MinlpPenaltyFunc Class Reference

A Penalty Function for a MinlpProblem. More...

#include <problem.h>

Inheritance diagram for MinlpPenaltyFunc:

[legend]

Collaboration diagram for MinlpPenaltyFunc:

[legend]

List of all members.

Public Member Functions
	MinlpPenaltyFunc (Pointer< MinlpProblem > minlp_, Pointer< Param > penalty_param=NULL)
	Constructor for a MinlpProblem.
double	eval (const UserVector< double > &x) const
	Evaluates the penalty function for a UserVector<double>.
int	valgrad (double &val, UserVector< double > &y, const UserVector< double > &x) const
	Computes the value and the gradient of this penalty function.
void	grad (UserVector< double > &y, const UserVector< double > &x) const
	Computes the gradient of this penalty function.
void	HessMult (UserVector< double > &y, const UserVector< double > &z, const UserVector< double > &x) const
	Computes the product of the hessian of this matrix and a UserVector<double>.
virtual void	set_curvature (CurvatureType ct)
virtual CurvatureType	get_curvature () const
void	print (ostream &out) const
	Prints some information about this function.
Public Attributes
dvector	delta
	The vector of delta-values.
Private Attributes
Pointer< MinlpProblem >	minlp
	The MinlpProblem.

---

Detailed Description

A Penalty Function for a MinlpProblem.

The function has the form

Definition at line 327 of file problem.h.

---

Constructor & Destructor Documentation

MinlpPenaltyFunc::MinlpPenaltyFunc	(	Pointer< MinlpProblem >	minlp_,
		Pointer< Param >	penalty_param = NULL
	)			[inline]

Constructor for a MinlpProblem.

Sets the delta-values to 1000.

Parameters:

	minlp_	The MinlpProblem, which PenaltyFunc should be represented.
	penalty_param	The delta-values for penalty of each constraints as parameters, default is "penaltyfunc.res".

Definition at line 343 of file problem.h.

---

Member Function Documentation

double MinlpPenaltyFunc::eval

(

const UserVector< double > &

x

)

const [virtual]

Evaluates the penalty function for a UserVector<double>.

Parameters:

x

The UserVector<double> to compute the value for.

Returns:

The value of this function.

See also:

MinlpPenaltyFunc

valgrad(double&, UserVector<double>&, const UserVector<double>&)

Implements Func.

int MinlpPenaltyFunc::valgrad	(	double &	val,
		UserVector< double > &	y,
		const UserVector< double > &	x
	)			const [virtual]

Computes the value and the gradient of this penalty function.

The gradient of the absolut of an equality constraint is handled as zero, if the gradient of the equality constraint is zero.

Parameters:

	val	A double to store the value in.
	y	A UserVector<double> to store the gradient in.
	x	The UserVector<double> to compute.

Returns:

0

See also:

eval(const UserVector<double>&)

grad(UserVector<double>&, const UserVector<double>&)

Reimplemented from Func.

void MinlpPenaltyFunc::grad	(	UserVector< double > &	y,
		const UserVector< double > &	x
	)			const [virtual]

Computes the gradient of this penalty function.

Calls valgrad(double&, UserVector<double>&, const UserVector<double>&).

Parameters:

	y	The UserVector<double> to store the gradient in.
	x	The UserVector<double> to compute.

See also:

valgrad(double&, UserVector<double>&, const UserVector<double>&)

Implements Func.

void MinlpPenaltyFunc::HessMult	(	UserVector< double > &	y,
		const UserVector< double > &	z,
		const UserVector< double > &	x
	)			const [virtual]

Computes the product of the hessian of this matrix and a UserVector<double>.

Parameters:

	y	The UserVector<double> to store the result in.
	z	The UserVector<double> to compute the hessian for.
	x	The UserVector<double> to multiply with the hessian.

Implements Func.

virtual void MinlpPenaltyFunc::set_curvature

(

CurvatureType

ct

)

[inline, virtual]

Implements Func.

Definition at line 390 of file problem.h.

virtual CurvatureType MinlpPenaltyFunc::get_curvature

(

)

const [inline, virtual]

Implements Func.

Definition at line 391 of file problem.h.

void MinlpPenaltyFunc::print

(

ostream &

out

)

const [inline, virtual]

Prints some information about this function.

Prints the delta vector.

Parameters:

out

The ostream to print to.

See also:

delta

Reimplemented from Func.

Definition at line 398 of file problem.h.

---

Member Data Documentation

Pointer<MinlpProblem> MinlpPenaltyFunc::minlp [private]

The MinlpProblem.

Definition at line 331 of file problem.h.

dvector MinlpPenaltyFunc::delta

The vector of delta-values.

Definition at line 336 of file problem.h.

---

The documentation for this class was generated from the following file:

• problem.h