Un problème courant de planification est le magasin de tâches, dans lequel plusieurs tâches sont traitées sur plusieurs machines.
Chaque tâche se compose d'une séquence de tâches, qui doivent être exécutées dans un ordre donné, et chacune d'elles doit être traitée sur une machine spécifique.
Par exemple, le travail pourrait être la fabrication d'un seul article grand public, tel qu'une voiture.
Le problème consiste à planifier les tâches sur les machines de manière à réduire la length de la planification, c'est-à-dire le temps nécessaire pour effectuer toutes les tâches.
Le problème de l'atelier de vente présente plusieurs contraintes:
- Aucune tâche ne peut être démarrée tant que la tâche précédente de cette tâche n'est pas terminée.
- Une machine ne peut travailler que sur une tâche à la fois.
- Une fois lancée, une tâche doit être exécutée jusqu'à la fin.
Exemple de problème
Vous trouverez ci-dessous un exemple simple de problème de magasin de tâches, dans lequel chaque tâche est étiquetée par une paire de chiffres (m, p), où m est le nombre de machines sur lesquelles la tâche doit être traitée et p est le temps de traitement de cette tâche, c'est-à-dire la durée nécessaire. (La numérotation des tâches et des machines commence à 0.)
- tâche 0 = [(0, 3), (1, 2), (2, 2)]
- tâche 1 = [(0, 2), (2, 1), (1, 4)]
- tâche 2 = [(1, 4), (2, 3)]
Dans cet exemple, la tâche 0 comporte trois tâches. Le premier (0, 3) doit être traité sur la machine 0 en trois unités de temps. La seconde, (1, 2), doit être traitée sur la machine 1 en deux unités de temps, et ainsi de suite. Au total, il y a huit tâches.
Solution au problème
Une solution au problème de la tâche est d'attribuer une heure de début à chaque tâche, en respectant les contraintes indiquées ci-dessus.
Le diagramme ci-dessous présente une solution possible au problème :
Vous pouvez vérifier que les tâches de chaque tâche sont planifiées à des intervalles de temps qui ne se chevauchent pas, dans l'ordre indiqué par le problème.
La longueur de cette solution est de 12, ce qui correspond à la première fois que les trois tâches sont terminées. Cependant, comme vous le verrez ci-dessous, il ne s'agit pas de la solution optimale au problème.
Variables et contraintes liées au problème
Cette section explique comment configurer les variables et les contraintes liées au problème.
Tout d'abord, laissez task(i, j) indiquer la tâche jth de la séquence pour la tâche i. Par exemple, task(0, 2) indique la deuxième tâche pour la tâche 0, ce qui correspond à la paire (1, 2) dans la description du problème.
Ensuite, définissez ti, j comme heure de début de task(i, j). ti, j sont les variables du problème de l'atelier. Pour trouver une solution, vous devez déterminer les valeurs de ces variables qui répondent à l'exigence du problème.
Il existe deux types de contraintes pour le problème de l'offre d'emploi:
- Contraintes de priorité : elles découlent de la condition que pour deux tâches consécutives dans la même tâche, la première doit être terminée avant que la seconde ne puisse être lancée. Par exemple,
task(0, 2)ettask(0, 3)sont des tâches consécutives pour la tâche 0. L'heure de traitement detask(0, 2)étant égale à 2, l'heure de début pourtask(0, 3)doit être au moins deux heures après l'heure de début de la tâche 2. (La tâche 2 consiste peut-être à peindre une porte, et il faut deux heures pour que la peinture sèche.) Par conséquent, vous obtenez la contrainte suivante :t0, 2 + 2 <=t0, 3
- Aucune contrainte de chevauchement : ces contraintes résultent de la restriction qu'une machine ne peut pas exécuter sur deux tâches en même temps.
Par exemple, les tâches task(0, 2) et task(2, 1) sont traitées sur la machine 1.
Étant donné que leur temps de traitement est respectivement de 2 et 4, l'une des contraintes suivantes doit contenir :
t0, 2 + 2 <=t2, 1 (sitask(0, 2)est planifié avanttask(2, 1)) out2, 1 + 4 <=t0, 2 (sitask(2, 1)est planifié avanttask(0, 2)).
Objectif du problème
L'objectif du problème de l'offre d'emploi est de réduire au maximum la durée, c'est-à-dire la durée entre le début et la dernière heure de la tâche.
Une solution de programme
Les sections suivantes décrivent les principaux éléments d'un programme qui résout le problème de la recherche d'emploi.
Importer les bibliothèques
Le code suivant importe la bibliothèque requise.
Python
import collections from ortools.sat.python import cp_model
C++
#include <stdlib.h> #include <algorithm> #include <cstdint> #include <map> #include <numeric> #include <string> #include <tuple> #include <vector> #include "absl/strings/str_format.h" #include "ortools/base/logging.h" #include "ortools/sat/cp_model.h" #include "ortools/sat/cp_model.pb.h" #include "ortools/sat/cp_model_solver.h"
Java
import static java.lang.Math.max; import com.google.ortools.Loader; import com.google.ortools.sat.CpModel; import com.google.ortools.sat.CpSolver; import com.google.ortools.sat.CpSolverStatus; import com.google.ortools.sat.IntVar; import com.google.ortools.sat.IntervalVar; import com.google.ortools.sat.LinearExpr; import java.util.ArrayList; import java.util.Arrays; import java.util.Collections; import java.util.Comparator; import java.util.HashMap; import java.util.List; import java.util.Map; import java.util.stream.IntStream;
C#
using System; using System.Collections; using System.Collections.Generic; using System.Linq; using Google.OrTools.Sat;
Définir les données
Le programme définit ensuite les données correspondant au problème.
Python
jobs_data = [ # task = (machine_id, processing_time).
