Here, you can find the code for the paper 'Stability Analysis of Complementarity Systems with Neural Network Controllers' from the conference HSCC 2021.
In order to run the examples, both MATLAB and Python are required.
The linear complementarity problems (LCPs) generated by this library are solved using http://pages.cs.wisc.edu/~ferris/path.html.
MOSEK is used in order to solve the optimization problems (https://www.mosek.com/).
Optimization problems are formulated using https://yalmip.github.io/ (YALMIP).
pathlcp, mosek, yalmip will need to be in the MATLAB path for the examples to run.
Python3 is used for all of the examples with the following additional packages:
numpy: https://numpy.org/
pytorch: https://pytorch.org/
scipy.io: https://docs.scipy.org/doc/scipy/reference/tutorial/io.html
numpy, pytorch, scipy.io should be imported for the examples to run.
The library can be used to verify stability of linear complementarity systems with neural network controllers.
double_int: Double integrator (Example 5.1)
cartpole: Cart-pole with soft walls (Example 5.2)
box_friction: Box with friction (Example 5.3)
five_carts: Five carts with soft contacts (Example 5.4)