The package is deprecated in favor of https://github.com/aimclub/evoguess-ai

EvoGuess

Framework for finding decomposition sets and estimating hardness of SAT instances. The search for decomposition sets is realized via the optimization of the special pseudo-Boolean black-box functions that estimate the hardness of the decomposition corresponding to the employed decomposition method and the considered set. To optimize the value of such functions the framework uses metaheuristic algorithms, in particular, the evolutionary ones.

Installation

At the moment, only manual installation is available.

$ git clone [email protected]:ctlab/evoguess.git

Dependencies

$ pip install numpy
$ pip install python-sat
$ pip install python-dotenv

To use EvoGuess in MPI mode, you also need to install:

$ pip install mpi4py

Environment

Create .env file using create_env.sh script

$ cd evoguess
$ ./create_env.sh

How to use

Run main.py using configuration file.

$ python3 main.py -f <configuration file>

Or use configuration json-string.

$ python3 main.py -l <configuration json-string>

MPI mode

The EvoGuess can be run in MPI mode as follows:

$ mpiexec -n <workers> -perhost <perhost> python3 -m mpi4py.futures main.py -f <configuration file>

Example

Run on example configuration config.json.

$ python3 main.py -f config.json

Documentation

Documentation is available here and includes installation instructions, base and advanced usage manual.