This repository contains supplementary data used in INF05010 course for the 2019/1. A presentation of problems and their mathematical formulation can be found here (in portuguese). A more complete specification of the project is available here (in portuguese). Further information can be found in Moodle platform.
Instance names are formatted as follows: for VFR60_10_3_Gap, 60 refers to the number of tasks, 5 refers to the number of machines, 3 refers to instance id within 60_10 family.
The format of files is described in last paragraph of Section 3.1 of Vallada et al. (2015).
| Instance | |N| | |M| | BKS |
|---|---|---|---|
| VFR10_15_1_Gap | 10 | 15 | 1307 |
| VFR20_10_3_Gap | 20 | 10 | 1592 |
| VFR20_20_1_Gap | 20 | 20 | 2270 |
| VFR60_5_10_Gap | 60 | 5 | 3663 |
| VFR60_10_3_Gap | 60 | 10 | 3423 |
| VFR100_60_1_Gap | 100 | 60 | 9395 |
| VFR500_40_1_Gap | 500 | 40 | 28548 |
| VFR500_60_3_Gap | 500 | 60 | 31125 |
| VFR600_20_1_Gap | 600 | 20 | 31433 |
| VFR700_20_10_Gap | 700 | 20 | 36417 |
Note : Instances mirrored from Web of Instances.
Note 2: Best known solutions are presented in supplementary material from Vallada, Ruiz e Framinan (2015).
Note 3: A mathematical formulation of the problem can be found in Tseng et al. (2004).
Instance names are similar to PFSP: for 20on4Rp50Rs50_1, 20 refers to the number of tasks, 4 refers to the number of machines, 1 refers to instance id within 20_4 family. Other components of instance name describe parameters used on instance generation. Those information can be ignored.
The third paragraph of Section 4 of Ezugwu (2019) gives a short description of instance file format.
| Instance | |N| | |M| | BKS |
|---|---|---|---|
| 20on4Rp50Rs50_1 | 20 | 4 | 527.80 ± 15.43 |
| 60on8Rp50Rs50_1 | 60 | 8 | 820.00 ± 9.62 |
| 60on4Rp50Rs50_1 | 60 | 4 | 1673.20 ± 43.67 |
| 80on8Rp50Rs50_1 | 80 | 8 | 1089.00 ± 7.25 |
| 80on12Rp50Rs50_1 | 80 | 12 | 711.60 ± 5.73 |
| 100on2Rp50Rs50_1 | 100 | 2 | 5872.00 ± 33.32 |
| 100on6Rp50Rs50_1 | 100 | 6 | 1858.40 ± 9.07 |
| 100on8Rp50Rs50_1 | 100 | 8 | 1371.00 ± 12.10 |
| 120on12Rp50Rs50_1 | 120 | 12 | 1087.80 ± 32.26 |
| 120on10Rp50Rs50_1 | 120 | 10 | 1326.80 ± 13.46 |
Note : Instances mirrored from Scheduling Research Virtual Center.
Note 2: Best known solutions are presented in Ezugwu (2019).
Note 3: A mathematical formulation of the problem can be found in Ezugwu (2019).
TSPDL instances are adaptations from classic datasets for TSP. Instance names use the following convention: for bayg29_10_1, bayg indicates the dataset to which the instance belongs, and29 indicates the number of vertices. The other informations can be ignored.
Instance format for bayg, gr, and ulysses is structured and self explanatory.
For KroA and pcb families, check the paper of Reinelt (1991).
| Instance | BKS (avg) | BKS (instance) |
|---|---|---|
| bayg29_10_1 | 1713.60 | 1610 |
| bayg29_50_1 | 2091.00 | 2743 |
| gr17_25_1 | 2237.70 | 2265 |
| gr48_10_1 | 6635.70 | 5046 |
| gr48_25_1 | 5800.30 | 5161.65 |
| KroA200_50_1 | 30665.20 | Unpublished |
| KroA200_75_1 | 30896.10 | Unpublished |
| pcb442_50_1 | 59858.30 | Unpublished |
| pcb442_75_1 | 61010.10 | Unpublished |
| Ulysses_22_50_1 | 8425.60 | 8290 |
Note : Instances mirrored from The TSPDL Lib.
Note 2: Best known solutions are presented in Todosijević et al. (2017).
Note 3: A mathematical formulation of the problem can be found in Rakke et al. (2012).
Note 4: The BKR (avg) column presents the average best known solution to the entire instance family (bayg29_10, KroA200_50, and so forth). When available, the BKS of individual instances are presented in column BKS (instance).
Students and other academic members of Institute of Informatics can request a copy of the proprietary solver CPLEX, under strictly academic conditions. The software can be downloaded from through OnTheHub website. To request credentials to the OnTheHub, contact Library Chief Beatriz Haro.