[(0, 3), (1, 2), (2, 2)], # Job0
[(0, 2), (2, 1), (1, 4)], # Job1
[(1, 4), (2, 3)], # Job2
]
machines_count = 1 + max(task[0] for job in jobs_data for task in job)
all_machines = range(machines_count)
# Computes horizon dynamically as the sum of all durations.
horizon = sum(task[1] for job in jobs_data for task in job)
C++
using Task = std::tuple<int64_t, int64_t>; // (machine_id, processing_time)
using Job = std::vector<Task>;
std::vector<Job> jobs_data = {
{{0, 3}, {1, 2}, {2, 2}}, // Job_0: Task_0 Task_1 Task_2
{{0, 2}, {2, 1}, {1, 4}}, // Job_1: Task_0 Task_1 Task_2
{{1, 4}, {2, 3}}, // Job_2: Task_0 Task_1
};
int64_t num_machines = 0;
for (const auto& job : jobs_data) {
for (const auto& [machine, _] : job) {
num_machines = std::max(num_machines, 1 + machine);
}
}
std::vector<int> all_machines(num_machines);
std::iota(all_machines.begin(), all_machines.end(), 0);
// Computes horizon dynamically as the sum of all durations.
int64_t horizon = 0;
for (const auto& job : jobs_data) {
for (const auto& [_, time] : job) {
horizon += time;
}
}
Java
class Task {
int machine;
int duration;
Task(int machine, int duration) {
this.machine = machine;
this.duration = duration;
}
}
final List<List<Task>> allJobs =
Arrays.asList(Arrays.asList(new Task(0, 3), new Task(1, 2), new Task(2, 2)), // Job0
Arrays.asList(new Task(0, 2), new Task(2, 1), new Task(1, 4)), // Job1
Arrays.asList(new Task(1, 4), new Task(2, 3)) // Job2
);
int numMachines = 1;
for (List<Task> job : allJobs) {
for (Task task : job) {
numMachines = max(numMachines, 1 + task.machine);
}
}
final int[] allMachines = IntStream.range(0, numMachines).toArray();
// Computes horizon dynamically as the sum of all durations.
int horizon = 0;
for (List<Task> job : allJobs) {
for (Task task : job) {
horizon += task.duration;
}
}
C#
var allJobs =
new[] {
new[] {
// job0
new { machine = 0, duration = 3 }, // task0
new { machine = 1, duration = 2 }, // task1
new { machine = 2, duration = 2 }, // task2
}
.ToList(),
new[] {
// job1
new { machine = 0, duration = 2 }, // task0
new { machine = 2, duration = 1 }, // task1
new { machine = 1, duration = 4 }, // task2
}
.ToList(),
new[] {
// job2
new { machine = 1, duration = 4 }, // task0
new { machine = 2, duration = 3 }, // task1
}
.ToList(),
}
.ToList();
int numMachines = 0;
foreach (var job in allJobs)
{
foreach (var task in job)
{
numMachines = Math.Max(numMachines, 1 + task.machine);
}
}
int[] allMachines = Enumerable.Range(0, numMachines).ToArray();
// Computes horizon dynamically as the sum of all durations.
int horizon = 0;
foreach (var job in allJobs)
{
foreach (var task in job)
{
horizon += task.duration;
}
}
Déclarer le modèle
Le code suivant déclare le modèle pour le problème.
Python
model = cp_model.CpModel()
C++
CpModelBuilder cp_model;
Java
CpModel model = new CpModel();
C#
CpModel model = new CpModel();
Définir les variables
Le code suivant définit les variables du problème.
Python
# Named tuple to store information about created variables.
task_type = collections.namedtuple("task_type", "start end interval")
# Named tuple to manipulate solution information.
assigned_task_type = collections.namedtuple(
"assigned_task_type", "start job index duration"
)
# Creates job intervals and add to the corresponding machine lists.
all_tasks = {}
machine_to_intervals = collections.defaultdict(list)
for job_id, job in enumerate(jobs_data):
for task_id, task in enumerate(job):
machine, duration = task
suffix = f"_{job_id}_{task_id}"
start_var = model.NewIntVar(0, horizon, "start" + suffix)
end_var = model.NewIntVar(0, horizon, "end" + suffix)
interval_var = model.NewIntervalVar(
start_var, duration, end_var, "interval" + suffix
)
all_tasks[job_id, task_id] = task_type(
start=start_var, end=end_var, interval=interval_var
)
machine_to_intervals[machine].append(interval_var)
C++
struct TaskType {
IntVar start;
IntVar end;
IntervalVar interval;
};
using TaskID = std::tuple<int, int>; // (job_id, task_id)
std::map<TaskID, TaskType> all_tasks;
std::map<int64_t, std::vector<IntervalVar>> machine_to_intervals;
for (int job_id = 0; job_id < jobs_data.size(); ++job_id) {
const auto& job = jobs_data[job_id];
for (int task_id = 0; task_id < job.size(); ++task_id) {
const auto [machine, duration] = job[task_id];
std::string suffix = absl::StrFormat("_%d_%d", job_id, task_id);
IntVar start = cp_model.NewIntVar({0, horizon})
.WithName(std::string("start") + suffix);
IntVar end = cp_model.NewIntVar({0, horizon})
.WithName(std::string("end") + suffix);
IntervalVar interval = cp_model.NewIntervalVar(start, duration, end)
.WithName(std::string("interval") + suffix);
TaskID key = std::make_tuple(job_id, task_id);
all_tasks.emplace(key, TaskType{/*.start=*/start,
/*.end=*/end,
/*.interval=*/interval});
machine_to_intervals[machine].push_back(interval);
}
}
Java
class TaskType {
IntVar start;
IntVar end;
IntervalVar interval;
}
Map<List<Integer>, TaskType> allTasks = new HashMap<>();
Map<Integer, List<IntervalVar>> machineToIntervals = new HashMap<>();
for (int jobID = 0; jobID < allJobs.size(); ++jobID) {
List<Task> job = allJobs.get(jobID);
for (int taskID = 0; taskID < job.size(); ++taskID) {
Task task = job.get(taskID);
String suffix = "_" + jobID + "_" + taskID;
TaskType taskType = new TaskType();
taskType.start = model.newIntVar(0, horizon, "start" + suffix);
taskType.end = model.newIntVar(0, horizon, "end" + suffix);
taskType.interval = model.newIntervalVar(
taskType.start, LinearExpr.constant(task.duration), taskType.end, "interval" + suffix);
List<Integer> key = Arrays.asList(jobID, taskID);
allTasks.put(key, taskType);
machineToIntervals.computeIfAbsent(task.machine, (Integer k) -> new ArrayList<>());
machineToIntervals.get(task.machine).add(taskType.interval);
}
}
C#
Dictionary<Tuple<int, int>, Tuple<IntVar, IntVar, IntervalVar>> allTasks =
new Dictionary<Tuple<int, int>, Tuple<IntVar, IntVar, IntervalVar>>(); // (start, end, duration)
Dictionary<int, List<IntervalVar>> machineToIntervals = new Dictionary<int, List<IntervalVar>>();
for (int jobID = 0; jobID < allJobs.Count(); ++jobID)
{
var job = allJobs[jobID];
for (int taskID = 0; taskID < job.Count(); ++taskID)
{
var task = job[taskID];
String suffix = $"_{jobID}_{taskID}";
IntVar start = model.NewIntVar(0, horizon, "start" + suffix);
IntVar end = model.NewIntVar(0, horizon, "end" + suffix);
IntervalVar interval = model.NewIntervalVar(start, task.duration, end, "interval" + suffix);
var key = Tuple.Create(jobID, taskID);
allTasks[key] = Tuple.Create(start, end, interval);
if (!machineToIntervals.ContainsKey(task.machine))
{
machineToIntervals.Add(task.machine, new List<IntervalVar>());
}
machineToIntervals[task.machine].Add(interval);
}
}
Pour chaque tâche et tâche, le programme utilise la méthode NewIntVar du résolveur afin de créer les variables:
start_var: heure de début de la tâche.end_var: heure de fin de la tâche.
