Python toolbox for abstraction and controller synthesis for mode-counting objectives

Working Python environment, with

• numpy
• scipy
• matplotlib (for visualization)
• networkx
• One of the following ILP solvers
  • Mosek
  • Gurobi

1. Download Gurobi Optimizer and request a license file. Place it in e.g. ~/gurobi.lic
2. Go to the installation path (OS X: /Library/gurobi650/mac64/) and run python setup.py install
3. Add an environment variable to where the license file is located: e.g. export GRB_LICENSE_FILE=~/gurobi.lic or echo 'export GRB_LICENSE_FILE=~/gurobi.lic' >> ~/.bashrc
4. Run a test: e.g. python /Library/gurobi650/mac64/examples/python/mip1.py

For more help see the quick start guides at http://www.gurobi.com/documentation/.

python example_simple.py
python example_abstraction_lin2d.py

python -m unittest discover

(currently outdated)

See the sphinx documentation.

• Write code that purges forbidden parts of graph: difficult because of order_fcn
• Add option to weigh different classes in counting constraints
• Allow for different counts in prefix
• Non-conservative suffix counting option