Prophesy is a tool set for parameter synthesis of parametric Markov models. It can work with a variety of backend tools. The release of Prophesy is accompanied by an overview paper. To get started, see the notes further below.
Please notice that prophesy is academic software, and mostly meant as a sandbox for developing new algorithms. Prophesy is licensed under the GPL License. If you are interested in other licensing options, do not hesitate to contact us!
We advise users to follow this guide. The following outline is very brief. Make sure you have CArL in the c++14 version installed.
Then:
python setup.py develop
installs the required dependencies and prophesy.
It will create prophesy/prophesy.cfg and prophesy/dependencies.cfg which you might want to extend.
Running:
python -m pytest tests
python -m pytest scripts/tests
executes varying tests. Any occurrences of s show that your support currently does not contain some optional dependencies.
The command line tools are available in the scripts folder.
Prophesy is mainly developed at RWTH Aachen University by:
Prophesy received notable contributions from:
- Harold Bruintjes
- Tom Janson
- Lutz Klinkenberg
We would like to thank Christian Hensel and Tim Quatmann for their contributions in the Storm backend, Murat Cubuktepe for his support in developing the QCQP-driven feasibility sampling, and Gereon Kremer for his support of CArL. Prophesy is developed in close cooperation with Nils Jansen, Joost-Pieter Katoen, and Erika Abraham.