La limite supérieure de start_var et end_var est horizon, soit la somme des temps de traitement de toutes les tâches de toutes les tâches.
horizon est suffisamment volumineux pour effectuer toutes les tâches pour la raison suivante : si vous planifiez les tâches à des intervalles qui ne se chevauchent pas (solution non optimale), la durée totale de la planification est exactement horizon. Ainsi, la durée de la solution optimale ne peut pas être supérieure à horizon.
Ensuite, le programme utilise la méthode NewIntervalVar pour créer une variable d'intervalle (dont la valeur correspond à un intervalle de temps variable) pour la tâche. Les entrées pour NewIntervalVar sont les suivantes:
start_var: variable de l'heure de début de la tâche.duration: durée de l'intervalle de temps de la tâche.end_var: variable pour l'heure de fin de la tâche.'interval_%i_%i' % (job, task_id)): nom de la variable d'intervalle.
Dans toutes les solutions, end_var moins start_var doit être égal à duration.
Définir les contraintes
Le code suivant définit les contraintes liées au problème.
Python
# Create and add disjunctive constraints.
for machine in all_machines:
model.AddNoOverlap(machine_to_intervals[machine])
# Precedences inside a job.
for job_id, job in enumerate(jobs_data):
for task_id in range(len(job) - 1):
model.Add(
all_tasks[job_id, task_id + 1].start >= all_tasks[job_id, task_id].end
)
C++
// Create and add disjunctive constraints.
for (const auto machine : all_machines) {
cp_model.AddNoOverlap(machine_to_intervals[machine]);
}
// Precedences inside a job.
for (int job_id = 0; job_id < jobs_data.size(); ++job_id) {
const auto& job = jobs_data[job_id];
for (int task_id = 0; task_id < job.size() - 1; ++task_id) {
TaskID key = std::make_tuple(job_id, task_id);
TaskID next_key = std::make_tuple(job_id, task_id + 1);
cp_model.AddGreaterOrEqual(all_tasks[next_key].start, all_tasks[key].end);
}
}
Java
// Create and add disjunctive constraints.
for (int machine : allMachines) {
List<IntervalVar> list = machineToIntervals.get(machine);
model.addNoOverlap(list);
}
// Precedences inside a job.
for (int jobID = 0; jobID < allJobs.size(); ++jobID) {
List<Task> job = allJobs.get(jobID);
for (int taskID = 0; taskID < job.size() - 1; ++taskID) {
List<Integer> prevKey = Arrays.asList(jobID, taskID);
List<Integer> nextKey = Arrays.asList(jobID, taskID + 1);
model.addGreaterOrEqual(allTasks.get(nextKey).start, allTasks.get(prevKey).end);
}
}
C#
// Create and add disjunctive constraints.
foreach (int machine in allMachines)
{
model.AddNoOverlap(machineToIntervals[machine]);
}
// Precedences inside a job.
for (int jobID = 0; jobID < allJobs.Count(); ++jobID)
{
var job = allJobs[jobID];
for (int taskID = 0; taskID < job.Count() - 1; ++taskID)
{
var key = Tuple.Create(jobID, taskID);
var nextKey = Tuple.Create(jobID, taskID + 1);
model.Add(allTasks[nextKey].Item1 >= allTasks[key].Item2);
}
}
Le programme utilise la méthode AddNoOverlap du résolveur pour créer des contraintes sans chevauchement, qui empêchent les tâches d'une même machine de se chevaucher dans le temps.
Ensuite, le programme ajoute les contraintes de priorité, qui empêchent les tâches consécutives pour une même tâche de se chevaucher dans le temps. Pour chaque tâche, la ligne model.Add(all_tasks[job, task_id + 1].start >= all_tasks[job, task_id].end) exige que l'heure de fin d'une tâche se produise avant l'heure de début de la tâche suivante.
Définir l'objectif
Le code suivant définit l'objectif du problème.
Python
# Makespan objective.
obj_var = model.NewIntVar(0, horizon, "makespan")
model.AddMaxEquality(
obj_var,
[all_tasks[job_id, len(job) - 1].end for job_id, job in enumerate(jobs_data)],
)
model.Minimize(obj_var)
C++
// Makespan objective.
IntVar obj_var = cp_model.NewIntVar({0, horizon}).WithName("makespan");
std::vector<IntVar> ends;
for (int job_id = 0; job_id < jobs_data.size(); ++job_id) {
const auto& job = jobs_data[job_id];
TaskID key = std::make_tuple(job_id, job.size() - 1);
ends.push_back(all_tasks[key].end);
}
cp_model.AddMaxEquality(obj_var, ends);
cp_model.Minimize(obj_var);
Java
// Makespan objective.
IntVar objVar = model.newIntVar(0, horizon, "makespan");
List<IntVar> ends = new ArrayList<>();
for (int jobID = 0; jobID < allJobs.size(); ++jobID) {
List<Task> job = allJobs.get(jobID);
List<Integer> key = Arrays.asList(jobID, job.size() - 1);
ends.add(allTasks.get(key).end);
}
model.addMaxEquality(objVar, ends);
model.minimize(objVar);
C#
// Makespan objective.
IntVar objVar = model.NewIntVar(0, horizon, "makespan");
List<IntVar> ends = new List<IntVar>();
for (int jobID = 0; jobID < allJobs.Count(); ++jobID)
{
var job = allJobs[jobID];
var key = Tuple.Create(jobID, job.Count() - 1);
ends.Add(allTasks[key].Item2);
}
model.AddMaxEquality(objVar, ends);
model.Minimize(objVar);
L'expression
model.AddMaxEquality(
obj_var,
[all_tasks[(job, len(jobs_data[job]) - 1)].end for job in all_jobs])
crée une variable obj_var dont la valeur correspond au nombre maximal de heures de fin pour toutes les tâches, à savoir le makespan.
