Flexible Job Shop Scheduling Problem with Sequence-Dependent Setup Times (FJSP-SDST)¶
Principles learned¶
Add multiple list decision variables
Use the find operator
Use interval decision variables
Order interval decision variables by pairing them up with a list variable
Problem¶
In the flexible job shop scheduling problem with sequence-dependent setup times (FJSP-SDST), a set of jobs has to be processed on the machines in the shop. Each job consists of an ordered sequence of tasks (called operations), and each operation must be performed by one of the machines compatible with that operation. An operation cannot begin until the previous operation in the job is completed. Each operation has a given processing time that depends on the chosen machine. Each machine can only process one operation at a time and must be set up between two consecutive operations. These setup times depend both on the operations and the machine considered.
The goal is to find a sequence of jobs that minimizes the makespan: the time when all jobs have been processed.
Download the exampleData¶
The format of the data files is as follows:
First line: number of jobs, number of machines (+ average number of machines per operations, not needed)
From the second line, for each job:
Number of operations in that job
For each operation:
Number of machines compatible with this operation
For each compatible machine: a pair of numbers (machine, processing time)
For each machine and each operation:
Setup time between the first operation and every other operation on the machine considered.
Program¶
The model is an extension from the Flexible Job Shop Problem with the use of the setup times on the machines. The decision variables are the following: we represent the time ranges of the tasks by interval decision variables and we model the order of the operations performed on each machine by a list decision variable.
Each operation of each job must be processed on one and only one machine,
hence the partition
operator on the lists.
The disjunctive resource contraints between tasks on a machine guarantee that an operation starts on a machine only after the previous operation is done and the setup between the two operations is completed.
The constraints of compatibility of the machines are modeled in the same way as for the flexible job shop problem, and the makespan to be minimized is the time when all tasks have been processed.
- Execution:
- localsolver flexible_jobshop_setup.lsp inFileName=instances/Fattahi_setup_01.fjs [outFileName=] [lsTimeLimit=]
use io;
/* Read instance data */
function input() {
local usage = "Usage: localsolver flexible_jobshop_setup.lsp inFileName=instanceFile"
+ " [outFileName=outputFile] [lsTimeLimit=timeLimit]";
if (inFileName == nil) throw usage;
// Constant for incompatible machines
INFINITE = 1000000;
inFile = io.openRead(inFileName);
// Number of jobs
nbJobs = inFile.readInt();
// Number of machines
nbMachines = inFile.readInt();
inFile.readln(); // skip last number
// Number of tasks
nbTasks = 0;
processingTime = {};
// Processing time for each task, for each machine
taskProcessingTime = {};
// For each job, for each operation, the corresponding task id
jobOperationTask = {};
for [j in 0...nbJobs] {
// Number of operations for each job
nbOperations[j] = inFile.readInt();
for [o in 0...nbOperations[j]] {
local nbMachinesOperation = inFile.readInt();
for [i in 0...nbMachinesOperation] {
local machine = inFile.readInt() - 1;
local time = inFile.readInt();
processingTime[j][o][machine] = time;
taskProcessingTime[nbTasks][machine] = time;
}
jobOperationTask[j][o] = nbTasks;
nbTasks += 1;
}
}
// Setup time between every two consecutive tasks, for each machine
taskSetupTime[m in 0...nbMachines][i in 0...nbTasks][j in 0...nbTasks] = inFile.readInt();
maxSetup = 0;
for [m in 0...nbMachines][i in 0...nbTasks][j in 0...nbTasks] {
if (taskSetupTime[m][i][j] != INFINITE && taskSetupTime[m][i][j] > maxSetup) {
maxSetup = taskSetupTime[m][i][j];
}
}
inFile.close();
// Trivial upper bound for the start times of the tasks
maxSumProcessingTimes = 0;
for [j in 0...nbJobs][o in 0...nbOperations[j]] {
local maxProcessingTime = 0;
for [m in 0...nbMachines] {
if (processingTime[j][o][m] == nil) {
local task = jobOperationTask[j][o];
taskProcessingTime[task][m] = INFINITE;
} else if (processingTime[j][o][m] >= maxProcessingTime) {
maxProcessingTime = processingTime[j][o][m];
}
}
maxSumProcessingTimes += maxProcessingTime;
}
maxStart = maxSumProcessingTimes + nbTasks * maxSetup;
