We tried to put these online materials along with the manuscript in the submission system, but alas, this is almost 800 MBytes compressed with 7zip. Get it here. There is a PDF inside with all explanations.

Benchmark of instances for the permutation flowshop scheduling problem. There are 240 small instances and 240 large instances. All sets, together with the best known solutions and lower bounds are given as a single file here. You will need the freeware compressor 7zip to open the file. 

These instances have been used in the paper:

Vallada, Eva, Ruiz, Rubén & Framinan, Jose M. (2017). New hard benchmark for flowshop scheduling problems minimising makespan. European Journal of Operational Research, 240, 666-677.

The complete solutions (permutations) are available here.

The online materials for a work in which we study an unrelated parallel machine problem with an additiona resource are provided in a single zip file here.

There is lots of information here:

• readme.txt file with quick instructions
• a Howto pdf file with detailed instructions
• All instances used in the paper
• A binary to generate those instances or other instances if wished
• A binary to replicate all results of all proposed algorithms and models
• Excel files with all detailed summarized results given in the paper
• Detailed complete solution files
• Statistics files used in the files for the analyses
•  

These instances consider flowshops with m stages and several parallel (identical) machines per stage. Instead of due dates for jobs we have due date windows. Each due date is different for each job. There are weights for earliness that are different from the weights of tardiness.

There are 4 sets of instances. Small, large, calibration small and calibration large. All sets, together with the best solutions known and best total weighted earliness and tardiness values are given a a single file here. There is also a complete explanation of the instances and files inside. Look for "readme.txt". You will need the freeware compressor 7zip to open the file. 

These instances have been used in the paper:

Quan-Ke, Pan, Ruiz, Rubén & Alfaro-Fernández, Pedro. (2017). Iterated search methods for earliness and tardiness minimization in hybrid flowshops with due windows. Computers & Operations Research, 80, 50-60.

The complete solutions for all tested algorithms and methods are available upon request from the authors.

These instances extend the distributed permutation flowshop problem by adding an additional assembly stage at the end and sequence dependent setup times

There are two sets calibration and test instances. The calibration set comprises 60 instances that are available here (7z compressed file, use 7Zip freeware program to open it). The best solutions known for these instances are here.

Then there is a much larger set of 540 test instances. They are here and the best solutions here.

Instances used in Rodríguez A., Ruiz R., "A study of the effect of the depot location in realistic and asymmetric vehicle routing problems", can be downloaded here:

• case ACVRP - depot (30 MB)
• 540 location files, for different number of nodes on several territories and depot locations.
• 540 distance matrix files (real asymmetric).
• 360 demand and max.capacity files.
•  
• » explanation of the ACVRP - depot  files format here.
•  

Note: The files were tested 100% ok in their extraction with several software. You can to extract the files with: 7-Zip (Windows), IZArc (Windows), Unarchiver (Mac OS X) or other software compatible with 7z compress format. For more information, please contact.

Related papers and publications.

Rodríguez, Alejandro & Ruiz, Rubén. (2012). A study on the effect of the asymmetry on real capacitated vehicle routing problems. Computers & Operations Research, 39(9), 2142–2151

The sets of small and large instances, along with best known solutions for each case are available here (7z compressed file, use 7Zip freeware program to open it).

The sets of small and large instances, along with best known solutions for each case are available here (7z compressed file, use 7Zip freeware program to open it).

These instances extend the pure no-idle instances by considering both idle as well as no-idle machines.

The basic benchmark contains 250 instances. All combinations of the following n and m values are used: n = {50, 100, 150, 200, 250, 300, 350, 400, 450, 500} and m = {10, 20, 30, 40, 50}. For each one of the 10 × 5 = 50 combinations, five replicates are obtained which results in 250 instances. 

Furthermore, for each one of these 250 instances we create 7 different sets:

Group 1: The first 50% of the machines have the no-idle constraint. The remaining 50% are regular idle machines.

Group 2: The second 50% of the machines have the no-idle constraint.

Group 3: The machines alternate, in order, between regular and no-idle constraints.

Group 4: A random 25% of the machines are no-idle. Group 5: 50% random no-idle machines.

Group 6: 75% random no-idle machines.

Group 7: This group contains the 250 original no-idle instances of Ruiz et al. (2009), i.e., in this group all machines have the no-idle constraint.

In total we have 1750 instances which are available here (7z compressed file, use 7Zip freeware program to open it). There is also a set of 100 random calibration instances here.  The best known solutions for the test benchmark are here. Note that at the end of each instance there is a vector with a boolean for each machine (NO-IDLE), a value of 0 means a regular idle machine and a value of 1 means a no-idle machine.

