{
// Arcs in the solution have a flow value of 1. Their start and end nodes
// give an assignment of worker to task.
if (minCostFlow.Flow(i) > 0)
{
Console.WriteLine("Worker " + minCostFlow.Tail(i) + " assigned to task " + minCostFlow.Head(i) +
" Cost: " + minCostFlow.UnitCost(i));
}
}
}
}
else
{
Console.WriteLine("Solving the min cost flow problem failed.");
Console.WriteLine("Solver status: " + status);
}
Confira a seguir a saída do programa.
Total cost = 250 Worker 1 assigned to task 9. Cost = 75 Worker 2 assigned to task 7. Cost = 35 Worker 5 assigned to task 10. Cost = 75 Worker 6 assigned to task 8. Cost = 65 Time = 0.00031 seconds
A equipe A recebe as tarefas 9 e 10, e a equipe B recebe as tarefas 7 e 8.
O solucionador de fluxo de custo mínimo é mais rápido para esse problema do que o solucionador MIP, que leva cerca de 0,006 segundo.
Todo o programa
Confira o programa inteiro abaixo.
Python
"""Assignment with teams of workers."""
from ortools.graph.python import min_cost_flow
def main():
"""Solving an Assignment with teams of worker."""
smcf = min_cost_flow.SimpleMinCostFlow()
# Define the directed graph for the flow.
team_a = [1, 3, 5]
team_b = [2, 4, 6]
start_nodes = (
# fmt: off
[0, 0]
+ [11, 11, 11]
+ [12, 12, 12]
+ [1, 1, 1, 1, 2, 2, 2, 2, 3, 3, 3, 3, 4, 4, 4, 4, 5, 5, 5, 5, 6, 6, 6, 6]
+ [7, 8, 9, 10]
# fmt: on
)
end_nodes = (
# fmt: off
[11, 12]
+ team_a
+ team_b
+ [7, 8, 9, 10, 7, 8, 9, 10, 7, 8, 9, 10, 7, 8, 9, 10, 7, 8, 9, 10, 7, 8, 9, 10]
+ [13, 13, 13, 13]
# fmt: on
)
capacities = (
# fmt: off
[2, 2]
+ [1, 1, 1]
+ [1, 1, 1]
+ [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]
+ [1, 1, 1, 1]
# fmt: on
)
costs = (
# fmt: off
[0, 0]
+ [0, 0, 0]
+ [0, 0, 0]
+ [90, 76, 75, 70, 35, 85, 55, 65, 125, 95, 90, 105, 45, 110, 95, 115, 60, 105, 80, 75, 45, 65, 110, 95]
+ [0, 0, 0, 0]
# fmt: on
)
source = 0
sink = 13
tasks = 4
# Define an array of supplies at each node.
supplies = [tasks, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, -tasks]
# Add each arc.
for i in range(0, len(start_nodes)):
smcf.add_arc_with_capacity_and_unit_cost(
start_nodes[i], end_nodes[i], capacities[i], costs[i]
)
# Add node supplies.
for i in range(0, len(supplies)):
smcf.set_node_supply(i, supplies[i])
# Find the minimum cost flow between node 0 and node 10.
status = smcf.solve()
if status == smcf.OPTIMAL:
print("Total cost = ", smcf.optimal_cost())
print()
for arc in range(smcf.num_arcs()):
# Can ignore arcs leading out of source or intermediate, or into sink.
if (
smcf.tail(arc) != source
and smcf.tail(arc) != 11
and smcf.tail(arc) != 12
and smcf.head(arc) != sink
):
# Arcs in the solution will have a flow value of 1.
# There start and end nodes give an assignment of worker to task.
if smcf.flow(arc) > 0:
print(
"Worker %d assigned to task %d. Cost = %d"
% (smcf.tail(arc), smcf.head(arc), smcf.unit_cost(arc))
)
else:
print("There was an issue with the min cost flow input.")
print(f"Status: {status}")
if __name__ == "__main__":
main()
C++
#include <cstdint>
#include <vector>
#include "ortools/graph/min_cost_flow.h"
namespace operations_research {
// MinCostFlow simple interface example.
void BalanceMinFlow() {
// Instantiate a SimpleMinCostFlow solver.