Appeler le résolveur
Le code suivant appelle le résolveur.
Python
solver = cp_model.CpSolver() status = solver.Solve(model)
C++
const CpSolverResponse response = Solve(cp_model.Build());
Java
CpSolver solver = new CpSolver(); CpSolverStatus status = solver.solve(model);
C#
CpSolver solver = new CpSolver();
CpSolverStatus status = solver.Solve(model);
Console.WriteLine($"Solve status: {status}");
Afficher les résultats
Le code suivant affiche les résultats, y compris la planification optimale et les intervalles de tâches.
Python
if status == cp_model.OPTIMAL or status == cp_model.FEASIBLE:
print("Solution:")
# Create one list of assigned tasks per machine.
assigned_jobs = collections.defaultdict(list)
for job_id, job in enumerate(jobs_data):
for task_id, task in enumerate(job):
machine = task[0]
assigned_jobs[machine].append(
assigned_task_type(
start=solver.Value(all_tasks[job_id, task_id].start),
job=job_id,
index=task_id,
duration=task[1],
)
)
# Create per machine output lines.
output = ""
for machine in all_machines:
# Sort by starting time.
assigned_jobs[machine].sort()
sol_line_tasks = "Machine " + str(machine) + ": "
sol_line = " "
for assigned_task in assigned_jobs[machine]:
name = f"job_{assigned_task.job}_task_{assigned_task.index}"
# Add spaces to output to align columns.
sol_line_tasks += f"{name:15}"
start = assigned_task.start
duration = assigned_task.duration
sol_tmp = f"[{start},{start + duration}]"
# Add spaces to output to align columns.
sol_line += f"{sol_tmp:15}"
sol_line += "\n"
sol_line_tasks += "\n"
output += sol_line_tasks
output += sol_line
# Finally print the solution found.
print(f"Optimal Schedule Length: {solver.ObjectiveValue()}")
print(output)
else:
print("No solution found.")
C++
if (response.status() == CpSolverStatus::OPTIMAL ||
response.status() == CpSolverStatus::FEASIBLE) {
LOG(INFO) << "Solution:";
// create one list of assigned tasks per machine.
struct AssignedTaskType {
int job_id;
int task_id;
int64_t start;
int64_t duration;
bool operator<(const AssignedTaskType& rhs) const {
return std::tie(this->start, this->duration) <
std::tie(rhs.start, rhs.duration);
}
};
std::map<int64_t, std::vector<AssignedTaskType>> assigned_jobs;
for (int job_id = 0; job_id < jobs_data.size(); ++job_id) {
const auto& job = jobs_data[job_id];
for (int task_id = 0; task_id < job.size(); ++task_id) {
const auto [machine, duration] = job[task_id];
TaskID key = std::make_tuple(job_id, task_id);
int64_t start = SolutionIntegerValue(response, all_tasks[key].start);
assigned_jobs[machine].push_back(
AssignedTaskType{/*.job_id=*/job_id,
/*.task_id=*/task_id,
/*.start=*/start,
/*.duration=*/duration});
}
}
// Create per machine output lines.
std::string output = "";
for (const auto machine : all_machines) {
// Sort by starting time.
std::sort(assigned_jobs[machine].begin(), assigned_jobs[machine].end());
std::string sol_line_tasks = "Machine " + std::to_string(machine) + ": ";
std::string sol_line = " ";
for (const auto& assigned_task : assigned_jobs[machine]) {
std::string name = absl::StrFormat(
"job_%d_task_%d", assigned_task.job_id, assigned_task.task_id);
// Add spaces to output to align columns.
sol_line_tasks += absl::StrFormat("%-15s", name);
int64_t start = assigned_task.start;
int64_t duration = assigned_task.duration;
std::string sol_tmp =
absl::StrFormat("[%i,%i]", start, start + duration);
// Add spaces to output to align columns.
sol_line += absl::StrFormat("%-15s", sol_tmp);
}
output += sol_line_tasks + "\n";
output += sol_line + "\n";
}
// Finally print the solution found.
LOG(INFO) << "Optimal Schedule Length: " << response.objective_value();
LOG(INFO) << "\n" << output;
} else {
LOG(INFO) << "No solution found.";
}
Java
if (status == CpSolverStatus.OPTIMAL || status == CpSolverStatus.FEASIBLE) {
class AssignedTask {
int jobID;
int taskID;
int start;
int duration;
// Ctor
AssignedTask(int jobID, int taskID, int start, int duration) {
this.jobID = jobID;
this.taskID = taskID;
this.start = start;
this.duration = duration;
}
}
class SortTasks implements Comparator<AssignedTask> {
@Override
public int compare(AssignedTask a, AssignedTask b) {
if (a.start != b.start) {
return a.start - b.start;
} else {
return a.duration - b.duration;
}
}
}
System.out.println("Solution:");
// Create one list of assigned tasks per machine.
Map<Integer, List<AssignedTask>> assignedJobs = new HashMap<>();
for (int jobID = 0; jobID < allJobs.size(); ++jobID) {
List<Task> job = allJobs.get(jobID);
for (int taskID = 0; taskID < job.size(); ++taskID) {
Task task = job.get(taskID);
List<Integer> key = Arrays.asList(jobID, taskID);
AssignedTask assignedTask = new AssignedTask(
jobID, taskID, (int) solver.value(allTasks.get(key).start), task.duration);
assignedJobs.computeIfAbsent(task.machine, (Integer k) -> new ArrayList<>());
assignedJobs.get(task.machine).add(assignedTask);
}
}
// Create per machine output lines.
String output = "";
for (int machine : allMachines) {
// Sort by starting time.