These instances extend the distributed permutation flowshop problem by adding an additional assembly stage at the end

There is a set of 900 small instances and a set of 810 larger instances. All instances are available here (7z compressed file, use 7Zip freeware program to open it). The best known solutions are here.

The base instances as regards jobs and machines are based on "instances for unrelated parallel machines problems and makespan criterion" here we only give the best known solutions in an excel file which is available here (Excel 2007 format).

We used these instances in the following paper:

• L. Fanjul, and R. Ruiz, (2012). Scheduling unrelated parallel machines with optional machines and jobs selection. Computers & Operations Research,39(7), 1745-1753

Instances used in the Ph.D. thesis of Alejandro Rodríguez Villalobos, named "Study of the effect of asymmetry in vehicle routing problems", can be downloaded here:

•  
• 
• case TSP (606 MB)
• 450 location files, for different number of nodes on several territories.
• 2250 distance matrix files (orthodromic, real asymmetric and others).
•  
• » explanation of the TSP files format here.
•  
•  

• case CVRP (712 MB)
• 1350 location files, for different number of nodes on several territories and depot locations.
• 2700 distance matrix files (orthodromic and real asymmetric).
• 900 demand and max.capacity files.
•  
• » explanation of the CVRP files format here.
•  

Note: The files were tested 100% ok in their extraction with several software. You can to extract the files with: 7-Zip (Windows), IZArc (Windows), Unarchiver (Mac OS X) or other software compatible with 7z compress format. For more information, please contact.

Related papers and publications.

Rodríguez, Alejandro & Ruiz, Rubén (2010). Estudio de la relación entre la asimetría de las redes de transporte por carretera, el territorio y la localización. En  de  Basurto  Uraga, Pablo  Díaz,  Carrasco, Ernesto  Cilleruelo,  Laburu, Carlos  Ochoa,  de  Arbulo  López, Patxi  Ruiz,  Bilbao, Enara  Zarrabeitia,  Meaza, Izaskun  Alvarez et al (editores), 4th   International   Conference   on   Industrial   Engineering   and   Industrial  Management.  XIV  Congreso  Ingeniería  de  Organización, páginas 1100-1109. Alameda  Urquijo  s/n, 48013  Bilbao   : Departamento  de  Organización  de  Empresas, Universidad  . [Más] [Versión digital]

Katragjini, Kostanca, Ruiz, Rubén & Rodríguez, Alejandro (2010). Using real world distances in logistics management. En de Basurto Uraga, Pablo Díaz, Carrasco, Ernesto Cilleruelo, Laburu, Carlos Ochoa, de Arbulo López, Patxi Ruiz, Bilbao, Enara Zarrabeitia, Meaza, Izaskun Alvarez et al (editores), 4th International Conference on Industrial Engineering and Industrial Management. XIV Congreso Ingeniería de Organización, páginas 1166-1176. Alameda Urquijo s/n, 48013 Bilbao : Departamento de Organización de Empresas, Universidad. [Más] [Versión digital] 

Rodríguez, Alejandro & Ruiz, Rubén. (2012a). A study on the effect of the asymmetry on real capacitated vehicle routing problems. Computers & Operations Research, 39(9), 2142–2151. [Más] [Versión digital]

Rodríguez, Alejandro & Ruiz, Rubén. (2012b). The effect of the asymmetry of road transportation networks on the Traveling Salesman Problem. Computers & Operations Research, 39(7), 1566-1576. [Más] [Versión digital]

•  
•  Instances for the multi-objective permutation flowshop problem with sequence dependent setups ara available here.
•  
• The results of the "Design-of-Experiment" (DoE) phase can be downloaded here.
•  
• Te results of the experiments for makespan-total flowtime and makespan total weighted tardines are available here.
•  
• Here there is a collection of images representing empirical attainment functions (EAFs) and differential empirical attainment functions (Diff-EAFs) for both (makespan - total flowtime) and (makespan - total weighted tardiness) objective pairs
•  
• Finally here can be found a 47 pages PDF file containing all the tables and figures for the ANOVA as well as Friedman rank-based experiments.  
•  
• Instances used in this paper:
•  
• M. Ciavotta, G. Minella and R. Ruiz, (2013). Multi-objective sequence dependent setup times flowshop scheduling: a new algorithm and a comprehensive study. European Journal of Operational Research, 227(2), 301-313

Large and small instances, both for calibration and for final tests, can be downloaded here.

The instances are used in the Ph.D. thesis of Thijs Urlings, named "Heuristics and metaheuristics for heavily constrained hybrid flowshop problems".

The complete set of events along with the initial baselines can be downloaded from here.

The instances have been used in this paper:

K. Katragjini, E. Vallada, Eva and R. Ruiz, (2013). Flowshop rescheduling under different types of disruption. International Journal of Production Research, 50(1), 780-797.

