Struct GaussNewton 

pub struct GaussNewton<T, R, C, F, J> { /* private fields */ }

Gauss-Newton

Solves x in a system of equation F(x) = 0 (where F: Rm ⟶ Rn) by minimizing ||F(x)|| in a least square sense using The Gauss-Newton algorithm (Wikipedia). This struct uses a closure that takes a nalgebra vector (size m) and outputs a nalgebra vector (size n) to represent the overdetermined system of equations. The struct also uses the Jacobian of F, that is represented by a closure taking a nalgebra vector (size m) and outputs a nalgebra matrix of size n x m.

Default Tolerance: 1e-6

Default Max Iterations: 50

Implementations

impl<T, R, C, F, J> GaussNewton<T, R, C, F, J>

where T: Float + ComplexField<RealField = T> + Signed, R: Dim, C: Dim, F: Fn(Vector<T, C, <DefaultAllocator as Allocator<C>>::Buffer<T>>) -> Vector<T, R, <DefaultAllocator as Allocator<R>>::Buffer<T>>, J: Fn(Vector<T, C, <DefaultAllocator as Allocator<C>>::Buffer<T>>) -> Matrix<T, R, C, <DefaultAllocator as Allocator<R, C>>::Buffer<T>>, DefaultAllocator: Allocator<C> + Allocator<R> + Allocator<R, C> + Allocator<C, R> + Allocator<C, C>,

pub fn new(f: F, j: J) -> Self

Create a new instance of the algorithm

Instantiates the Gauss-Newton algorithm using the system of equation F and its Jacobian J.

pub fn with_tol(&mut self, tol: T) -> &mut Self

Updates the solver’s tolerance (Magnitude of Error).

Default Tolerance: 1e-6

pub fn with_itermax(&mut self, max: usize) -> &mut Self

Updates the solver’s amount of iterations done before terminating the iteration

Default Max Iterations: 50

pub fn solve( &self, x0: Vector<T, C, <DefaultAllocator as Allocator<C>>::Buffer<T>>, ) -> SolverResult<Vector<T, C, <DefaultAllocator as Allocator<C>>::Buffer<T>>>

Run the algorithm

Finds x such that ||F(x)|| is minimized where F is the overdetermined system of equations.