MTToolBox::AlgorithmEquidistribution< U, V > Class Template Reference

Calculate dimension of equi-distribution of output of pseudo random number generators. More...

#include <AlgorithmEquidistribution.hpp>

Public Member Functions
	AlgorithmEquidistribution (const ECGenerator &rand, int bit_length, int mexp=0)
	Constructor. More...
	~AlgorithmEquidistribution ()
	Destructor. More...
int	get_all_equidist (int veq[])
	Calculate dimension of equi-distribution with v-bit accuracy. More...
int	get_equidist (int *sum_equidist)
	Calculate dimension of equi-distribution with v-bit accuracy. More...

Detailed Description

template<typename U, typename V = U>
class MTToolBox::AlgorithmEquidistribution< U, V >

Calculate dimension of equi-distribution of output of pseudo random number generators.

Algorithm that calculates dimension of equi-distribution of output of pseudo random number generators using PIS method1.

Template Parameters

type	of output of pseudo random number generator.
type	of output of parameter generator.

Constructor & Destructor Documentation

template<typename U, typename V = U>

MTToolBox::AlgorithmEquidistribution< U, V >::AlgorithmEquidistribution ( const ECGenerator &  rand, int  bit_length, int  mexp = 0  )

inline

Constructor.

PIS method can calculates dimension of equi-distribution of v-bit accuracy k(v) using intermediate result of calculation of k(v+1). This class calculate multiple k(v) varying v from bit_length to 1.

Parameters

rand	pseudo random number generator
bit_length	bit length from MSB to calculate dimension of equi-distribution. This is first v of k(v).
mexp	Mersenne Exponent. If not specified, internal state size is used. If generator is reducible, this parameter should be specified.

References MTToolBox::AbstractGenerator< U >::bitSize(), and MTToolBox::linear_generator_vector< U, V >::next_state().

template<typename U, typename V = U>

MTToolBox::AlgorithmEquidistribution< U, V >::~AlgorithmEquidistribution ( )

inline

Destructor.

Member Function Documentation

template<typename U , typename V >

int MTToolBox::AlgorithmEquidistribution< U, V >::get_all_equidist

(

int

veq[]

)

Calculate dimension of equi-distribution with v-bit accuracy.

Calculate dimension of equi-distribution with v-bit accuracy k(v) for v = bit_length to 1, and set them into an array veq[]. The return value is sum of d(v)s, which are difference between k(v) and theoretical upper bound at v.

Warning

Only one of get_all_equidist() or get_equidist() can be called only one time. (Bad interface)

Parameters

[out]

veq

an array of k(v)

Returns

sum of d(v)s

Referenced by MTToolBox::AlgorithmReducibleRecursionAndTempering< U, G >::search(), and MTToolBox::AlgorithmRecursionAndTempering< U, V >::search().

template<typename U , typename V >

int MTToolBox::AlgorithmEquidistribution< U, V >::get_equidist

(

int *

sum_equidist

)

Calculate dimension of equi-distribution with v-bit accuracy.

Calculate dimension of equi-distribution with v-bit accuracy k(v) for v = bit_length. sum_equidist is sum of d(v)s, which are difference between k(v) and theoretical upper bound at v.

Warning

Only one of get_all_equidist() or get_equidist() can be called only one time. (Bad interface)

Parameters

[out]

sum_equidist

sum of d(v)s

Returns

k(bit_length)

---

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

• AlgorithmEquidistribution.hpp