Efficient time stepping for numerical integration using reinforcement learning

We learn step-size controllers for (1) adaptive quadrature rules and (2) adaptive integration schemes for ODEs.
Please refer to the Jupyter notebook files "quadrature.ipynb" and "time_stepper_ODE.ipynb".

These files accompany the following preprint:

Efficient time stepping for numerical integration using reinforcement learning
Michael Dellnitz, Eyke Hüllermeier, Marvin Lücke, Sina Ober-Blöbaum, Christian Offen, Sebastian Peitz, Karlson Pfannschmidt
https://arxiv.org/abs/2104.03562