Numerical Optimization Module

Numerical optimization routines. More...

Classes
class	itpp::Newton_Search
	Newton Search. More...
class	itpp::Line_Search
	Line Search. More...
Enumerations
enum	itpp::Newton_Search_Method { BFGS }
	Newton Search method.
enum	itpp::Line_Search_Method { Soft, Exact }
	Line Search method.
Functions
vec	itpp::fminunc (double(*function)(const vec &), vec(*gradient)(const vec &), const vec &x0)
	Unconstrained minimization.

---

Detailed Description

Numerical optimization routines.

---

Function Documentation

vec itpp::fminunc	(	double(*)(const vec &)	function,
		vec(*)(const vec &)	gradient,
		const vec &	x0
	)

Unconstrained minimization.

Unconstrained minimization using a Newton or Quasi-Newton optimization method that try to minimize the objective function given an initial guess .

The function and the gradient need to be known and supplied.

The default algorithm is a Quasi-Newton search using BFGS updates of the inverse Hessian matrix.

References itpp::Newton_Search::search(), and itpp::Newton_Search::set_functions().