This repository provides a generator for parity games that bundles different ways to generate parity games, and uses parallel processing to generate over 1000 parity games. Currently four classes of games are included: random games, synthetic games that are hard for certain algorithms, model checking problems and equivalence checking problems.
If you use the parity game generator or the generated games in your paper, please cite
J.J.A. Keiren. Benchmarks for Parity Games. FSEN 2015, Lecture Notes in Computer Science 9392, Springer-Verlag, pp. 126-142, 2015.
The generator also collects structural information about the parity game and stores this in the structured YAML format. For convenience some utilities are provided to check consistency of the output, and to transform the YAML data into an SQLite database. This latter database allows quick querying, and plotting of data.
The current version of the parity game generator was developed to evaluate equivalence reductions of parity games, and therefore includes the possibility to reduce games using 4 types of equivalences (see also the "thesis" tag).
Using this script, a set of parity games was generated as a benchmark suite. The games are reported in:
- Chapter 5 of J.J.A. Keiren. Advanced Reduction Techniques for Model Checking. PhD thesis, Eindhoven University of Technology, 2013.
Furthermore, a description of the set of benchmarks has been published separately as:
- J.J.A. Keiren. Benchmarks for Parity Games. FSEN 2015, Lecture Notes in Computer Science 9392, Springer-Verlag, pp. 126-142, 2015.
- J.J.A. Keiren. Benchmarks for Parity Games (extended version). arXiv:1407.3121 [cs.LO]
The generated games, and the corresponding data is available from figshare, DOI: 10.6084/m9.figshare.6004130. You can download the following:
- Model checking (
modelchecking.zip) Parity games encoding model checking problems. - Equivalence checking (
equivalencechecking.zipandequivalencechecking-hesselink.zip) Parity games encoding equivalence checking problems. This comes in two files due to filesize restrictions on figshare. - PGSolver (
pgsolver.zip) Parity games generated using PGSolver. These include: random games, and games that are hard for certain algorithms. - MLSolver (
mlsolver.zip) Parity games generated using MLSolver. These are games that encode satisfiability and validity problems for modal logics. - Data (yaml) (
data.yaml) This is a file in YAML format that includes high-level information about all parity games described before. Detailed (structural) information about the games can be found as separate YAML files in the respectipe ZIP archives. - Data (sqlite) (
data.sqlite) This contains a database in SQLite format that was derived from the YAML files provided in the archives, and the previous data file. The sole purpose of this derived data is to enable efficient plotting of results. The plots in the paper have been based on this file.
For the time being, the data is also available from here. It is, however, recommended to use the link on figshare.
All parity games are in max parity game format, i.e. they are based on the assumption that even wins the game if the greatest priority that occurs infinitely often is even (opposed to min parity games in which the least priority is considered).
If you would like to add parity games to this benchmark suite, or would like to add extra statistics, feel free to contact me e.g. through Jeroen.Keiren (you know what goes here) ou.nl.
The tool is currently Linux-only due to the use of a platform-specific script to limit time and memory usage.
Before the script can be used, all tools that are called by the tool need to be installed. We assume that the check out lives in /path/to/paritygame-generator. The tools can be installed by executing the following command in the root of the paritygame-generator directory:
./install_prerequisites.sh
In this case, make sure to execute the following before calling run.py below.
export PATH=/path/to/paritygame-generator/tools/install/bin:$PATH
export LD_LIBRARY_PATH=/path/to/paritygame-generator/tools/install/lib:$LD_LIBRARY_PATH
Alternatively, all tools can be installed manually; in this case you need to
make sure that the following tools can be found in the $PATH.
gistgoallps2ltslps2pbesmcrl22lpsmlsolverpbes2bespginfopgsolver
Experiments should be run only if the tools listed in the previous section are in the
PATH. Then invoke
python run.py [-v[v[v[...]]]] [-jN] [yamlfile]
to run the experiments. More v's means more verbose. For N a positive integer, N jobs are started if -jN is given. If a filename is given on the command line, results are put in that file. The file is not overwritten: if it already exists, then the experiments for which it contains results are skipped.