Collections.sort(assignedJobs.get(machine), new SortTasks());
String solLineTasks = "Machine " + machine + ": ";
String solLine = " ";
for (AssignedTask assignedTask : assignedJobs.get(machine)) {
String name = "job_" + assignedTask.jobID + "_task_" + assignedTask.taskID;
// Add spaces to output to align columns.
solLineTasks += String.format("%-15s", name);
String solTmp =
"[" + assignedTask.start + "," + (assignedTask.start + assignedTask.duration) + "]";
// Add spaces to output to align columns.
solLine += String.format("%-15s", solTmp);
}
output += solLineTasks + "%n";
output += solLine + "%n";
}
System.out.printf("Optimal Schedule Length: %f%n", solver.objectiveValue());
System.out.printf(output);
} else {
System.out.println("No solution found.");
}
C#
if (status == CpSolverStatus.Optimal || status == CpSolverStatus.Feasible)
{
Console.WriteLine("Solution:");
Dictionary<int, List<AssignedTask>> assignedJobs = new Dictionary<int, List<AssignedTask>>();
for (int jobID = 0; jobID < allJobs.Count(); ++jobID)
{
var job = allJobs[jobID];
for (int taskID = 0; taskID < job.Count(); ++taskID)
{
var task = job[taskID];
var key = Tuple.Create(jobID, taskID);
int start = (int)solver.Value(allTasks[key].Item1);
if (!assignedJobs.ContainsKey(task.machine))
{
assignedJobs.Add(task.machine, new List<AssignedTask>());
}
assignedJobs[task.machine].Add(new AssignedTask(jobID, taskID, start, task.duration));
}
}
// Create per machine output lines.
String output = "";
foreach (int machine in allMachines)
{
// Sort by starting time.
assignedJobs[machine].Sort();
String solLineTasks = $"Machine {machine}: ";
String solLine = " ";
foreach (var assignedTask in assignedJobs[machine])
{
String name = $"job_{assignedTask.jobID}_task_{assignedTask.taskID}";
// Add spaces to output to align columns.
solLineTasks += $"{name,-15}";
String solTmp = $"[{assignedTask.start},{assignedTask.start+assignedTask.duration}]";
// Add spaces to output to align columns.
solLine += $"{solTmp,-15}";
}
output += solLineTasks + "\n";
output += solLine + "\n";
}
// Finally print the solution found.
Console.WriteLine($"Optimal Schedule Length: {solver.ObjectiveValue}");
Console.WriteLine($"\n{output}");
}
else
{
Console.WriteLine("No solution found.");
}
Le calendrier optimal est indiqué ci-dessous:
Optimal Schedule Length: 11
Machine 0: job_0_0 job_1_0
[0,3] [3,5]
Machine 1: job_2_0 job_0_1 job_1_2
[0,4] [4,6] [7,11]
Machine 2: job_1_1 job_0_2 job_2_1
[5,6] [6,8] [8,11]
Lorsqu'ils examinent la machine 1 avec des yeux oculaires, la tâche 1 peut se demander à la date 7 et non 6. Ce sont toutes deux des solutions valides, mais n'oubliez pas que l'objectif est de minimiser la portée. Le déplacement de job_1_2 plus tôt ne réduira pas la durée de la tâche. Par conséquent, les deux solutions sont égales du point de vue du résolveur.
Intégralité du programme
Enfin, voici le programme complet pour le problème de l'emploi.
Python
"""Minimal jobshop example."""
import collections
from ortools.sat.python import cp_model
def main():
"""Minimal jobshop problem."""
# Data.
jobs_data = [ # task = (machine_id, processing_time).
[(0, 3), (1, 2), (2, 2)], # Job0
[(0, 2), (2, 1), (1, 4)], # Job1
[(1, 4), (2, 3)], # Job2
]
machines_count = 1 + max(task[0] for job in jobs_data for task in job)
all_machines = range(machines_count)
# Computes horizon dynamically as the sum of all durations.
horizon = sum(task[1] for job in jobs_data for task in job)
# Create the model.
model = cp_model.CpModel()
# Named tuple to store information about created variables.
task_type = collections.namedtuple("task_type", "start end interval")
# Named tuple to manipulate solution information.
assigned_task_type = collections.namedtuple(
"assigned_task_type", "start job index duration"
)
# Creates job intervals and add to the corresponding machine lists.
all_tasks = {}
machine_to_intervals = collections.defaultdict(list)
for job_id, job in enumerate(jobs_data):
for task_id, task in enumerate(job):
machine, duration = task
suffix = f"_{job_id}_{task_id}"
start_var = model.NewIntVar(0, horizon, "start" + suffix)
end_var = model.NewIntVar(0, horizon, "end" + suffix)
interval_var = model.NewIntervalVar(
start_var, duration, end_var, "interval" + suffix
)
all_tasks[job_id, task_id] = task_type(
start=start_var, end=end_var, interval=interval_var
)
machine_to_intervals[machine].append(interval_var)
# Create and add disjunctive constraints.
for machine in all_machines:
model.AddNoOverlap(machine_to_intervals[machine])
# Precedences inside a job.
for job_id, job in enumerate(jobs_data):
for task_id in range(len(job) - 1):
model.Add(
all_tasks[job_id, task_id + 1].start >= all_tasks[job_id, task_id].end
)
# Makespan objective.
obj_var = model.NewIntVar(0, horizon, "makespan")
model.AddMaxEquality(
obj_var,
[all_tasks[job_id, len(job) - 1].end for job_id, job in enumerate(jobs_data)],
)
model.Minimize(obj_var)
# Creates the solver and solve.
solver = cp_model.CpSolver()
status = solver.Solve(model)
if status == cp_model.OPTIMAL or status == cp_model.FEASIBLE:
print("Solution:")
# Create one list of assigned tasks per machine.
assigned_jobs = collections.defaultdict(list)
for job_id, job in enumerate(jobs_data):
for task_id, task in enumerate(job):
machine = task[0]
assigned_jobs[machine].append(
assigned_task_type(
start=solver.Value(all_tasks[job_id, task_id].start),
job=job_id,
index=task_id,
duration=task[1],
)
)
# Create per machine output lines.
output = ""
for machine in all_machines:
# Sort by starting time.
assigned_jobs[machine].sort()
sol_line_tasks = "Machine " + str(machine) + ": "
sol_line = " "
for assigned_task in assigned_jobs[machine]:
name = f"job_{assigned_task.job}_task_{assigned_task.index}"
# Add spaces to output to align columns.
sol_line_tasks += f"{name:15}"
start = assigned_task.start
duration = assigned_task.duration
sol_tmp = f"[{start},{start + duration}]"
# Add spaces to output to align columns.
sol_line += f"{sol_tmp:15}"
sol_line += "\n"
sol_line_tasks += "\n"
output += sol_line_tasks
output += sol_line
# Finally print the solution found.
print(f"Optimal Schedule Length: {solver.ObjectiveValue()}")
print(output)
else:
print("No solution found.")