}
/* Declare the optimization model */
function model() {
// Sequence of tasks on each machine
jobsOrder[m in 0...nbMachines] <- list(nbTasks);
// Each task is scheduled on a machine
constraint partition[m in 0...nbMachines](jobsOrder[m]);
// Only compatible machines can be selected for a task
for [i in 0...nbTasks][m in 0...nbMachines : taskProcessingTime[i][m] == INFINITE]
constraint !contains(jobsOrder[m], i);
// For each task, the selected machine
taskMachine[i in 0...nbTasks] <- find(jobsOrder, i);
// Interval decisions: time range of each task
tasks[i in 0...nbTasks] <- interval(0, maxStart);
// The task duration depends on the selected machine
duration[i in 0...nbTasks] <- taskProcessingTime[i][taskMachine[i]];
for [i in 0...nbTasks]
constraint length(tasks[i]) == duration[i];
// Precedence constraints between the operations of a job
for [j in 0...nbJobs][o in 0...nbOperations[j]-1] {
local i1 = jobOperationTask[j][o];
local i2 = jobOperationTask[j][o + 1];
constraint tasks[i1] < tasks[i2];
}
// Disjunctive resource constraints between the tasks on a machine
for [m in 0...nbMachines] {
local sequence <- jobsOrder[m];
constraint and(0...count(sequence)-1, i =>
start(tasks[sequence[i + 1]]) >= end(tasks[sequence[i]])
+ taskSetupTime[m][sequence[i]][sequence[i + 1]]);
}
// Minimize the makespan: end of the last task
makespan <- max[i in 0...nbTasks](end(tasks[i]));
minimize makespan;
}
/* Parameterize the solver */
function param() {
if (lsTimeLimit == nil) lsTimeLimit = 60;
}
/* Write the solution in a file with the following format:
* - for each operation of each job, the selected machine, the start and end dates */
function output() {
if (outFileName != nil) {
outFile = io.openWrite(outFileName);
println("Solution written in file ", outFileName);
for [j in 0...nbJobs][o in 0...nbOperations[j]] {
local taskIndex = jobOperationTask[j][o];
outFile.println(j + 1, "\t", o + 1, "\t", taskMachine[taskIndex].value + 1,
"\t", tasks[taskIndex].value.start, "\t", tasks[taskIndex].value.end);
}
}
}
- Execution (Windows)
- set PYTHONPATH=%LS_HOME%\bin\pythonpython flexible_jobshop_setup.py instances\Fattahi_setup_01.fjs
- Execution (Linux)
- export PYTHONPATH=/opt/localsolver_12_5/bin/pythonpython flexible_jobshop_setup.py instances/Fattahi_setup_01.fjs
import localsolver
import sys
# Constant for incompatible machines
INFINITE = 1000000
def read_instance(filename):
with open(filename) as f:
lines = f.readlines()
first_line = lines[0].split()
# Number of jobs
nb_jobs = int(first_line[0])
# Number of machines
nb_machines = int(first_line[1])
# Number of operations for each job
nb_operations = [int(lines[j + 1].split()[0]) for j in range(nb_jobs)]
# Number of tasks
nb_tasks = sum(nb_operations[j] for j in range(nb_jobs))
# Processing time for each task, for each machine
task_processing_time = [[INFINITE for m in range(nb_machines)] for i in range(nb_tasks)]
# For each job, for each operation, the corresponding task id
job_operation_task = [[0 for o in range(nb_operations[j])] for j in range(nb_jobs)]
# Setup time between every two consecutive tasks, for each machine
task_setup_time = [[[-1 for r in range(nb_tasks)] for i in range(nb_tasks)] for m in range(nb_machines)]
id = 0
for j in range(nb_jobs):
line = lines[j + 1].split()
tmp = 0
for o in range(nb_operations[j]):
nb_machines_operation = int(line[tmp + o + 1])
for i in range(nb_machines_operation):
machine = int(line[tmp + o + 2 * i + 2]) - 1
time = int(line[tmp + o + 2 * i + 3])
task_processing_time[id][machine] = time
job_operation_task[j][o] = id
id = id + 1
tmp = tmp + 2 * nb_machines_operation
id_line = nb_jobs + 2
max_setup = 0
for m in range(nb_machines):
for i1 in range(nb_tasks):
task_setup_time[m][i1] = list(map(int, lines[id_line].split()))
max_setup = max(max_setup, max(s if s != INFINITE else 0 for s in task_setup_time[m][i1]))
id_line += 1
# Trivial upper bound for the start times of the tasks
max_sum_processing_times = sum(
max(task_processing_time[i][m] for m in range(nb_machines) if task_processing_time[i][m] != INFINITE)
for i in range(nb_tasks))
max_start = max_sum_processing_times + nb_tasks * max_setup
return nb_jobs, nb_machines, nb_tasks, task_processing_time, job_operation_task, \
nb_operations, task_setup_time, max_start
def main(instance_file, output_file, time_limit):
nb_jobs, nb_machines, nb_tasks, task_processing_time_data, job_operation_task, \
nb_operations, task_setup_time_data, max_start = read_instance(instance_file)
with localsolver.LocalSolver() as ls:
