1Blurb::
2Quasi-Newton optimization method
3Description::
4This is a Newton method that expects a gradient and computes a
5low-rank approximation to the Hessian.  Each of the Newton-based
6methods are automatically bound to the appropriate OPT++ algorithm
7based on the user constraint specification (unconstrained,
8bound-constrained, or generally-constrained). In the
9generally-constrained case, the Newton methods use a nonlinear
10interior-point approach to manage the constraints.
11
12See \ref topic-package_optpp for info related to all \c optpp methods.
13
14<b>Expected HDF5 Output</b>
15
16If Dakota was built with HDF5 support and run with the
17\ref environment-results_output-hdf5 keyword, this method
18writes the following results to HDF5:
19
20- \ref hdf5_results-best_params
21- \ref hdf5_results-best_obj_fncs (when \ref responses-objective_functions) are specified)
22- \ref hdf5_results-best_constraints
23- \ref hdf5_results-calibration (when \ref responses-calibration_terms are specified)
24
25Topics::	package_optpp, local_optimization_methods
26Examples::
27Theory::
28Faq::
29