Function constrained_least_squares

pub fn constrained_least_squares(
    a: &Matrix,
    b: &[f64],
    c: &Matrix,
    d: &[f64],
) -> NumericResult<ConstrainedSolution>

Minimise ||A x - b||_2 subject to C x = d by the null-space method.

C (p x n) must have full row rank p <= n, and A must have full column rank on the null space of C, so the minimiser is unique. Use this to fit with interpolated points or prescribed end tangents.

Errors

Refuses mismatched sizes, non-finite input, a rank-deficient constraint matrix (redundant or conflicting constraints) and an objective that does not determine the solution on the constraints’ null space.