#
# Declare the optimization model
#
model = ls.model
# Sequence of tasks on each machine
jobs_order = [model.list(nb_tasks) for _ in range(nb_machines)]
machines = model.array(jobs_order)
# Each task is scheduled on a machine
model.constraint(model.partition(machines))
# Only compatible machines can be selected for a task
for i in range(nb_tasks):
for m in range(nb_machines):
if task_processing_time_data[i][m] == INFINITE:
model.constraint(model.not_(model.contains(jobs_order[m], i)))
# For each task, the selected machine
task_machine = [model.find(machines, i) for i in range(nb_tasks)]
task_processing_time = model.array(task_processing_time_data)
task_setup_time = model.array(task_setup_time_data)
# Interval decisions: time range of each task
tasks = [model.interval(0, max_start) for _ in range(nb_tasks)]
# The task duration depends on the selected machine
duration = [model.at(task_processing_time, i, task_machine[i]) for i in range(nb_tasks)]
for i in range(nb_tasks):
model.constraint(model.length(tasks[i]) == duration[i])
task_array = model.array(tasks)
# Precedence constraints between the operations of a job
for j in range(nb_jobs):
for o in range(nb_operations[j] - 1):
i1 = job_operation_task[j][o]
i2 = job_operation_task[j][o + 1]
model.constraint(tasks[i1] < tasks[i2])
# Disjunctive resource constraints between the tasks on a machine
for m in range(nb_machines):
sequence = jobs_order[m]
sequence_lambda = model.lambda_function(
lambda i: model.start(task_array[sequence[i + 1]]) >= model.end(task_array[sequence[i]])
+ model.at(task_setup_time, m, sequence[i], sequence[i + 1]))
model.constraint(model.and_(model.range(0, model.count(sequence) - 1), sequence_lambda))
# Minimize the makespan: end of the last task
makespan = model.max([model.end(tasks[i]) for i in range(nb_tasks)])
model.minimize(makespan)
model.close()
# Parameterize the solver
ls.param.time_limit = time_limit
ls.solve()
# Write the solution in a file with the following format:
# - for each operation of each job, the selected machine, the start and end dates
if output_file != None:
with open(output_file, "w") as f:
print("Solution written in file", output_file)
for j in range(nb_jobs):
for o in range(0, nb_operations[j]):
taskIndex = job_operation_task[j][o]
f.write(str(j + 1) + "\t" + str(o + 1) + "\t" + str(task_machine[taskIndex].value + 1)
+ "\t" + str(tasks[taskIndex].value.start())
+ "\t" + str(tasks[taskIndex].value.end()) + "\n")
if __name__ == '__main__':
if len(sys.argv) < 2:
print("Usage: python flexible_jobshop_setup.py instance_file [output_file] [time_limit]")
sys.exit(1)
instance_file = sys.argv[1]
output_file = sys.argv[2] if len(sys.argv) >= 3 else None
time_limit = int(sys.argv[3]) if len(sys.argv) >= 4 else 60
main(instance_file, output_file, time_limit)
- Compilation / Execution (Windows)
- cl /EHsc flexible_jobshop_setup.cpp -I%LS_HOME%\include /link %LS_HOME%\bin\localsolver125.libflexible_jobshop_setup instances\Fattahi_setup_01.fjs
- Compilation / Execution (Linux)
- g++ flexible_jobshop_setup.cpp -I/opt/localsolver_12_5/include -llocalsolver125 -lpthread -o flexible_jobshop_setup./flexible_jobshop_setup instances/Fattahi_setup_01.fjs
#include "localsolver.h"
#include <algorithm>
#include <fstream>
#include <iostream>
#include <limits>
#include <numeric>
#include <vector>
using namespace localsolver;
class FlexibleJobshopSetup {
private:
// Number of jobs
int nbJobs;
// Number of machines
int nbMachines;
// Number of tasks
int nbTasks;
// Processing time for each task, for each machine
std::vector<std::vector<int>> taskProcessingTimeData;
// Setup time between every two consecutive tasks, for each machine
std::vector<std::vector<std::vector<int>>> taskSetupTimeData;
// For each job, for each operation, the corresponding task id
std::vector<std::vector<int>> jobOperationTask;
// Number of operations for each job
std::vector<int> nbOperations;
// Trivial upper bound for the start times of the tasks
int maxStart;
// Constant for incompatible machines
const int INFINITE = 1000000;
// LocalSolver
LocalSolver localsolver;
// Decision variables: time range of each task
std::vector<LSExpression> tasks;
// Decision variables: sequence of tasks on each machine
std::vector<LSExpression> jobsOrder;
// For each task, the selected machine
std::vector<LSExpression> taskMachine;
// Objective = minimize the makespan: end of the last task
LSExpression makespan;
public:
FlexibleJobshopSetup() : localsolver() {}
void readInstance(const std::string& fileName) {