# Statistics.
print("\nStatistics")
print(f" - conflicts: {solver.NumConflicts()}")
print(f" - branches : {solver.NumBranches()}")
print(f" - wall time: {solver.WallTime()}s")
if __name__ == "__main__":
main()
C++
// Nurse scheduling problem with shift requests.
#include <stdlib.h>
#include <algorithm>
#include <cstdint>
#include <map>
#include <numeric>
#include <string>
#include <tuple>
#include <vector>
#include "absl/strings/str_format.h"
#include "ortools/base/logging.h"
#include "ortools/sat/cp_model.h"
#include "ortools/sat/cp_model.pb.h"
#include "ortools/sat/cp_model_solver.h"
namespace operations_research {
namespace sat {
void MinimalJobshopSat() {
using Task = std::tuple<int64_t, int64_t>; // (machine_id, processing_time)
using Job = std::vector<Task>;
std::vector<Job> jobs_data = {
{{0, 3}, {1, 2}, {2, 2}}, // Job_0: Task_0 Task_1 Task_2
{{0, 2}, {2, 1}, {1, 4}}, // Job_1: Task_0 Task_1 Task_2
{{1, 4}, {2, 3}}, // Job_2: Task_0 Task_1
};
int64_t num_machines = 0;
for (const auto& job : jobs_data) {
for (const auto& [machine, _] : job) {
num_machines = std::max(num_machines, 1 + machine);
}
}
std::vector<int> all_machines(num_machines);
std::iota(all_machines.begin(), all_machines.end(), 0);
// Computes horizon dynamically as the sum of all durations.
int64_t horizon = 0;
for (const auto& job : jobs_data) {
for (const auto& [_, time] : job) {
horizon += time;
}
}
// Creates the model.
CpModelBuilder cp_model;
struct TaskType {
IntVar start;
IntVar end;
IntervalVar interval;
};
using TaskID = std::tuple<int, int>; // (job_id, task_id)
std::map<TaskID, TaskType> all_tasks;
std::map<int64_t, std::vector<IntervalVar>> machine_to_intervals;
for (int job_id = 0; job_id < jobs_data.size(); ++job_id) {
const auto& job = jobs_data[job_id];
for (int task_id = 0; task_id < job.size(); ++task_id) {
const auto [machine, duration] = job[task_id];
std::string suffix = absl::StrFormat("_%d_%d", job_id, task_id);
IntVar start = cp_model.NewIntVar({0, horizon})
.WithName(std::string("start") + suffix);
IntVar end = cp_model.NewIntVar({0, horizon})
.WithName(std::string("end") + suffix);
IntervalVar interval = cp_model.NewIntervalVar(start, duration, end)
.WithName(std::string("interval") + suffix);
TaskID key = std::make_tuple(job_id, task_id);
all_tasks.emplace(key, TaskType{/*.start=*/start,
/*.end=*/end,
/*.interval=*/interval});
machine_to_intervals[machine].push_back(interval);
}
}
// Create and add disjunctive constraints.
for (const auto machine : all_machines) {
cp_model.AddNoOverlap(machine_to_intervals[machine]);
}
// Precedences inside a job.
for (int job_id = 0; job_id < jobs_data.size(); ++job_id) {
const auto& job = jobs_data[job_id];
for (int task_id = 0; task_id < job.size() - 1; ++task_id) {
TaskID key = std::make_tuple(job_id, task_id);
TaskID next_key = std::make_tuple(job_id, task_id + 1);
cp_model.AddGreaterOrEqual(all_tasks[next_key].start, all_tasks[key].end);
}
}
// Makespan objective.
IntVar obj_var = cp_model.NewIntVar({0, horizon}).WithName("makespan");
std::vector<IntVar> ends;
for (int job_id = 0; job_id < jobs_data.size(); ++job_id) {
const auto& job = jobs_data[job_id];
TaskID key = std::make_tuple(job_id, job.size() - 1);
ends.push_back(all_tasks[key].end);
}
cp_model.AddMaxEquality(obj_var, ends);
cp_model.Minimize(obj_var);
const CpSolverResponse response = Solve(cp_model.Build());
if (response.status() == CpSolverStatus::OPTIMAL ||
response.status() == CpSolverStatus::FEASIBLE) {
LOG(INFO) << "Solution:";
// create one list of assigned tasks per machine.
struct AssignedTaskType {
int job_id;
int task_id;
int64_t start;
int64_t duration;
bool operator<(const AssignedTaskType& rhs) const {
return std::tie(this->start, this->duration) <
std::tie(rhs.start, rhs.duration);
}
};
std::map<int64_t, std::vector<AssignedTaskType>> assigned_jobs;
for (int job_id = 0; job_id < jobs_data.size(); ++job_id) {
const auto& job = jobs_data[job_id];
for (int task_id = 0; task_id < job.size(); ++task_id) {
const auto [machine, duration] = job[task_id];
TaskID key = std::make_tuple(job_id, task_id);
int64_t start = SolutionIntegerValue(response, all_tasks[key].start);
assigned_jobs[machine].push_back(
AssignedTaskType{/*.job_id=*/job_id,
/*.task_id=*/task_id,
/*.start=*/start,
/*.duration=*/duration});
}
}
// Create per machine output lines.
std::string output = "";
for (const auto machine : all_machines) {
// Sort by starting time.
std::sort(assigned_jobs[machine].begin(), assigned_jobs[machine].end());
std::string sol_line_tasks = "Machine " + std::to_string(machine) + ": ";
std::string sol_line = " ";
for (const auto& assigned_task : assigned_jobs[machine]) {
std::string name = absl::StrFormat(
"job_%d_task_%d", assigned_task.job_id, assigned_task.task_id);
// Add spaces to output to align columns.
sol_line_tasks += absl::StrFormat("%-15s", name);
int64_t start = assigned_task.start;
int64_t duration = assigned_task.duration;
std::string sol_tmp =
absl::StrFormat("[%i,%i]", start, start + duration);
// Add spaces to output to align columns.
sol_line += absl::StrFormat("%-15s", sol_tmp);
}
output += sol_line_tasks + "\n";
output += sol_line + "\n";
}
// Finally print the solution found.
LOG(INFO) << "Optimal Schedule Length: " << response.objective_value();
LOG(INFO) << "\n" << output;
} else {
LOG(INFO) << "No solution found.";
}
// Statistics.
