Function for Phase I (feasibility) of interior-point optimisation. More...

#include <PhaseIBoundedVariableFunctionAndDerivatives.hpp>

Inheritance diagram for ipo_function::detail::PhaseIBoundedVariableFunctionAndDerivatives:

Collaboration diagram for ipo_function::detail::PhaseIBoundedVariableFunctionAndDerivatives:

Public Member Functions
	PhaseIBoundedVariableFunctionAndDerivatives (double const b, bool const upperBound)
	Construct from a function and bound. More...

double	operator() (gsl::vector const &vector)
	The function operator: computes or according as is an upper or lower bound. More...

gsl::vector	gradient (gsl::vector const &vector)
	The function operator: returns the gradient. More...

gsl::matrix	hessian (gsl::vector const &vector)
	The function operator: returns the Hessian. More...

void	setVector (gsl::vector const &vector)
	Set a vector and compute function value, gradient and Hessian efficiently. More...

Public Member Functions inherited from ipo_function::Function
	Function (size_t const size=0)
	Constructor. More...

virtual std::tuple < gsl::vector, gsl::matrix >	derivatives (gsl::vector const &vector)
	Compute or estimate derivative values. More...

Public Member Functions inherited from ipo_function::detail::FunctionBase
	FunctionBase (size_t const size=0)
	Constructor. More...

virtual	~FunctionBase ()=0
	Make the class abstract. More...

size_t	getSize () const
	Get size of vector for function arguments or zero for arbitrary size. More...

Public Member Functions inherited from ipo_function::DerivativesEstimates
	DerivativesEstimates (size_t const size=0)
	Constructor. More...

virtual gsl::matrix	hessian () const

Public Member Functions inherited from ipo_function::GradientEstimate
	GradientEstimate (size_t const size=0)
	Constructor. More...

virtual double	value () const

virtual gsl::vector	gradient () const

Private Attributes
Function &	function
	The function. More...

double const	b
	The bound. More...

bool const	upperBound
	A boolean value: true or false according as b is an upper bound or lower. More...

Additional Inherited Members
Protected Attributes inherited from ipo_function::Function
double	h = std::sqrt( std::numeric_limits<double>::epsilon() )
	Value used for default gradient estimates. More...

double	h2 = std::sqrt( h )
	Value used for default Hessian estimates. More...

Protected Attributes inherited from ipo_function::detail::FunctionBase
size_t const	size
	Size of vector arguments to supply to subclass functions. More...

Protected Attributes inherited from ipo_function::DerivativesEstimates
gsl::matrix	functionHessian
	The Hessian value. More...

Protected Attributes inherited from ipo_function::GradientEstimate
double	functionValue
	The function value. More...

gsl::vector	functionGradient
	The gradient value. More...

Detailed Description

Function for Phase I (feasibility) of interior-point optimisation.

This represents a variable with an upper or lower bound b. The function takes as an argument a vector of size 2 whose first element represents the variable $x$ and whose second represents the variable $s$ used in phase I. If the variable has bound $b$ , the function is $x-s-b$ or $-x-s-b$ according as $b$ is an upper or lower bound.

The variable $x$ is not supplied, but should be assigned in a Constraint object. The extra cost of creating a vector of size 2 to supply as an argument to setVector() should normally be low in comparison to the cost of copying large vectors and matrices. And, if we passed a vector of all the variables of a PhaseIModel object as argument to setVector, we would have to copy a large vector and matrix. Hence, this class is designed for a reasonable compromise. It gives better efficiency for larger problems.

This class also computes derivatives.

Definition at line 47 of file PhaseIBoundedVariableFunctionAndDerivatives.hpp.