jonasbreuling / scipy_dae

Python implementation of solvers for differential algebraic equation's (DAEs) that should be added to scipy one day.

scipy_dae - solving differential algebraic equations in Python

Python implementation of solvers for differential algebraic equation's (DAEs) that should be added to scipy one day. See the GitHub repository.

Install

An editable developer mode can be installed via

python -m pip install -e .[dev]

The tests can be started using

python -m pytest --cov

scipy_dae's Issues

Add step size `h` to `fun` and `jac` for optimal preconditioning.

Add `Radau` for `M y' = f(x, y)` since the error estimate is flawed for `F(x, y, y') = 0`

Add correct embedded method proposed by Fabien

See Section 5.3.5 in Fabien2009.

Add weighting between embedded error and dense output error AS done by RADAR5

Enable `vectorization`.

Dense output for `yp`.

Improve tests and test coverage.

• Add all tests of scipy.integrate._ivp
• Add index 1 DAE of Robertson
• Add index 1 DAE of Jay
• Find Pendulum index 1 problem again

Locate tests as done by scipy

Initial step size

Chose initial step size as proposed by Petzold for DASSL and compare with Shampine.

Add events with signature `event(t, y, yp) = 0`

Use Hermite interpolation for dense output and error estimate

See wikipedia and Shampine's article.

Use `scipy.integrate._ivp.common.num_jac` if possible.

Build with `meson-python`

See https://mesonbuild.com/meson-python/tutorials/introduction.html