LOG(INFO) << "Statistics";
LOG(INFO) << CpSolverResponseStats(response);
}
} // namespace sat
} // namespace operations_research
int main() {
operations_research::sat::MinimalJobshopSat();
return EXIT_SUCCESS;
}
Java
package com.google.ortools.sat.samples;
import static java.lang.Math.max;
import com.google.ortools.Loader;
import com.google.ortools.sat.CpModel;
import com.google.ortools.sat.CpSolver;
import com.google.ortools.sat.CpSolverStatus;
import com.google.ortools.sat.IntVar;
import com.google.ortools.sat.IntervalVar;
import com.google.ortools.sat.LinearExpr;
import java.util.ArrayList;
import java.util.Arrays;
import java.util.Collections;
import java.util.Comparator;
import java.util.HashMap;
import java.util.List;
import java.util.Map;
import java.util.stream.IntStream;
/** Minimal Jobshop problem. */
public class MinimalJobshopSat {
public static void main(String[] args) {
Loader.loadNativeLibraries();
class Task {
int machine;
int duration;
Task(int machine, int duration) {
this.machine = machine;
this.duration = duration;
}
}
final List<List<Task>> allJobs =
Arrays.asList(Arrays.asList(new Task(0, 3), new Task(1, 2), new Task(2, 2)), // Job0
Arrays.asList(new Task(0, 2), new Task(2, 1), new Task(1, 4)), // Job1
Arrays.asList(new Task(1, 4), new Task(2, 3)) // Job2
);
int numMachines = 1;
for (List<Task> job : allJobs) {
for (Task task : job) {
numMachines = max(numMachines, 1 + task.machine);
}
}
final int[] allMachines = IntStream.range(0, numMachines).toArray();
// Computes horizon dynamically as the sum of all durations.
int horizon = 0;
for (List<Task> job : allJobs) {
for (Task task : job) {
horizon += task.duration;
}
}
// Creates the model.
CpModel model = new CpModel();
class TaskType {
IntVar start;
IntVar end;
IntervalVar interval;
}
Map<List<Integer>, TaskType> allTasks = new HashMap<>();
Map<Integer, List<IntervalVar>> machineToIntervals = new HashMap<>();
for (int jobID = 0; jobID < allJobs.size(); ++jobID) {
List<Task> job = allJobs.get(jobID);
for (int taskID = 0; taskID < job.size(); ++taskID) {
Task task = job.get(taskID);
String suffix = "_" + jobID + "_" + taskID;
TaskType taskType = new TaskType();
taskType.start = model.newIntVar(0, horizon, "start" + suffix);
taskType.end = model.newIntVar(0, horizon, "end" + suffix);
taskType.interval = model.newIntervalVar(
taskType.start, LinearExpr.constant(task.duration), taskType.end, "interval" + suffix);
List<Integer> key = Arrays.asList(jobID, taskID);
allTasks.put(key, taskType);
machineToIntervals.computeIfAbsent(task.machine, (Integer k) -> new ArrayList<>());
machineToIntervals.get(task.machine).add(taskType.interval);
}
}
// Create and add disjunctive constraints.
for (int machine : allMachines) {
List<IntervalVar> list = machineToIntervals.get(machine);
model.addNoOverlap(list);
}
// Precedences inside a job.
for (int jobID = 0; jobID < allJobs.size(); ++jobID) {
List<Task> job = allJobs.get(jobID);
for (int taskID = 0; taskID < job.size() - 1; ++taskID) {
List<Integer> prevKey = Arrays.asList(jobID, taskID);
List<Integer> nextKey = Arrays.asList(jobID, taskID + 1);
model.addGreaterOrEqual(allTasks.get(nextKey).start, allTasks.get(prevKey).end);
}
}
// Makespan objective.
IntVar objVar = model.newIntVar(0, horizon, "makespan");
List<IntVar> ends = new ArrayList<>();
for (int jobID = 0; jobID < allJobs.size(); ++jobID) {
List<Task> job = allJobs.get(jobID);
List<Integer> key = Arrays.asList(jobID, job.size() - 1);
ends.add(allTasks.get(key).end);
}
model.addMaxEquality(objVar, ends);
model.minimize(objVar);
// Creates a solver and solves the model.
CpSolver solver = new CpSolver();
CpSolverStatus status = solver.solve(model);
if (status == CpSolverStatus.OPTIMAL || status == CpSolverStatus.FEASIBLE) {
class AssignedTask {
int jobID;
int taskID;
int start;
int duration;
// Ctor
AssignedTask(int jobID, int taskID, int start, int duration) {
this.jobID = jobID;
this.taskID = taskID;
this.start = start;
this.duration = duration;
}
}
class SortTasks implements Comparator<AssignedTask> {
@Override
public int compare(AssignedTask a, AssignedTask b) {
if (a.start != b.start) {
return a.start - b.start;
} else {
return a.duration - b.duration;
}
}
}
System.out.println("Solution:");
// Create one list of assigned tasks per machine.
Map<Integer, List<AssignedTask>> assignedJobs = new HashMap<>();
for (int jobID = 0; jobID < allJobs.size(); ++jobID) {
List<Task> job = allJobs.get(jobID);
for (int taskID = 0; taskID < job.size(); ++taskID) {
Task task = job.get(taskID);
List<Integer> key = Arrays.asList(jobID, taskID);
AssignedTask assignedTask = new AssignedTask(
jobID, taskID, (int) solver.value(allTasks.get(key).start), task.duration);
assignedJobs.computeIfAbsent(task.machine, (Integer k) -> new ArrayList<>());
assignedJobs.get(task.machine).add(assignedTask);
}
}
// Create per machine output lines.
String output = "";
for (int machine : allMachines) {
// Sort by starting time.
Collections.sort(assignedJobs.get(machine), new SortTasks());
String solLineTasks = "Machine " + machine + ": ";
String solLine = " ";
for (AssignedTask assignedTask : assignedJobs.get(machine)) {
String name = "job_" + assignedTask.jobID + "_task_" + assignedTask.taskID;
// Add spaces to output to align columns.
solLineTasks += String.format("%-15s", name);
String solTmp =
"[" + assignedTask.start + "," + (assignedTask.start + assignedTask.duration) + "]";
// Add spaces to output to align columns.
solLine += String.format("%-15s", solTmp);
}
output += solLineTasks + "%n";
output += solLine + "%n";
}
System.out.printf("Optimal Schedule Length: %f%n", solver.objectiveValue());
System.out.printf(output);
} else {
System.out.println("No solution found.");
}
// Statistics.
