PSL is a minimalistic library for simulating dynamical systems in Python using SciPy library.
Authors: Jan Drgona, Aaron Tuor, Stefan Dernbach, James Koch, Soumya Vasisht, Wenceslao Shaw Cortez, Draguna Vrabie
- Optional pytorch backend via torchdiffeq for all systems in autonomous.py
- Faster numpy integration for systems in autonomous.py
- Sampling of initial conditions and control inputs over dynamic range via get_U, and get_x0 functions for all systems.
- Coupled dynamical systems in coupled_systems.py module
See online Documentation.
$ conda create -n psl python=3.8
$ conda activate psl
(psl) $ conda install numpy
(psl) $ conda install scipy
(psl) $ conda install matplotlib
(psl) $ pip install pyts
(psl) $ pip install tqdm
(psl) $ conda install torchdiffeqimport psl
# instantiate selected dynamical system model
model = psl.systems['Duffing'](ts=0.01)
# simulate the dynamical system over nsim steps
out = model.simulate(nsim=2000)
# plot time series and phase portrait
psl.plot.pltOL(Y=out['Y'], X=out['X'])
psl.plot.pltPhase(X=out['Y'])
See folder tests.