Collaboration diagram for Minotaur::CxUnivarConstraintData:

Public Member Functions
void	initRelax (RelaxationPtr rel, DoubleVector &tmpX, DoubleVector &grad)
	Creates initial relaxations

void	updateRelax (RelaxationPtr rel, DoubleVector &tmpX, DoubleVector &grad, ModVector &mods)
	Update the current relaxation based on current variable bounds.

bool	isFeasible (const double *x)

double	getViol (const std::vector< double > &x)

void	addSecant (RelaxationPtr rel, ConstVariablePtr iv, ConstVariablePtr ov, FunctionPtr fn, DoubleVector &tmpX, bool init, ModVector &mods)

void	addLin (RelaxationPtr rel, ConstVariablePtr iv, ConstVariablePtr ov, FunctionPtr fn, DoubleVector &tmpX, DoubleVector &grad, bool init, ModVector &mods)

ConstraintPtr	getOriginalCon () const

ConstraintPtr	getSecantCon () const

ConstVariablePtr	getROutVar () const

ConstVariablePtr	getRInputVar () const

ConstraintIterator	linConsBegin ()

ConstraintIterator	linConsEnd ()

char	getSense ()

	CxUnivarConstraintData (double eTol, double vTol, ConstraintPtr newcon, ConstVariablePtr ovar, ConstVariablePtr ivar, char sense)
	Default constructor.

	~CxUnivarConstraintData ()
	Destroy.

Protected Attributes
double	eTol_
	Tolerance for constraint violation.

double	vTol_
	Tolerance for when upper and lower bounds considered equal.

ConstraintPtr	con_
	This is the constraint of the orginal problem.

ConstVariablePtr	iv_

VariablePtr	riv_

ConstVariablePtr	ov_

VariablePtr	rov_

char	sense_

ConstraintPtr	secCon_
	Secant constraint in the relaxation.

ConstraintVector	linCons_
	Array of linearization constraints in the relaxation.

Member Data Documentation

◆ iv_

ConstVariablePtr Minotaur::CxUnivarConstraintData::iv_

protected

Input variable, i.e., x in y = f(x) This should be a pointer to the ‘original’ variable, look at corresponding bounds in the relaxation to get bounds

ConstVariablePtr Minotaur::CxUnivarConstraintData::ov_

protected

Output variable, i.e., y in y = f(x) This is a pointer to the ‘original’ variable

char Minotaur::CxUnivarConstraintData::sense_

protected

'L' => y <= f(x), so only add secants 'G' -> y >= f(x), so only add linearizations 'E' -> y == f(x), so do both

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