std::ifstream infile;
infile.exceptions(std::ifstream::failbit | std::ifstream::badbit);
infile.open(fileName.c_str());
infile >> nbJobs;
infile >> nbMachines;
infile.ignore(std::numeric_limits<std::streamsize>::max(), '\n'); // skip last number
nbTasks = 0;
std::vector<std::vector<std::vector<int>>> processingTime = std::vector<std::vector<std::vector<int>>>(nbJobs);
jobOperationTask.resize(nbJobs);
nbOperations.resize(nbJobs);
for (unsigned int j = 0; j < nbJobs; ++j) {
infile >> nbOperations[j];
jobOperationTask[j].resize(nbOperations[j]);
processingTime[j].resize(nbOperations[j]);
for (unsigned int o = 0; o < nbOperations[j]; ++o) {
int nbMachinesOperation;
infile >> nbMachinesOperation;
taskProcessingTimeData.push_back(std::vector<int>(nbMachines, INFINITE));
processingTime[j][o].resize(nbMachines, INFINITE);
for (int m = 0; m < nbMachinesOperation; ++m) {
int machine;
int time;
infile >> machine;
infile >> time;
processingTime[j][o][machine - 1] = time;
taskProcessingTimeData[nbTasks][machine - 1] = time;
}
jobOperationTask[j][o] = nbTasks;
nbTasks += 1;
}
}
taskSetupTimeData = std::vector<std::vector<std::vector<int>>>(
nbMachines, std::vector<std::vector<int>>(nbTasks, std::vector<int>(nbTasks)));
int maxSetup = 0;
for (unsigned int m = 0; m < nbMachines; m++) {
for (unsigned int i = 0; i < nbTasks; i++) {
for (unsigned int j = 0; j < nbTasks; j++) {
infile >> taskSetupTimeData[m][i][j];
if (taskSetupTimeData[m][i][j] != INFINITE && taskSetupTimeData[m][i][j] > maxSetup)
maxSetup = taskSetupTimeData[m][i][j];
}
}
}
infile.close();
// Trivial upper bound for the start times of the tasks
int maxSumProcessingTimes = 0;
for (unsigned int j = 0; j < nbJobs; ++j) {
for (unsigned int o = 0; o < nbOperations[j]; ++o) {
int maxProcessingTime = 0;
for (unsigned int m = 0; m < nbMachines; ++m) {
if (processingTime[j][o][m] != INFINITE && processingTime[j][o][m] > maxProcessingTime)
maxProcessingTime = processingTime[j][o][m];
}
maxSumProcessingTimes += maxProcessingTime;
}
}
maxStart = maxSumProcessingTimes + nbTasks * maxSetup;
}
void solve(int timeLimit) {
// Declare the optimization model
LSModel model = localsolver.getModel();
// Sequence of tasks on each machine
jobsOrder.resize(nbMachines);
LSExpression machines = model.array();
for (unsigned int m = 0; m < nbMachines; ++m) {
jobsOrder[m] = model.listVar(nbTasks);
machines.addOperand(jobsOrder[m]);
}
// Each task is scheduled on a machine
model.constraint(model.partition(machines));
// Only compatible machines can be selected for a task
for (int i = 0; i < nbTasks; ++i) {
for (unsigned int m = 0; m < nbMachines; ++m) {
if (taskProcessingTimeData[i][m] == INFINITE) {
model.constraint(!model.contains(jobsOrder[m], i));
}
}
}
taskMachine.resize(nbTasks);
LSExpression taskProcessingTime = model.array();
for (int i = 0; i < nbTasks; ++i) {
// For each task, the selected machine
taskMachine[i] = model.find(machines, i);
taskProcessingTime.addOperand(
model.array(taskProcessingTimeData[i].begin(), taskProcessingTimeData[i].end()));
}
LSExpression taskSetupTime = model.array();
for (int m = 0; m < nbMachines; ++m) {
LSExpression setupTimeMachine = model.array();
for (int i = 0; i < nbTasks; ++i) {
setupTimeMachine.addOperand(
model.array(taskSetupTimeData[m][i].begin(), taskSetupTimeData[m][i].end()));
}
taskSetupTime.addOperand(setupTimeMachine);
}
tasks.resize(nbTasks);
std::vector<LSExpression> duration(nbTasks);
for (int i = 0; i < nbTasks; ++i) {
// Interval decisions: time range of each task
tasks[i] = model.intervalVar(0, maxStart);
// The task duration depends on the selected machine
duration[i] = model.at(taskProcessingTime, i, taskMachine[i]);
model.constraint(model.length(tasks[i]) == duration[i]);
}
LSExpression taskArray = model.array(tasks.begin(), tasks.end());
// Precedence constraints between the operations of a job
for (unsigned int j = 0; j < nbJobs; ++j) {
for (unsigned int o = 0; o < nbOperations[j] - 1; ++o) {
int i1 = jobOperationTask[j][o];
int i2 = jobOperationTask[j][o + 1];
model.constraint(taskArray[i1] < taskArray[i2]);
}
}
// Disjunctive resource constraints between the tasks on a machine
for (int m = 0; m < nbMachines; ++m) {
LSExpression sequence = jobsOrder[m];
LSExpression sequenceLambda = model.createLambdaFunction([&](LSExpression i) {
return model.start(taskArray[sequence[i + 1]]) >=
model.end(taskArray[sequence[i]]) + model.at(taskSetupTime, m, sequence[i], sequence[i + 1]);
});
model.constraint(model.and_(model.range(0, model.count(sequence) - 1), sequenceLambda));
