pub fn solve_numeric_system_with(
eqs: &[Ex],
vars: &[Ex],
x0: &[f64],
opts: &NewtonOpts,
) -> Result<Vec<f64>, SymplexError>Expand description
Newton’s method with explicit NewtonOpts.
The Jacobian is computed symbolically (crate::matrix::jacobian) and
compiled to native closures where possible. Each step solves
J·Δ = −F by Gaussian elimination with partial pivoting; with
damping on, the step is halved until the residual norm decreases.
§Errors
SymplexError::InvalidArgumentif the system is not square orx0has the wrong length.SymplexError::ComputationFailedif the Jacobian becomes singular or the iteration does not converge withinmax_itersteps; the message reports the final residual norm and iterate.