Defines bound classes for real numbers. More...

#include <es/eoIntBounds.h>

Inheritance diagram for eoIntBounds:

Public Member Functions
virtual bool	isBounded (void) const =0
	Self-Test: true if *both* a min and a max.
virtual bool	hasNoBoundAtAll (void) const =0
	Self-Test: true if no min *and* no max hence no further need to test/truncate/fold anything.
virtual bool	isMinBounded (void) const =0
	Self-Test: bounded from below???
virtual bool	isMaxBounded (void) const =0
	Self-Test: bounded from above???
virtual bool	isInBounds (double) const =0
	Test on a value: is it in bounds?
virtual void	foldsInBounds (double &) const =0
	Put value back into bounds - by folding back and forth.
virtual void	foldsInBounds (long int &i) const
	foldsInBounds for ints: call the method for double and convert back
virtual void	truncate (double &) const =0
	Put value back into bounds - by truncating to a boundary value.
virtual void	truncate (long int &i) const
	truncate for ints: call the method for double and convert back
virtual long int	minimum () const =0
	get minimum value std::exception if does not exist
virtual long int	maximum () const =0
	get maximum value std::exception if does not exist
virtual long int	range () const =0
	get range std::exception if unbounded
virtual double	uniform (eoRng &_rng=eo::rng) const =0
	random generator of uniform numbers in bounds uses same naming convention than eo::rng std::exception if unbounded
virtual long int	random (eoRng &_rng=eo::rng) const =0
virtual eoIntBounds *	dup () const =0
	for memory managements - ugly

Detailed Description

Defines bound classes for real numbers.

Scalar type: ------------ Basic class is eoIntBounds, a pure virtual.

The following pure virtual methods are to be used in mutations:

void foldsInBounds(long int &) that folds any value that falls out of the bounds back into the bounds, by bouncing on the limit (if any)
bool isInBounds(long int) that simply says whether or not the argument is in the bounds
void truncate(long int &) that set the argument to the bound value it it exceeds it

So mutation can choose

iterate trying until they fall in bounds,
only try once and "restd::pair" by using the foldsInBounds method
only try once and restd::pair using the truncate method (will create a huge bias toward the bound if the soluiton is not far from the bounds)

There is also a uniform() method that generates a uniform value (if possible, i.e. if bounded) in the interval.

Derived class are eoIntInterval that holds a minimum and maximum value, eoIntNoBounds the "unbounded bounds" (-infinity, +infinity) eoIntBelowBound the half-bounded interval [min, +infinity) eoIntAboveBound the half-bounded interval (-infinity, max]

THis file also contains the declaration of *the* global object that is the unbounded bound

Definition at line 70 of file eoIntBounds.h.

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

eoIntBounds.h

Public Member Functions

Detailed Description