}
// Minimize the makespan: end of the last task
makespan = model.max();
for (int i = 0; i < nbTasks; ++i) {
makespan.addOperand(model.end(tasks[i]));
}
model.minimize(makespan);
model.close();
// Parameterize the solver
localsolver.getParam().setTimeLimit(timeLimit);
localsolver.solve();
}
/* Write the solution in a file with the following format:
* - for each operation of each job, the selected machine, the start and end dates */
void writeSolution(const std::string& fileName) {
std::ofstream outfile(fileName.c_str());
if (!outfile.is_open()) {
std::cerr << "File " << fileName << " cannot be opened." << std::endl;
exit(1);
}
std::cout << "Solution written in file " << fileName << std::endl;
for (unsigned int j = 0; j < nbJobs; ++j) {
for (unsigned int o = 0; o < nbOperations[j]; ++o) {
int taskIndex = jobOperationTask[j][o];
outfile << j + 1 << "\t" << o + 1 << "\t" << taskMachine[taskIndex].getValue() + 1 << "\t"
<< tasks[taskIndex].getIntervalValue().getStart() << "\t"
<< tasks[taskIndex].getIntervalValue().getEnd() << std::endl;
}
}
outfile.close();
}
};
int main(int argc, char** argv) {
if (argc < 2) {
std::cout << "Usage: flexible_jobshop_setup instanceFile [outputFile] [timeLimit]" << std::endl;
exit(1);
}
const char* instanceFile = argv[1];
const char* outputFile = argc > 2 ? argv[2] : NULL;
const char* strTimeLimit = argc > 3 ? argv[3] : "60";
FlexibleJobshopSetup model;
try {
model.readInstance(instanceFile);
const int timeLimit = atoi(strTimeLimit);
model.solve(timeLimit);
if (outputFile != NULL)
model.writeSolution(outputFile);
return 0;
} catch (const std::exception& e) {
std::cerr << "An error occurred: " << e.what() << std::endl;
return 1;
}
}
- Compilation / Execution (Windows)
- copy %LS_HOME%\bin\localsolvernet.dll .csc FlexibleJobshopSetup.cs /reference:localsolvernet.dllFlexibleJobshopSetup instances\Fattahi_setup_01.fjs
using System;
using System.IO;
using System.Linq;
using localsolver;
public class FlexibleJobshopSetup : IDisposable
{
// Number of jobs
private int nbJobs;
// Number of machines
private int nbMachines;
// Number of tasks
private int nbTasks;
// Processing time for each task, for each machine
private long[][] taskProcessingTimeData;
// Setup time between every two consecutive tasks, for each machine
private int[][][] taskSetupTimeData;
// For each job, for each operation, the corresponding task id
private int[][] jobOperationTask;
// Number of operations for each job;
private int[] nbOperations;
// Trivial upper bound for the start times of the tasks
private long maxStart;
// Constant for incompatible machines
private const long INFINITE = 1000000;
// LocalSolver
private LocalSolver localsolver;
// Decision variables: time range of each task
private LSExpression[] tasks;
// Decision variables: sequence of tasks on each machine
private LSExpression[] jobsOrder;
// For each task, the selected machine
private LSExpression[] taskMachine;
// Objective = minimize the makespan: end of the last task
private LSExpression makespan;
public FlexibleJobshopSetup()
{
localsolver = new LocalSolver();
}
public void ReadInstance(string fileName)
{
using (StreamReader input = new StreamReader(fileName))
{
char[] separators = new char[] { '\t', ' ' };
string[] splitted = input
.ReadLine()
.Split(separators, StringSplitOptions.RemoveEmptyEntries);
nbJobs = int.Parse(splitted[0]);
nbMachines = int.Parse(splitted[1]);
nbTasks = 0;
long[][][] processingTime = new long[nbJobs][][];
jobOperationTask = new int[nbJobs][];
nbOperations = new int[nbJobs];
for (int j = 0; j < nbJobs; ++j)
{
splitted = input
.ReadLine()
.Split(separators, StringSplitOptions.RemoveEmptyEntries);
nbOperations[j] = int.Parse(splitted[0]);
jobOperationTask[j] = new int[nbOperations[j]];
processingTime[j] = new long[nbOperations[j]][];
int k = 1;
for (int o = 0; o < nbOperations[j]; ++o)
{
int nbMachinesOperation = int.Parse(splitted[k]);
k++;
processingTime[j][o] = Enumerable.Repeat((long)INFINITE, nbMachines).ToArray();
for (int m = 0; m < nbMachinesOperation; ++m)
{
int machine = int.Parse(splitted[k]) - 1;
long time = long.Parse(splitted[k + 1]);
processingTime[j][o][machine] = time;
k += 2;
}
jobOperationTask[j][o] = nbTasks;
nbTasks++;
}
}
input.ReadLine();
taskSetupTimeData = new int[nbMachines][][];
int maxSetup = 0;
for (int m = 0; m < nbMachines; ++m)
{
taskSetupTimeData[m] = new int[nbTasks][];
for (int i = 0; i < nbTasks; ++i)
{
taskSetupTimeData[m][i] = new int[nbTasks];
splitted = input.