SimpleMinCostFlow min_cost_flow;
// Define the directed graph for the flow.
const std::vector<int64_t> team_A = {1, 3, 5};
const std::vector<int64_t> team_B = {2, 4, 6};
const std::vector<int64_t> start_nodes = {
0, 0, 11, 11, 11, 12, 12, 12, 1, 1, 1, 1, 2, 2, 2, 2, 3, 3,
3, 3, 4, 4, 4, 4, 5, 5, 5, 5, 6, 6, 6, 6, 7, 8, 9, 10};
const std::vector<int64_t> end_nodes = {
11, 12, 1, 3, 5, 2, 4, 6, 7, 8, 9, 10, 7, 8, 9, 10, 7, 8,
9, 10, 7, 8, 9, 10, 7, 8, 9, 10, 7, 8, 9, 10, 13, 13, 13, 13};
const std::vector<int64_t> capacities = {2, 2, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1};
const std::vector<int64_t> unit_costs = {
0, 0, 0, 0, 0, 0, 0, 0, 90, 76, 75, 70,
35, 85, 55, 65, 125, 95, 90, 105, 45, 110, 95, 115,
60, 105, 80, 75, 45, 65, 110, 95, 0, 0, 0, 0};
const int64_t source = 0;
const int64_t sink = 13;
const int64_t tasks = 4;
// Define an array of supplies at each node.
const std::vector<int64_t> supplies = {tasks, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, -tasks};
// Add each arc.
for (int i = 0; i < start_nodes.size(); ++i) {
int arc = min_cost_flow.AddArcWithCapacityAndUnitCost(
start_nodes[i], end_nodes[i], capacities[i], unit_costs[i]);
if (arc != i) LOG(FATAL) << "Internal error";
}
// Add node supplies.
for (int i = 0; i < supplies.size(); ++i) {
min_cost_flow.SetNodeSupply(i, supplies[i]);
}
// Find the min cost flow.
int status = min_cost_flow.Solve();
if (status == MinCostFlow::OPTIMAL) {
LOG(INFO) << "Total cost: " << min_cost_flow.OptimalCost();
LOG(INFO) << "";
for (std::size_t i = 0; i < min_cost_flow.NumArcs(); ++i) {
// Can ignore arcs leading out of source or intermediate nodes, or into
// sink.
if (min_cost_flow.Tail(i) != source && min_cost_flow.Tail(i) != 11 &&
min_cost_flow.Tail(i) != 12 && min_cost_flow.Head(i) != sink) {
// Arcs in the solution have a flow value of 1. Their start and end
// nodes give an assignment of worker to task.
if (min_cost_flow.Flow(i) > 0) {
LOG(INFO) << "Worker " << min_cost_flow.Tail(i)
<< " assigned to task " << min_cost_flow.Head(i)
<< " Cost: " << min_cost_flow.UnitCost(i);
}
}
}
} else {
LOG(INFO) << "Solving the min cost flow problem failed.";
LOG(INFO) << "Solver status: " << status;
}
}
} // namespace operations_research
int main() {
operations_research::BalanceMinFlow();
return EXIT_SUCCESS;
}
Java
package com.google.ortools.graph.samples;
import com.google.ortools.Loader;
import com.google.ortools.graph.MinCostFlow;
import com.google.ortools.graph.MinCostFlowBase;
/** Minimal Assignment Min Flow. */
public class BalanceMinFlow {
public static void main(String[] args) throws Exception {
Loader.loadNativeLibraries();
// Instantiate a SimpleMinCostFlow solver.
MinCostFlow minCostFlow = new MinCostFlow();
// Define the directed graph for the flow.
// int[] teamA = new int[] {1, 3, 5};
// int[] teamB = new int[] {2, 4, 6};
int[] startNodes = new int[] {0, 0, 11, 11, 11, 12, 12, 12, 1, 1, 1, 1, 2, 2, 2, 2, 3, 3, 3, 3,
4, 4, 4, 4, 5, 5, 5, 5, 6, 6, 6, 6, 7, 8, 9, 10};
int[] endNodes = new int[] {11, 12, 1, 3, 5, 2, 4, 6, 7, 8, 9, 10, 7, 8, 9, 10, 7, 8, 9, 10, 7,
8, 9, 10, 7, 8, 9, 10, 7, 8, 9, 10, 13, 13, 13, 13};
int[] capacities = new int[] {2, 2, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1};
int[] unitCosts = new int[] {0, 0, 0, 0, 0, 0, 0, 0, 90, 76, 75, 70, 35, 85, 55, 65, 125, 95,
90, 105, 45, 110, 95, 115, 60, 105, 80, 75, 45, 65, 110, 95, 0, 0, 0, 0};
int source = 0;
int sink = 13;
int tasks = 4;
// Define an array of supplies at each node.
int[] supplies = new int[] {tasks, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, -tasks};
// Add each arc.
for (int i = 0; i < startNodes.length; ++i) {
int arc = minCostFlow.addArcWithCapacityAndUnitCost(
startNodes[i], endNodes[i], capacities[i], unitCosts[i]);
if (arc != i) {
throw new Exception("Internal error");
}
}
// Add node supplies.
for (int i = 0; i < supplies.length; ++i) {
minCostFlow.setNodeSupply(i, supplies[i]);
}
// Find the min cost flow.
