RandomNumberGenerator< Engine > Class Template Reference

#include <vigra/random.hxx>

Inherits Engine.

Public Member Functions
double	normal () const
double	normal (double mean, double stddev) const
UInt32	operator() () const
	RandomNumberGenerator ()
	RandomNumberGenerator (RandomSeedTag)
	RandomNumberGenerator (UInt32 theSeed, bool ignoreSeed=false)
template<class Iterator >
	RandomNumberGenerator (Iterator init, UInt32 length)
void	seed (RandomSeedTag)
void	seed (UInt32 theSeed, bool ignoreSeed=false)
template<class Iterator >
void	seed (Iterator init, UInt32 length)
double	uniform () const
double	uniform (double lower, double upper) const
double	uniform53 () const
UInt32	uniformInt () const
UInt32	uniformInt (UInt32 beyond) const

Static Public Member Functions
static RandomNumberGenerator &	global ()

Detailed Description

template<class Engine = detail::RandomState<detail::MT19937>>
class vigra::RandomNumberGenerator< Engine >

Generic random number generator.

The actual generator is passed in the template argument Engine. Two generators are currently available:

• RandomMT19937: The state-of-the-art Mersenne Twister with a state length of 2¹⁹⁹³⁷ and very high statistical quality.
• RandomTT800: (default) The Tempered Twister, a simpler predecessor of the Mersenne Twister with period length 2⁸⁰⁰.

Both generators have been designed by Makoto Matsumoto.

Traits defined:

FunctorTraits<RandomNumberGenerator<Engine> >::isInitializer 

is true (VigraTrueType).

Examples:

voronoi.cxx.

Constructor & Destructor Documentation

RandomNumberGenerator

(

)

Create a new random generator object with standard seed.

Due to standard seeding, the random numbers generated will always be the same. This is useful for debugging.

RandomNumberGenerator

(

RandomSeedTag

)

Create a new random generator object with a random seed.

The seed is obtained from the machines current clock() and time() values.

Usage:

RandomNumberGenerator<> rnd = RandomNumberGenerator<>(RandomSeed);

RandomNumberGenerator	(	UInt32	theSeed,
		bool	ignoreSeed = false
	)

Create a new random generator object from the given seed.

The same seed will always produce identical random sequences. If ignoreSeed is true, the given seed is ignored, and the generator is seeded randomly (as if it was constructed with RandomNumberGenerator<>(RandomSeed)). This allows you to switch between random and deterministic seeding at run-time.

RandomNumberGenerator	(	Iterator	init,
		UInt32	length
	)

Create a new random generator object from the given seed sequence.

Longer seed sequences lead to better initialization in the sense that the generator's state space is covered much better than is possible with 32-bit seeds alone.

Member Function Documentation

void seed

(

RandomSeedTag

)

Re-initialize the random generator object with a random seed.

The seed is obtained from the machines current clock() and time() values.

Usage:

RandomNumberGenerator<> rnd = ...;
...
rnd.seed(RandomSeed);

void seed	(	UInt32	theSeed,
		bool	ignoreSeed = false
	)

Re-initialize the random generator object from the given seed.

The same seed will always produce identical random sequences. If ignoreSeed is true, the given seed is ignored, and the generator is seeded randomly (as if seed(RandomSeed) was called). This allows you to switch between random and deterministic seeding at run-time.

void seed	(	Iterator	init,
		UInt32	length
	)

Re-initialize the random generator object from the given seed sequence.

Longer seed sequences lead to better initialization in the sense that the generator's state space is covered much better than is possible with 32-bit seeds alone.

UInt32 operator()

(

)

const

Return a uniformly distributed integer random number in [0, 2³²).

That is, 0 <= i < 2³².

UInt32 uniformInt

(

)

const

Return a uniformly distributed integer random number in [0, 2³²).

That is, 0 <= i < 2³².

Examples:

voronoi.cxx.

UInt32 uniformInt

(

UInt32

beyond

)

const

Return a uniformly distributed integer random number in [0, beyond).

That is, 0 <= i < beyond.

double uniform53

(

)

const

Return a uniformly distributed double-precision random number in [0.0, 1.0).

That is, 0.0 <= i < 1.0. All 53-bit bits of the mantissa are random (two 32-bit integers are used to create this number).

double uniform

(

)

const

Return a uniformly distributed double-precision random number in [0.0, 1.0].

That is, 0.0 <= i <= 1.0. This number is computed by uniformInt() / (2³² - 1), so it has effectively only 32 random bits.

double uniform	(	double	lower,
		double	upper
	)		const

Return a uniformly distributed double-precision random number in [lower, upper].

That is, lower <= i <= upper. This number is computed from uniform(), so it has effectively only 32 random bits.

double normal	(	double	mean,
		double	stddev
	)		const

Return a normal variate (Gaussian) random number with the given mean and standard deviation.

It uses the polar form of the Box-Muller transform.

static RandomNumberGenerator& global ( )

static

Access the global (program-wide) instance of the present random number generator.

Normally, you will create a local generator by one of the constructor calls. But sometimes it is useful to have all program parts access the same generator.

---

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

• vigra/random.hxx