.. _responses-numerical_hessians:

""""""""""""""""""
numerical_hessians
""""""""""""""""""


Hessians are needed and will be approximated by finite differences


.. toctree::
   :hidden:
   :maxdepth: 1

   responses-numerical_hessians-fd_step_size
   responses-numerical_hessians-relative
   responses-numerical_hessians-absolute
   responses-numerical_hessians-bounds
   responses-numerical_hessians-forward
   responses-numerical_hessians-central


**Specification**

- *Alias:* None

- *Arguments:* None


**Child Keywords:**

+-------------------------+--------------------+--------------------+---------------------------------------------+
| Required/Optional       | Description of     | Dakota Keyword     | Dakota Keyword Description                  |
|                         | Group              |                    |                                             |
+=========================+====================+====================+=============================================+
| Optional                                     | `fd_step_size`__   | Step size used when computing gradients and |
|                                              |                    | Hessians                                    |
+-------------------------+--------------------+--------------------+---------------------------------------------+
| Optional (Choose One)   | Step Scaling       | `relative`__       | (Default) Scale step size by the parameter  |
|                         |                    |                    | value                                       |
|                         |                    +--------------------+---------------------------------------------+
|                         |                    | `absolute`__       | Do not scale step-size                      |
|                         |                    +--------------------+---------------------------------------------+
|                         |                    | `bounds`__         | Scale step-size by the domain of the        |
|                         |                    |                    | parameter                                   |
+-------------------------+--------------------+--------------------+---------------------------------------------+
| Optional (Choose One)   | Finite Difference  | `forward`__        | (Default) Use forward differences           |
|                         | Type               +--------------------+---------------------------------------------+
|                         |                    | `central`__        | Use central differences                     |
+-------------------------+--------------------+--------------------+---------------------------------------------+

.. __: responses-numerical_hessians-fd_step_size.html
__ responses-numerical_hessians-relative.html
__ responses-numerical_hessians-absolute.html
__ responses-numerical_hessians-bounds.html
__ responses-numerical_hessians-forward.html
__ responses-numerical_hessians-central.html



**Description**


The ``numerical_hessians`` specification means that Hessian information
is needed and will be computed with finite differences using either
first-order gradient differencing (for the cases of
``analytic_gradients`` or for the functions identified by
``id_analytic_gradients`` in the case of ``mixed_gradients``) or first- or
second-order function value differencing (all other gradient
specifications). In the former case, the following expression

.. math:: 

   \nabla^2 f ({\bf x})_i ong
   \frac{\nabla f ({\bf x} + h {\bf e}_i) - \nabla f ({\bf x})}{h}

estimates the :math:`i^{th}`  Hessian column, and in the latter case, the
following expressions

.. math:: 

   \nabla^2 f ({\bf x})_{i,j} ong ````\frac{f({\bf x} + h_i {\bf e}_i + h_j {\bf e}_j) -
   f({\bf x} + h_i {\bf e}_i) -
   f({\bf x} - h_j {\bf e}_j) +
   f({\bf x})}{h_i h_j}

and

.. math:: 

   \nabla^2 f ({\bf x})_{i,j} ong ````\frac{f({\bf x} + h {\bf e}_i + h {\bf e}_j) -
   f({\bf x} + h {\bf e}_i - h {\bf e}_j) -
   f({\bf x} - h {\bf e}_i + h {\bf e}_j) +
   f({\bf x} - h {\bf e}_i - h {\bf e}_j)}{4h^2}

provide first- and second-order estimates of the :math:`ij^{th}`  Hessian term.
Prior to Dakota 5.0, Dakota always used second-order estimates.
In Dakota 5.0 and newer, the default is to use first-order estimates
(which honor bounds on the variables and
require only about a quarter as many function evaluations
as do the second-order estimates), but specifying ``central``
after ``numerical_hessians`` causes Dakota to use the old second-order
estimates, which do not honor bounds. In optimization algorithms that
use Hessians, there is little reason to use second-order differences in
computing Hessian approximations.