This repository contains all of the instances used in the paper Optimal Torpedo Scheduling, they are organised as follows.
small/ # Instances used in small dataset
comp-test/ # Small instances from the competition (included in small)
medium/ # Instances used in medium dataset
large/ # Instances used in large dataset
massive/ # Instances used to test the limits of our approach
comp/ # ACP 2016 Competition instances
configs/ # Configuration files used for generating instances
All instances were generated using the instance generator available from the ACP website and the configuration files included. Problem details are available at http://cp2016.a4cp.org/program/acp-challenge/problem.html.
Results are summurised below, for full details, see results.csv file. Zeros for both nTorpedoes and desulfTime indicate that the instance was infeasible.
| Instance | nTorpedoes | desulfTime | total runtime |
|---|---|---|---|
| config1 30x20 | 3 | 125 | 0.315 |
| config2 30x20 | 3 | 144 | 0.377 |
| config3 30x20 | 3 | 84 | 0.153 |
| config1 100x50 | 3 | 77 | 0.400 |
| config2 100x50 | 4 | 154 | 0.340 |
| config3 100x50 | 3 | 190 | 0.390 |
| config1 300x200 | 3 | 1482 | 1.911 |
| config2 300x200 | 3 | 720 | 2.330 |
| config3 300x200 | 4 | 615 | 1.873 |
| instance01 | 4 | 7695 | 41.331 |
| instance02 | 4 | 5302 | 118.720 |
| instance03 | 3 | 27150 | 415.389 |
| instance04 | 3 | 10676 | 34.754 |
| instance05 | 4 | 16308 | 574.852 |
| instance06 | 4 | 7755 | 1133.542 |
| config1 300x100 | 4 | 216 | 1.615 |
| config2 300x100 | 4 | 52 | 1.701 |
| config3 300x100 | 4 | 56 | 1.642 |
| config1 500x200 | 4 | 130 | 5.337 |
| config2 500x200 | 3 | 1150 | 5.381 |
| config3 500x200 | 4 | 238 | 7.579 |
| config1 1000x500 | 4 | 546 | 28.414 |
| config2 1000x500 | 0 | 0 | 0.482 |
| config3 1000x500 | 4 | 432 | 25.661 |
| config1 1500x1000 | 4 | 1250 | 87.387 |
| config2 1500x1000 | 0 | 0 | 1.297 |
| config3 1500x1000 | 4 | 1815 | 88.609 |
| config1 2000x1000 | 4 | 1638 | 194.296 |
| config2 2000x1000 | 5 | 870 | 552.648 |
| config3 2000x1000 | 5 | 845 | 354.375 |
| config1 3000x500 | 5 | 25 | 1897.999 |
| config2 3000x500 | 5 | 140 | 412.892 |
| config3 3000x500 | 5 | 60 | 440.939 |
| config1 3000x1000 | 5 | 468 | 519.113 |
| config2 3000x1000 | 0 | 0 | 3.254 |
| config3 3000x1000 | 4 | 564 | 529.804 |
| config1 3000x2000 | 4 | 5301 | 773.672 |
| config2 3000x2000 | 0 | 0 | 5.217 |
| config3 3000x2000 | 4 | 3180 | 876.566 |
| config1 10000x5000 | 4 | 10374 | 102591.355 |
| config2 10000x5000 | 0 | 0 | 57.092 |
| config3 10000x5000 | 5 | 2587 | 145594.688 |
| config1 100000x50000 | 5 | 28237 | 4310.742 ^ |
| config2 100000x50000 | 0 | 0 | 9.224 ^ |
| config3 100000x50000 | 5 | 20860 | 13275.286 ^ |
^ Note that these were run with a forward limit of 20
- Adrian Goldwaser, Andreas Schutt. Optimal Torpedo Scheduling. In Proceedings of Principles and Practice of Constraint Programming - CP2017, 2017.