ReadLine()
.Split(separators, StringSplitOptions.RemoveEmptyEntries);
for (int j = 0; j < nbTasks; ++j)
{
taskSetupTimeData[m][i][j] = int.Parse(splitted[j]);
if (
taskSetupTimeData[m][i][j] != INFINITE
&& taskSetupTimeData[m][i][j] > maxSetup
)
maxSetup = taskSetupTimeData[m][i][j];
}
}
}
// Trivial upper bound for the start times of the tasks
long maxSumProcessingTimes = 0;
taskProcessingTimeData = new long[nbTasks][];
for (int j = 0; j < nbJobs; ++j)
{
long maxProcessingTime = 0;
for (int o = 0; o < nbOperations[j]; ++o)
{
int task = jobOperationTask[j][o];
taskProcessingTimeData[task] = new long[nbMachines];
for (int m = 0; m < nbMachines; ++m)
{
taskProcessingTimeData[task][m] = processingTime[j][o][m];
if (
processingTime[j][o][m] != INFINITE
&& processingTime[j][o][m] > maxProcessingTime
)
{
maxProcessingTime = processingTime[j][o][m];
}
}
maxSumProcessingTimes += maxProcessingTime;
}
}
maxStart = maxSumProcessingTimes + nbTasks * maxSetup;
}
}
public void Dispose()
{
localsolver.Dispose();
}
public void Solve(int timeLimit)
{
// Declare the optimization model
LSModel model = localsolver.GetModel();
// Sequence of tasks on each machine
jobsOrder = new LSExpression[nbMachines];
LSExpression machines = model.Array();
for (int m = 0; m < nbMachines; ++m)
{
jobsOrder[m] = model.List(nbTasks);
machines.AddOperand(jobsOrder[m]);
}
// Each task is scheduled on a machine
model.Constraint(model.Partition(machines));
// Only compatible machines can be selected for a task
for (int i = 0; i < nbTasks; ++i)
{
for (int m = 0; m < nbMachines; ++m)
{
if (taskProcessingTimeData[i][m] == INFINITE)
model.Constraint(!model.Contains(jobsOrder[m], i));
}
}
// For each task, the selected machine
taskMachine = new LSExpression[nbTasks];
for (int i = 0; i < nbTasks; ++i)
{
taskMachine[i] = model.Find(machines, i);
}
tasks = new LSExpression[nbTasks];
LSExpression[] duration = new LSExpression[nbTasks];
LSExpression taskProcessingTime = model.Array(taskProcessingTimeData);
for (int i = 0; i < nbTasks; ++i)
{
// Interval decisions: time range of each task
tasks[i] = model.Interval(0, maxStart);
// The task duration depends on the selected machine
LSExpression iExpr = model.CreateConstant(i);
duration[i] = model.At(taskProcessingTime, iExpr, taskMachine[i]);
model.Constraint(model.Length(tasks[i]) == duration[i]);
}
LSExpression taskArray = model.Array(tasks);
// Precedence constraints between the operations of a job
for (int j = 0; j < nbJobs; ++j)
{
for (int o = 0; o < nbOperations[j] - 1; ++o)
{
int i1 = jobOperationTask[j][o];
int i2 = jobOperationTask[j][o + 1];
model.Constraint(tasks[i1] < tasks[i2]);
}
}
LSExpression taskSetupTime = model.Array(taskSetupTimeData);
// Disjunctive resource constraints between the tasks on a machine
for (int m = 0; m < nbMachines; ++m)
{
LSExpression sequence = jobsOrder[m];
LSExpression mexpr = model.CreateConstant(m);
LSExpression sequenceLambda = model.LambdaFunction(
i =>
model.Start(taskArray[sequence[i + 1]])
>= (model.End(taskArray[sequence[i]])
+ model.At(taskSetupTime, mexpr, sequence[i], sequence[i + 1])));
model.Constraint(model.And(model.Range(0, model.Count(sequence) - 1), sequenceLambda));
}
// Minimize the makespan: end of the last task
makespan = model.Max();
for (int i = 0; i < nbTasks; ++i)
{
makespan.AddOperand(model.End(tasks[i]));
}
model.Minimize(makespan);
model.Close();
// Parameterize the solver
localsolver.GetParam().SetTimeLimit(timeLimit);
localsolver.Solve();
}
/* Write the solution in a file with the following format:
* - for each operation of each job, the selected machine, the start and end dates */
public void WriteSolution(string fileName)
{
using (StreamWriter output = new StreamWriter(fileName))
{
Console.WriteLine("Solution written in file " + fileName);
for (int j = 1; j <= nbJobs; ++j)
{
for (int o = 1; o <= nbOperations[j - 1]; ++o)
{
int taskIndex = jobOperationTask[j - 1][o - 1];
output.WriteLine(
j
+ "\t"
+ o
+ "\t"
+ taskMachine[taskIndex].GetValue()
+ "\t"
+ tasks[taskIndex].GetIntervalValue().Start()