MinCostFlowBase.Status status = minCostFlow.solve();
if (status == MinCostFlow.Status.OPTIMAL) {
System.out.println("Total cost: " + minCostFlow.getOptimalCost());
System.out.println();
for (int i = 0; i < minCostFlow.getNumArcs(); ++i) {
// Can ignore arcs leading out of source or intermediate nodes, or into sink.
if (minCostFlow.getTail(i) != source && minCostFlow.getTail(i) != 11
&& minCostFlow.getTail(i) != 12 && minCostFlow.getHead(i) != sink) {
// Arcs in the solution have a flow value of 1. Their start and end nodes
// give an assignment of worker to task.
if (minCostFlow.getFlow(i) > 0) {
System.out.println("Worker " + minCostFlow.getTail(i) + " assigned to task "
+ minCostFlow.getHead(i) + " Cost: " + minCostFlow.getUnitCost(i));
}
}
}
} else {
System.out.println("Solving the min cost flow problem failed.");
System.out.println("Solver status: " + status);
}
}
private BalanceMinFlow() {}
}
C#
using System;
using Google.OrTools.Graph;
public class BalanceMinFlow
{
static void Main()
{
// Instantiate a SimpleMinCostFlow solver.
MinCostFlow minCostFlow = new MinCostFlow();
// Define the directed graph for the flow.
int[] teamA = { 1, 3, 5 };
int[] teamB = { 2, 4, 6 };
// Define four parallel arrays: sources, destinations, capacities, and unit costs
// between each pair.
int[] startNodes = { 0, 0, 11, 11, 11, 12, 12, 12, 1, 1, 1, 1, 2, 2, 2, 2, 3, 3,
3, 3, 4, 4, 4, 4, 5, 5, 5, 5, 6, 6, 6, 6, 7, 8, 9, 10 };
int[] endNodes = { 11, 12, 1, 3, 5, 2, 4, 6, 7, 8, 9, 10, 7, 8, 9, 10, 7, 8,
9, 10, 7, 8, 9, 10, 7, 8, 9, 10, 7, 8, 9, 10, 13, 13, 13, 13 };
int[] capacities = { 2, 2, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1 };
int[] unitCosts = { 0, 0, 0, 0, 0, 0, 0, 0, 90, 76, 75, 70, 35, 85, 55, 65, 125, 95,
90, 105, 45, 110, 95, 115, 60, 105, 80, 75, 45, 65, 110, 95, 0, 0, 0, 0 };
int source = 0;
int sink = 13;
int tasks = 4;
// Define an array of supplies at each node.
int[] supplies = { tasks, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, -tasks };
// Add each arc.
for (int i = 0; i < startNodes.Length; ++i)
{
int arc =
minCostFlow.AddArcWithCapacityAndUnitCost(startNodes[i], endNodes[i], capacities[i], unitCosts[i]);
if (arc != i)
throw new Exception("Internal error");
}
// Add node supplies.
for (int i = 0; i < supplies.Length; ++i)
{
minCostFlow.SetNodeSupply(i, supplies[i]);
}
// Find the min cost flow.
MinCostFlow.Status status = minCostFlow.Solve();
if (status == MinCostFlow.Status.OPTIMAL)
{
Console.WriteLine("Total cost: " + minCostFlow.OptimalCost());
Console.WriteLine("");
for (int i = 0; i < minCostFlow.NumArcs(); ++i)
{
// Can ignore arcs leading out of source or into sink.
if (minCostFlow.Tail(i) != source && minCostFlow.Tail(i) != 11 && minCostFlow.Tail(i) != 12 &&
minCostFlow.Head(i) != sink)
{
// Arcs in the solution have a flow value of 1. Their start and end nodes
// give an assignment of worker to task.
if (minCostFlow.Flow(i) > 0)
{
Console.WriteLine("Worker " + minCostFlow.Tail(i) + " assigned to task " + minCostFlow.Head(i) +
" Cost: " + minCostFlow.UnitCost(i));
}
}
}
}
else
{
Console.WriteLine("Solving the min cost flow problem failed.");
Console.WriteLine("Solver status: " + status);
}
}
}