作为最低费用流问题进行分配

您可以使用最小成本流求解器来解决 分配问题。

事实上，与 MIP 或 CP-SAT 求解器。然而，相较于 MIP 和 CP-SAT， 最小成本流，因此在大多数情况下，MIP 或 CP-SAT 是最佳选择。

以下部分介绍了可解决以下问题的 Python 程序： 使用最小成本流求解器的分配问题：

• 极简线性分配示例。
• 工作器团队出现分配问题。

线性分配示例

本部分介绍如何求解示例，如 线性分配求解器（以最小值 成本流问题。

导入库

以下代码会导入所需的库。

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;

声明求解器

以下代码可创建最低成本流求解器。

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();

创建数据

该问题的流程图由两部分表示开销图组成 （请参阅 作业概览 添加了来源和接收器。

注意 ：工作器和任务的编号与 线性分配求解器部分， 因为最小成本流求解器要求图中的所有节点都是编号 非常明显。

数据包含以下四个数组，分别对应起始节点。 端点节点、容量和成本。每个数组的长度为 图表中弧形的数量。

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 };

为了说明数据的设置方式，每个数组分成了三个 子数组：

• 第一个数组对应于离开源的弧线。
• 第二个数组对应于 worker 和任务之间的弧线。 对于 costs，这只是 费用矩阵 （供线性赋值求解器使用），展平为向量。
• 第三个数组对应于进入接收器的弧线。

该数据还包含矢量 supplies，它为每个 节点。

最低费用流问题如何表示分配问题

上述最低费用流问题如何代表分配问题？首先， 由于每条弧形的容量都是 1，因此在来源处供应 4 个弧形会迫使每个弧形 通向工作器的四条弧线中，流量为 1。

接下来，flow-in-equals-flow-out 条件强制数据流离开每个 worker 为 1。如果可能，求解器会引导该流通过最小成本 每个 Worker 的弧线。但是，求解器无法引导流 处理一个任务。如果确实如此，那么 该任务的流向为 2，无法跨单个弧线发送， 1,000。 这意味着求解器只能将任务分配给单个 worker， 所需的资源。

最后，flow-in-equals-flow-out 条件迫使每个任务 流出 1，因此每个任务都由某个工作器执行。

创建图和约束条件

以下代码将创建图表和约束条件。

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]);
}

调用求解器

以下代码会调用求解器并显示求解结果。

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();

该解决方案由工作器和被分配了 求解器的 1 流。（连接到来源或接收器的弧线不属于 解决方案。）

此程序会检查每条弧线，确认其是否有 flow 1，如果有，则输出 弧形的 Tail（起始节点）和 Head（结束节点），分别对应 工作器和任务。

程序的输出

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);
}

以下是程序的输出。

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

其结果与 线性赋值求解器（不同的是， 工作器和成本的编号）。线性作业求解器的速度略快一些 与最低费用流相比 - 0.000147 秒对比 0.000458 秒。

整个计划

整个计划如下所示。

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);
        }
    }
}

与员工团队进行分配

本部分介绍了一个更常见的分配问题。在本题中，六个 员工分成了两个团队问题在于，你需要为 让团队之间的工作量均衡， 每个团队分别执行两项任务

有关此问题的 MIP 求解器解决方案，请参阅 向员工团队分配。

以下部分描述了一个程序，该程序使用 成本流求解器。

导入库

以下代码会导入所需的库。

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;

声明求解器

以下代码可创建最低成本流求解器。

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();

创建数据

以下代码会为程序创建数据。

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 };

工作器对应于节点 1-6。团队 A 由工作器 1、3 和 5 组成， 而团队 B 则由员工 2、4 和 6 组成。任务编号为 7 - 10。

在源节点和工作器之间有两个新节点：11 和 12。“节点 11”现为 节点 12 连接到团队 A 的节点， B 组，弧形容量为 1。 下图仅显示从源到工作器的节点和弧线。

平衡工作负载的关键是来源 0 连接到节点 11 以及 12 乘以容量 2 的弧线这意味着节点 11 和 12（因此 小组 A 和小组 B）的最大数据流为 2。 因此，每个团队最多只能执行两项任务。

创建限制条件

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]);
}

调用求解器

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();

程序的输出

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);
}

下面显示了程序的输出。

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

为团队 A 分配任务 9 和 10，而为团队 B 分配任务 7 和 8。

请注意，相较于 MIP 求解器， 大约需要 0.006 秒。

整个计划

整个计划如下所示。

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);
        }
    }
}