+ "\t"
+ tasks[taskIndex].GetIntervalValue().End()
);
}
}
}
}
public static void Main(string[] args)
{
if (args.Length < 1)
{
Console.WriteLine("Usage: FlexibleJobshopSetup instanceFile [outputFile] [timeLimit]");
System.Environment.Exit(1);
}
string instanceFile = args[0];
string outputFile = args.Length > 1 ? args[1] : null;
string strTimeLimit = args.Length > 2 ? args[2] : "60";
using (FlexibleJobshopSetup model = new FlexibleJobshopSetup())
{
model.ReadInstance(instanceFile);
model.Solve(int.Parse(strTimeLimit));
if (outputFile != null)
model.WriteSolution(outputFile);
}
}
}
- Compilation / Execution (Windows)
- javac FlexibleJobshopSetup.java -cp %LS_HOME%\bin\localsolver.jarjava -cp %LS_HOME%\bin\localsolver.jar;. FlexibleJobshopSetup instances\Fattahi_setup_01.fjs
- Compilation / Execution (Linux)
- javac FlexibleJobshopSetup.java -cp /opt/localsolver_12_5/bin/localsolver.jarjava -cp /opt/localsolver_12_5/bin/localsolver.jar:. FlexibleJobshopSetup instances/Fattahi_setup_01.fjs
import java.util.*;
import java.io.*;
import localsolver.*;
public class FlexibleJobshopSetup {
// Number of jobs
private int nbJobs;
// Number of machines
private int nbMachines;
// Number of tasks
private int nbTasks;
// Processing time for each task, for each machine
private long[][] taskProcessingTimeData;
// Setup time between every two consecutive tasks, for each machine
private int[][][] taskSetupTimeData;
// For each job, for each operation, the corresponding task id
private int[][] jobOperationTask;
// Number of operations for each job;
private int[] nbOperations;
// Trivial upper bound for the start times of the tasks
private long maxStart;
// Constant for incompatible machines
private final int INFINITE = 1000000;
// LocalSolver
final LocalSolver localsolver;
// Decision variables: time range of each task
private LSExpression[] tasks;
// Decision variables: sequence of tasks on each machine
private LSExpression[] jobsOrder;
// For each task, the selected machine
private LSExpression[] taskMachine;
// Objective = minimize the makespan: end of the last task
private LSExpression makespan;
public FlexibleJobshopSetup(LocalSolver localsolver) throws IOException {
this.localsolver = localsolver;
}
public void readInstance(String fileName) throws IOException {
try (Scanner input = new Scanner(new File(fileName))) {
nbJobs = input.nextInt();
nbMachines = input.nextInt();
input.next(); // skip last number
nbTasks = 0;
long[][][] processingTime = new long[nbJobs][][];
jobOperationTask = new int[nbJobs][];
nbOperations = new int[nbJobs];
for (int j = 0; j < nbJobs; ++j) {
nbOperations[j] = input.nextInt();
jobOperationTask[j] = new int[nbOperations[j]];
processingTime[j] = new long[nbOperations[j]][nbMachines];
for (int o = 0; o < nbOperations[j]; ++o) {
int nbMachinesOperation = input.nextInt();
Arrays.fill(processingTime[j][o], INFINITE);
for (int m = 0; m < nbMachinesOperation; ++m) {
int machine = input.nextInt() - 1;
long time = input.nextLong();
processingTime[j][o][machine] = time;
}
jobOperationTask[j][o] = nbTasks;
nbTasks++;
}
}
taskSetupTimeData = new int[nbMachines][nbTasks][nbTasks];
int maxSetup = 0;
for (int m = 0; m < nbMachines; ++m) {
for (int i = 0; i < nbTasks; ++i) {
for (int j = 0; j < nbTasks; ++j) {
taskSetupTimeData[m][i][j] = input.nextInt();
if (taskSetupTimeData[m][i][j] != INFINITE && taskSetupTimeData[m][i][j] > maxSetup)
maxSetup = taskSetupTimeData[m][i][j];
}
}
}
// Trivial upper bound for the start times of the tasks
long maxSumProcessingTimes = 0;
taskProcessingTimeData = new long[nbTasks][];
for (int j = 0; j < nbJobs; ++j) {
long maxProcessingTime = 0;
for (int o = 0; o < nbOperations[j]; ++o) {
int task = jobOperationTask[j][o];
taskProcessingTimeData[task] = new long[nbMachines];
for (int m = 0; m < nbMachines; ++m) {
taskProcessingTimeData[task][m] = processingTime[j][o][m];
if (processingTime[j][o][m] != INFINITE && processingTime[j][o][m] > maxProcessingTime) {
maxProcessingTime = processingTime[j][o][m];
}
}