System.out.println("Statistics");
System.out.printf(" conflicts: %d%n", solver.numConflicts());
System.out.printf(" branches : %d%n", solver.numBranches());
System.out.printf(" wall time: %f s%n", solver.wallTime());
}
private MinimalJobshopSat() {}
}
C#
using System;
using System.Collections;
using System.Collections.Generic;
using System.Linq;
using Google.OrTools.Sat;
public class ScheduleRequestsSat
{
private class AssignedTask : IComparable
{
public int jobID;
public int taskID;
public int start;
public int duration;
public AssignedTask(int jobID, int taskID, int start, int duration)
{
this.jobID = jobID;
this.taskID = taskID;
this.start = start;
this.duration = duration;
}
public int CompareTo(object obj)
{
if (obj == null)
return 1;
AssignedTask otherTask = obj as AssignedTask;
if (otherTask != null)
{
if (this.start != otherTask.start)
return this.start.CompareTo(otherTask.start);
else
return this.duration.CompareTo(otherTask.duration);
}
else
throw new ArgumentException("Object is not a Temperature");
}
}
public static void Main(String[] args)
{
var allJobs =
new[] {
new[] {
// job0
new { machine = 0, duration = 3 }, // task0
new { machine = 1, duration = 2 }, // task1
new { machine = 2, duration = 2 }, // task2
}
.ToList(),
new[] {
// job1
new { machine = 0, duration = 2 }, // task0
new { machine = 2, duration = 1 }, // task1
new { machine = 1, duration = 4 }, // task2
}
.ToList(),
new[] {
// job2
new { machine = 1, duration = 4 }, // task0
new { machine = 2, duration = 3 }, // task1
}
.ToList(),
}
.ToList();
int numMachines = 0;
foreach (var job in allJobs)
{
foreach (var task in job)
{
numMachines = Math.Max(numMachines, 1 + task.machine);
}
}
int[] allMachines = Enumerable.Range(0, numMachines).ToArray();
// Computes horizon dynamically as the sum of all durations.
int horizon = 0;
foreach (var job in allJobs)
{
foreach (var task in job)
{
horizon += task.duration;
}
}
// Creates the model.
CpModel model = new CpModel();
Dictionary<Tuple<int, int>, Tuple<IntVar, IntVar, IntervalVar>> allTasks =
new Dictionary<Tuple<int, int>, Tuple<IntVar, IntVar, IntervalVar>>(); // (start, end, duration)
Dictionary<int, List<IntervalVar>> machineToIntervals = new Dictionary<int, List<IntervalVar>>();
for (int jobID = 0; jobID < allJobs.Count(); ++jobID)
{
var job = allJobs[jobID];
for (int taskID = 0; taskID < job.Count(); ++taskID)
{
var task = job[taskID];
String suffix = $"_{jobID}_{taskID}";
IntVar start = model.NewIntVar(0, horizon, "start" + suffix);
IntVar end = model.NewIntVar(0, horizon, "end" + suffix);
IntervalVar interval = model.NewIntervalVar(start, task.duration, end, "interval" + suffix);
var key = Tuple.Create(jobID, taskID);
allTasks[key] = Tuple.Create(start, end, interval);
if (!machineToIntervals.ContainsKey(task.machine))
{
machineToIntervals.Add(task.machine, new List<IntervalVar>());
}
machineToIntervals[task.machine].Add(interval);
}
}
// Create and add disjunctive constraints.
foreach (int machine in allMachines)
{
model.AddNoOverlap(machineToIntervals[machine]);
}
// Precedences inside a job.
for (int jobID = 0; jobID < allJobs.Count(); ++jobID)
{
var job = allJobs[jobID];
for (int taskID = 0; taskID < job.Count() - 1; ++taskID)
{
var key = Tuple.Create(jobID, taskID);
var nextKey = Tuple.Create(jobID, taskID + 1);
model.Add(allTasks[nextKey].Item1 >= allTasks[key].Item2);
}
}
// Makespan objective.
IntVar objVar = model.NewIntVar(0, horizon, "makespan");
List<IntVar> ends = new List<IntVar>();
for (int jobID = 0; jobID < allJobs.Count(); ++jobID)
{
var job = allJobs[jobID];
var key = Tuple.Create(jobID, job.Count() - 1);
ends.Add(allTasks[key].Item2);
}
model.AddMaxEquality(objVar, ends);
model.Minimize(objVar);
// Solve
CpSolver solver = new CpSolver();
CpSolverStatus status = solver.Solve(model);
Console.WriteLine($"Solve status: {status}");
if (status == CpSolverStatus.Optimal || status == CpSolverStatus.Feasible)
{
Console.WriteLine("Solution:");
Dictionary<int, List<AssignedTask>> assignedJobs = new Dictionary<int, List<AssignedTask>>();
for (int jobID = 0; jobID < allJobs.Count(); ++jobID)
{
var job = allJobs[jobID];
for (int taskID = 0; taskID < job.Count(); ++taskID)
{
var task = job[taskID];
var key = Tuple.Create(jobID, taskID);
int start = (int)solver.Value(allTasks[key].Item1);
if (!assignedJobs.ContainsKey(task.machine))
{
assignedJobs.Add(task.machine, new List<AssignedTask>());
}
assignedJobs[task.machine].Add(new AssignedTask(jobID, taskID, start, task.duration));
}
}
// Create per machine output lines.
String output = "";
foreach (int machine in allMachines)
{
// Sort by starting time.
assignedJobs[machine].Sort();
String solLineTasks = $"Machine {machine}: ";
String solLine = " ";
foreach (var assignedTask in assignedJobs[machine])
{
String name = $"job_{assignedTask.jobID}_task_{assignedTask.taskID}";
// Add spaces to output to align columns.
solLineTasks += $"{name,-15}";
String solTmp = $"[{assignedTask.start},{assignedTask.start+assignedTask.duration}]";
// Add spaces to output to align columns.
solLine += $"{solTmp,-15}";
}
output += solLineTasks + "\n";
output += solLine + "\n";
}
// Finally print the solution found.
Console.WriteLine($"Optimal Schedule Length: {solver.ObjectiveValue}");
Console.WriteLine($"\n{output}");
}
else
{
Console.WriteLine("No solution found.");
}
Console.WriteLine("Statistics");
Console.WriteLine($" conflicts: {solver.NumConflicts()}");
Console.WriteLine($" branches : {solver.NumBranches()}");
Console.WriteLine($" wall time: {solver.WallTime()}s");
}
}