These instances extend the regular flowshop instances with multiple factories

There is a set of 420 small instances and a set of 720 larger instances. There are also two small sets (small and large) of test instances for calibration. All instances are here (7z compressed file, use 7Zip freeware program to open it).

These instances have been used in this paper:

 Naderi, B. & Ruiz, Rubén, (2010). The distributed permutation flowshop scheduling problem. Computers & Operations Research, 37(4), 754-768.

The best solutions for this initial paper are here.

Later, the instances were used in this paper

Naderi, B. & Ruiz, Rubén, (2014). A scatter search algorithm for the distributed permutation flowshop scheduling problem. European Journal of Operational Research, 239(2), 323-334.

New best solutions here.

Recently, we have used the instances again:

Ruiz, Rubén, Pan, Q.-K. & Naderi, B. (2018). Iterated Greedy methods for the distributed permutation flowshop scheduling problem. OMEGA, The International Journal of Management Science. In press.

See the latest best solutions here.

The instances are divided into 7 groups of 200 instances each, totaling 1,400 instances.

All instances with the best solutions known are here (large 47Mb file, use 7Zip freeware program to open it). The best known solutions have been obtained with CPLEX 11.1 with 2 hours running time on a Core 2 Duo 2.4 GHz computer.

The instances have been used in these papers:

L. Fanjul, and R. Ruiz, (2010). Iterated greedy local search methods for unrelated parallel machine scheduling. European Journal of Operational Research, 207(1), 55-69.

L. Fanjul, and R. Ruiz, (2011). Size-reduction heuristics for the unrelated parallel machines scheduling problem. Computers & Operations Research, 38(1), 301-309.

L. Fanjul, and R. Ruiz, (2012). Scheduling unrelated parallel machines with optional machines and jobs selection. Computers & Operations Research,39(7), 1745-1753.

Much later, in 2021, some colleagues have presented excellent results using an exact solver in the paper:

E. Åblad, A.-B. Strömberg, D. Spensieri. (2021) Exact makespan minimization of unrelated parallel machines OJMO, 2, 15 p,  doi : 10.5802/ojmo.4

They have been using the aforementioned instances, resulting in many incumbent (often optimal) solutions. The new results and the solver source code is found here, including a summary on the results, a compressed file with all results, and all scripts that produced the results.

These problems are hybrid flowshops with identical parallel machines per stage where job stage skipping is allowed. Furthermore, stage and job sequence dependent setup times are present.

You can find all instances here (20,3 Mbyte file, use 7Zip freeware program to open it).

We used these instances in the following paper:

• B. Naderi, R. Ruiz and M. Zandieh. (2008), Algorithms for a realistic variant of flowshop scheduling,  Computers & Operations Research, 37(2), 236-246.

Full set of instances (including test instances) here (7z compressed file, use 7Zip freeware program to open it). The best known solutions are here.

These instances have been used in the following book chapter

• R. Ruiz, E. Vallada and Fernández-Martínez, C. (2009). Scheduling in flowshops with no-idle machines. Scheduling in flowshops with no-idle machines, chapter 2, pages 21-52 of the book "Computational Intelligence in Flow Shop and Job Shop Scheduling" edited by Uday K. Chakraborty. New York : Springer.

Large and small instances here (HUGE 425Mb file!, use 7Zip freeware program to open it). Test instances for calibration experiments are here. The best known solutions so far are here.

We used these instances in the following paper:

• E. Vallada, R. Ruiz (2011). Genetic algorithms for the unrelated parallel machine scheduling problem with sequence dependent setup times. European Journal of Operational Research, 211(3), 612-622.

Later, the same problem was studied in this paper:

   L. Fanjul, R. Ruiz & F. Perea. (2019). Reformulations and an exact algorithm for unrelated parallel machine scheduling problems with setup times. Computers and Operations Research, 101, 173-182.

This last paper uses three sets of instances of three different sizes, small, medium and large. Large instances go all the way up to 1000 jobs and 8 machines. These instances are so large in total (more than 8 Gigabytes) that we are providing here a 7zip file of almost 1 Terabyte with all of them.

Large and small instances, along with best solutions here (large 78 Mb file!, use 7Zip freeware program to open it).

 We have used these instances in the following paper:

• R. Ruiz and C. Andrés, (2011). Scheduling unrelated parallel machines with resource-assignable sequence dependent setup times. International Journal of Advanced Manufacturing Technology, 57(5-8), 777-794

The instances (weighted makespan and maximum lateness, with different weights) can be downloaded here.

We provide also the best solutions for those instances that were actually used in the paper (U[1,00] and m≥3) here.

 We used these instances in the following paper:

• R. Ruiz and A. Allahverdi, (2009). New heuristics for no-wait flowshops with bicriteria of makespan and maximum lateness. International Journal of Production Research, 47(20), 5717-5738

These problems contain many constraints.

