Collaboration diagram for newton_mod::solve:

Private Member Functions
subroutine	newton_solve (x, n, F)
	Newton algorithm to solve F(x) = 0. More...

Detailed Description

Definition at line 44 of file newton_mod.f90.

Member Function/Subroutine Documentation

subroutine newton_mod::solve::newton_solve	(	real(rp), dimension(:), intent(inout)	x,
		type(newton), intent(inout)	n,
		procedure(rntorn)	F
	)

private

Newton algorithm to solve F(x) = 0.

Approximate conputation of the Jacobian

Krylov iterative solver to solve DF(x)^{-1}.b = GMRes

x = initial guess = numerical solution at algorithm's end

neps = parameter to evaluate DF(x).y :

   DF(x).y \simeq (F(x + epsilon*y - F(x)))/epsilon

stopping criteria = residual < tol with residual = |F(x)|_2/|x|_2

returns : x = numerical solution iter = number of performed iterations res = residual Feval = number of performed operation z–>F(z)

Definition at line 92 of file newton_mod.f90.

The documentation for this interface was generated from the following file: