braincell.quad.diffrax_bwd_euler_step

braincell.quad.diffrax_bwd_euler_step(target, *args, tol=1e-05)

Advance one step with diffrax’s implicit (backward) Euler solver.

Wraps diffrax.ImplicitEuler, an \(L\)-stable first-order implicit method. The implicit residual is solved by a diffrax.VeryChord chord iteration with matching relative and absolute tolerances.

Parameters:

• target (DiffEqModule) – Module whose differential states will be advanced.

• *args – Extra positional arguments forwarded to target’s integration hooks.

• tol (float, optional) – Combined relative and absolute tolerance for the chord root-finder. Defaults to 1e-5.

Returns:

target’s state is updated in place.

Return type:

None

Raises:

ModuleNotFoundError – If diffrax is not installed.

See also

backward_euler_step

Native (no-diffrax) linearised backward Euler.

implicit_euler_step

Native Newton-iteration backward Euler.