The complete set of instances is incredibly large. This set contains no less than 9,216 small instances and 3,072 large instances. It was used in the following paper:

• R. Ruiz, F. Sivrikaya Şerifoğlu and T. Urlings, (2008). Modeling realistic hybrid flexible flowshop scheduling problems. Computers & Operations Research 35, 1151-1175.

 If you want the complete set of instances (about 1.52 Gigabytes!!) open up an FTP account or some shared storage and drop us an email. Alternatively, we can send you a DVD disc if you provide us with a full mailing address.

A subset of these instances (576 small instances and 192 large instances) is here (12,1 Mbyte file, use 7Zip freeware program to open it). A list of best known solution values for the small and the large instancesis available as well.

We used these latter instances in the following papers:

• T. Urlings, R. Ruiz and T. Stützle, (2010), Shifting representation search for hybrid flexible flowline problems, European Journal of Operational Research 207, 1086-1095.
  

• T. Urlings, R. Ruiz and F. Sivrikaya Şerifoğlu, (2010), Genetic algorithms with different representation schemes for complex hybrid flexible flow line problems, International Journal of Metaheuristics 1, 30-54.

You can download the instances here (use 7Zip freeware program to open it).

More or less, these instances are those of Taillard's with the addition of due dates. 

We have compiled the best known Pareto sets for each instance here. These are given for three different CPU time stopping criteria and also for each one of the three possible criteria combinations (makespan and total tardiness, makespan and total flowtime and total flowtime and total tardiness).

 We have used these instances in the following papers:

• G. Minella, R. Ruiz and M. Ciavotta (2008). A Review and Evaluation of Multi-Objective Algorithms for the Flowshop Scheduling Problem. INFORMS Journal on Computing, 20(3): 451-471.
• G. Minella, R. Ruiz and M. Ciavotta (2011). Restarted Iterated Pareto Greedy algorithm for Multi-Objective Flowshop Scheduling problems.Computers & Operations Research, 38(11), 1521-1533

You can download the instances here (28.5 Mb file!, use 7Zip freeware program to open it).

Basically, we use extended the instances described above for the flowshop with setup times with due dates, so there are also 4 groups of instances with 120 problems each (480 total). 

 We have used these instances in the following paper:

The complete benchmark (540 instances) along with the best solutions can be downloaded here.

We used these instances in the following papers:

• E. Vallada, R. Ruiz and G. Minella, (2008). Minimizing total tardiness in the m-machine flowshop problem: a review and evaluation of heuristics and metaheuristics. Computers & Operations Research 35, 1350-1373.
• E. Vallada and R. Ruiz, (2009). Cooperative Metaheuristics for the Permutation Flowshop Scheduling Problem.  European Journal of Operational Research, 193(2): 365-376.
• E. Vallada and R. Ruiz, (2010).Genetic algorithms with path relinking for the minimum tardiness permutation flowshop problem. OMEGA, The International Journal of Management Science, 38(1-2), 57-67.

The best known solutions (maximum lateness criterion) can be obtained from this small excel file.

We used these instances in the following paper:

There is a set of 3,000 small instances available here.

The set of 2,400 large instances can be downloaded from here (19.5 Mbyte file). 

The best known solutions (total completion time criterion) for the 5,400 instances can be obtained from this small excel file.

 We used these instances in the following paper:

The sets of instances, along with the best known solutions are available here (22 Mb file!, use 7Zip freeware program to open it)

There are two sets of instances, with 3,600 instances each (7,200 total).  

 We used these instances in the following paper: 

You can download the instances here (28.6 Mb file!, use 7Zip freeware program to open it).

Basically, we use the well known code of Taillard to "extend" Taillard's intances with setup times. 

There are four sets of instances with 120 instances each (480 instances in total):

 SET 1 (SDST10): Setup times are uniformly distributed between 1 and 9 (10% of the processing times)

 SET 2 (SDST50): Setup times are uniformly distributed between 1 and 49 (50% of the processing times)

 SET 3 (SDST100): Setup times are uniformly distributed between 1 and 99 (100% of the processing times)

 SET 4 (SDST125): Setup times are uniformly distributed between 1 and 124 (125% of the processing times)

The best known solutions for the instances are here (small .xls file)

 We have used these instances in the following papers:

•  R. Ruiz, C. Maroto and J. Alcaraz (2005). Solving the flowshop scheduling problem with sequence dependent setup times using advanced metaheuristics. European Journal of Operational Research 165, 34-54.
• R. Ruiz and T. Stützle, (2008) An Iterated Greedy heuristic for the sequence dependent setup times flowshop with makespan and weighted tardiness objectives. European Journal of Operational Research 187, 1143-1159.