ReachMM

Clone the Repo and its Submodules

git clone --recurse-submodules https://github.com/gtfactslab/ReachMM.git
cd ReachMM

Installing ReachMM into a Conda Environment

conda create -n ReachMM python=3.10
conda activate ReachMM

Install auto_LiRPA (information taken from https://github.com/Verified-Intelligence/auto_LiRPA).

cd auto_LiRPA
python setup.py install

Step back into the root folder and install the ReachMM package and its dependencies.

cd ..
pip install -e .

Reproducing Figures from CDC 2023 Submission

Vehicle Model

cd examples/vehicle

To reproduce Figure 3 for the vehicle model, run the following:

python cdc2023.py

and to reproduce Table I, run the following:

python cdc2023.py --table -N 10

where N specifies the number of runs to average over. This can take a while for large values of N.

Double Integrator Model

cd examples/doubleintegrator

To reproduce Figures 4 and 5 for the double integrator model, and the rows for ReachMM and ReachMM-CG for Table II, run the following:

python cdc2023.py -N 1

where N specifies the number of runs to average over. This can take a while for large values of N.

To reproduce the tree on Figure 2, you'll need to install PyGraphviz, which will require Graphviz (https://pygraphviz.github.io/documentation/stable/install.html).

python cdc2023.py --tree