maxSumProcessingTimes += maxProcessingTime;
}
}
maxStart = maxSumProcessingTimes + nbTasks * maxSetup;
}
}
public void solve(int timeLimit) {
// Declare the optimization model
LSModel model = localsolver.getModel();
// Sequence of tasks on each machine
jobsOrder = new LSExpression[nbMachines];
LSExpression machines = model.array();
for (int m = 0; m < nbMachines; ++m) {
jobsOrder[m] = model.listVar(nbTasks);
machines.addOperand(jobsOrder[m]);
}
// Each task is scheduled on a machine
model.constraint(model.partition(machines));
// Only compatible machines can be selected for a task
for (int i = 0; i < nbTasks; ++i) {
for (int m = 0; m < nbMachines; ++m) {
if (taskProcessingTimeData[i][m] == INFINITE) {
model.constraint(model.not(model.contains(jobsOrder[m], i)));
}
}
}
// For each task, the selected machine
taskMachine = new LSExpression[nbTasks];
for (int i = 0; i < nbTasks; ++i) {
taskMachine[i] = model.find(machines, i);
}
LSExpression taskProcessingTime = model.array(taskProcessingTimeData);
tasks = new LSExpression[nbTasks];
LSExpression[] duration = new LSExpression[nbTasks];
for (int i = 0; i < nbTasks; ++i) {
// Interval decisions: time range of each task
tasks[i] = model.intervalVar(0, maxStart);
// The task duration depends on the selected machine
LSExpression iExpr = model.createConstant(i);
duration[i] = model.at(taskProcessingTime, iExpr, taskMachine[i]);
model.constraint(model.eq(model.length(tasks[i]), duration[i]));
}
LSExpression taskArray = model.array(tasks);
// Precedence constraints between the operations of a job
for (int j = 0; j < nbJobs; ++j) {
for (int o = 0; o < nbOperations[j] - 1; ++o) {
int i1 = jobOperationTask[j][o];
int i2 = jobOperationTask[j][o + 1];
model.constraint(model.lt(tasks[i1], tasks[i2]));
}
}
LSExpression taskSetupTime = model.array(taskSetupTimeData);
// Disjunctive resource constraints between the tasks on a machine
for (int m = 0; m < nbMachines; ++m) {
LSExpression sequence = jobsOrder[m];
LSExpression mExpr = model.createConstant(m);
LSExpression sequenceLambda = model
.lambdaFunction(
i -> model.geq(model.start(model.at(taskArray, model.at(sequence, model.sum(i, 1)))),
model.sum(model.end(model.at(taskArray, model.at(sequence, i))),
model.at(taskSetupTime, mExpr,
model.at(sequence, i), model.at(sequence, model.sum(i, 1))))));
model.constraint(model.and(model.range(0, model.sub(model.count(sequence), 1)), sequenceLambda));
}
// Minimize the makespan: end of the last task
makespan = model.max();
for (int i = 0; i < nbTasks; ++i) {
makespan.addOperand(model.end(tasks[i]));
}
model.minimize(makespan);
model.close();
// Parameterize the solver
localsolver.getParam().setTimeLimit(timeLimit);
localsolver.solve();
}
/*
* Write the solution in a file with the following format:
* - for each operation of each job, the selected machine, the start and end
* dates
*/
public void writeSolution(String fileName) throws IOException {
try (PrintWriter output = new PrintWriter(fileName)) {
System.out.println("Solution written in file " + fileName);
for (int j = 1; j <= nbJobs; ++j) {
for (int o = 1; o <= nbOperations[j - 1]; ++o) {
int taskIndex = jobOperationTask[j - 1][o - 1];
output.write(j + "\t" + o + "\t" + taskMachine[taskIndex].getValue() + "\t"
+ tasks[taskIndex].getIntervalValue().getStart()
+ "\t" + tasks[taskIndex].getIntervalValue().getEnd() + "\n");
}
}
}
}
public static void main(String[] args) {
if (args.length < 1) {
System.out.println("Usage: java FlexibleJobshopSetup instanceFile [outputFile] [timeLimit]");
System.exit(1);
}
String instanceFile = args[0];
String outputFile = args.length > 1 ? args[1] : null;
String strTimeLimit = args.length > 2 ? args[2] : "60";
try (LocalSolver localsolver = new LocalSolver()) {
FlexibleJobshopSetup model = new FlexibleJobshopSetup(localsolver);
model.readInstance(instanceFile);
model.solve(Integer.parseInt(strTimeLimit));
if (outputFile != null) {
model.writeSolution(outputFile);
}
} catch (Exception ex) {
System.err.println(ex);
ex.printStackTrace();
System.exit(1);
}
}
}