|
Heidelberg Educational Numerics Library Version 0.27 (from 15 March 2021)
|
Solve nonlinear problem using a damped Newton method. More...
#include <newton.hh>
Public Member Functions | |
| Newton () | |
| constructor stores reference to the model | |
| void | set_maxit (size_type n) |
| maximum number of iterations before giving up | |
| void | set_sigma (double sigma_) |
| void | set_linesearchsteps (size_type n) |
| maximum number of steps in linesearch before giving up | |
| void | set_verbosity (size_type n) |
| control output given 0=nothing, 1=summary, 2=every step, 3=include line search | |
| void | set_abslimit (double l) |
| basolute limit for defect | |
| void | set_reduction (double l) |
| reduction factor | |
| template<class M > | |
| void | solve (const M &model, Vector< typename M::number_type > &x) const |
| do one step | |
| bool | has_converged () const |
| size_type | iterations () const |
Solve nonlinear problem using a damped Newton method.
The Newton solver is parametrized by a model. The model also exports all relevant types for types.