OptFDNewton is a derived class of OptNewtonLike. More...
Public Member Functions | |
int | checkDeriv () |
Compare the analytic gradient with the finite difference gradient. | |
OptFDNewton (NLP1 *p, TOLS t) | |
OptFDNewton (NLP1 *p, UPDATEFCN u) | |
OptFDNewton (NLP1 *p) | |
OptFDNewton () | |
Default Constructor. | |
NEWMAT::SymmetricMatrix | updateH (NEWMAT::SymmetricMatrix &H, int k) |
Compute finite-difference approximation to Hessian of the objective function. | |
virtual | ~OptFDNewton () |
Destructors. |
OptFDNewton is a derived class of OptNewtonLike.
This class implements an unconstrained Newton's Method with a finite-difference approximation to the Hessian. The user can select from the following globalization strategies: linesearch, trust-region, and trustpds.
Definition at line 26 of file OptFDNewton.h.
OPTPP::OptFDNewton::OptFDNewton | ( | ) | [inline] |
Default Constructor.
Definition at line 35 of file OptFDNewton.h.
References OPTPP::OptimizeClass::method.
OPTPP::OptFDNewton::OptFDNewton | ( | NLP1 * | p | ) | [inline] |
p | a pointer to an NLP1. |
Definition at line 39 of file OptFDNewton.h.
References OPTPP::OptimizeClass::method.
OPTPP::OptFDNewton::OptFDNewton | ( | NLP1 * | p, | |
UPDATEFCN | u | |||
) | [inline] |
p | a pointer to an NLP1. | |
u | a function pointer. |
Definition at line 45 of file OptFDNewton.h.
References OPTPP::OptimizeClass::method.
OPTPP::OptFDNewton::OptFDNewton | ( | NLP1 * | p, | |
TOLS | t | |||
) | [inline] |
p | a pointer to an NLP1. | |
t | tolerance class reference. |
Definition at line 51 of file OptFDNewton.h.
References OPTPP::OptimizeClass::method.
virtual OPTPP::OptFDNewton::~OptFDNewton | ( | ) | [inline, virtual] |
Destructors.
Definition at line 57 of file OptFDNewton.h.
int OPTPP::OptFDNewton::checkDeriv | ( | ) | [virtual] |
Compare the analytic gradient with the finite difference gradient.
Reimplemented from OPTPP::OptNewtonLike.
NEWMAT::SymmetricMatrix OPTPP::OptFDNewton::updateH | ( | NEWMAT::SymmetricMatrix & | H, | |
int | k | |||
) | [virtual] |
Compute finite-difference approximation to Hessian of the objective function.
Implements OPTPP::OptNewtonLike.