647 |
|
|
648 |
@deffn {Solver} gsl_multiroot_fsolver_dnewton |
@deffn {Solver} gsl_multiroot_fsolver_dnewton |
649 |
|
|
650 |
@cindex Discrete newton algorithm for multidimensional roots |
@cindex Discrete Newton algorithm for multidimensional roots |
651 |
@cindex Newton algorithm, discrete |
@cindex Newton algorithm, discrete |
|
@cindex Discrete Newton algorithm for nonlinear systems |
|
652 |
|
|
653 |
The @dfn{discrete Newton algorithm} is the simplest method of solving a |
The @dfn{discrete Newton algorithm} is the simplest method of solving a |
654 |
multidimensional system. It uses the Newton iteration |
multidimensional system. It uses the Newton iteration |
1001 |
not make any significant difference to its behavior, since it able to |
not make any significant difference to its behavior, since it able to |
1002 |
approximate the Jacobian numerically with sufficient accuracy. To |
approximate the Jacobian numerically with sufficient accuracy. To |
1003 |
illustrate the behavior of a different derivative solver we switch to |
illustrate the behavior of a different derivative solver we switch to |
1004 |
@code{gnewton}. This is a traditional newton solver with the constraint |
@code{gnewton}. This is a traditional Newton solver with the constraint |
1005 |
that it scales back its step if the full step would lead "uphill". Here |
that it scales back its step if the full step would lead "uphill". Here |
1006 |
is the output for the @code{gnewton} algorithm, |
is the output for the @code{gnewton} algorithm, |
1007 |
|
|