Potassco
Association:
All members of our team are affiliated with the University of Potsdam, Germany.
Description:
Potassco,
the "Potsdam Answer Set Solving Collection",
bundles tools for Answer Set Programming developed at the University of Potsdam.
It includes:
- solver clasp
- solver claspD
- grounder Gringo
- grounder + solver Clingo
- incremental grounder + solver iClingo
- grounder + (constraint) solver Clingcon
- experimental grounder Bingo (acronym for "Bottom-up Gringo")
Sources (and binaries) of the systems are publicly available at the
Potassco project page.
We publish the encodings we use on the
Asparagus 2 (prototype) platform.
Further information can be found
here.
Participants:
- Christian Drescher
- Martin Gebser
- Torsten Grote
- Roland Kaminski
- Benjamin Kaufmann
- Arne König
- Max Ostrowski
- Torsten Schaub
- Sven Thiele
System settings:
All of our systems have grounder Gringo and solver clasp at their core,
except for Bingo being a self-contained grounder.
For benchmarks
having particular requirements, we used the following Potassco tools:
- DisjunctiveScheduling: Clingcon
- GrammarBasedInformationExtraction: Bingo
- Reachability: Bingo
- StrategicCompanies: claspD
The following table provides an overview of solver parameter settings we used:
| Benchmark Class | Solver Parameters |
|---|
| HydraulicPlanning | Default |
| HydraulicLeaking |
| CompanyControls |
| Reachability |
| GrammarBasedInformationExtraction |
| HamiltonianPath |
| SchurNumbers |
| TravellingSalesperson |
| GeneralizedSlitherlink |
| HierarchicalClustering |
| Sudoku |
| EdgeMatching |
| FastfoodOptimize |
| CompanyControlsOptimize |
| GraphPartitioning | --sat-pre=20,25,120 --heu=VSIDS |
| Fastfood |
| ChannelRouting |
| BlockedNQueens | --sat-pre=20,25,120 --trans-ext=dynamic --initial-look=10 |
| 15Puzzle |
| KnightTour | --restarts=256 --save-p |
| Solitaire |
| ConnectedDominatingSet | --contr=120 --eq=10 --heuristic=Vmtf --rand-freq=0.02 --rand-watches=no --restarts=no --sat-prepro=yes --strengthen=no --trans-ext=all |
| Sokoban | --heu=VSIDS |
| WeightBoundedDominatingSet | --restarts=256 |
| Labyrinth | --hParam=512,1 |
| Hanoi | --sat-pre=20,25,120 --heu=VSIDS --trans-ext=dynamic |
| WireRouting | --sat-pre=20,25,120 --trans-ext=dynamic |
| DisjunctiveScheduling | --restarts=no --csp-num-as=1 |
| GraphColouring | --sat-pre=20,25,120 --trans-ext=dynamic --initial-l=10 --restarts=no |
| MazeGeneration | --heu=VSIDS --del=3,1.1,1000 --restarts=100,1.5,20000 --local-r |
| StrategicCompanies | --approx-usc=no --contraction=0 --del=3,1.1,100 --eq=no --evaluate-loops --forward-check=1.0 --heuristic=Vmtf --restarts=no
--initial-look=yes --minimize=no --minimize-loops --rand-prop=0.05 --rand-watches=no --reduce-on-restart |
| TravellingSalespersonOptimize | --solution-r |
| SokobanOptimize |
| GolombRuler | --restart-on-model --reset-r --opt-heu --heu=VSIDS |
| MaximalClique | --restart-on-model --reset-r |
| 15PuzzleOpt | --opt-heu --sat-pre=20,25,120 |
| WeightBoundedDominatingSetOptimize | --opt-heu --solution-r --reset-r |
| LabyrinthOptimize | --opt-heu |
Note that option names are automatically expanded if unambiguos, e.g.,
--heu stands for option --heuristic.
|