#include <RandomNumberGenerator.hpp>

Inheritance diagram for RandomNumberGenerator:

Collaboration diagram for RandomNumberGenerator:

Public Member Functions
double	StandardNormalRandomDeviate ()

double	NormalRandomDeviate (double mean, double stdDev)

double	ranf ()

double	GammaRandomDeviate (double shape, double scale)

double	ExponentialRandomDeviate (double scale)

unsigned	randMod (unsigned base)

template<class T >
void	Shuffle (std::vector< boost::shared_ptr< T > > &rValues)

void	Shuffle (unsigned num, std::vector< unsigned > &rValues)

void	Reseed (unsigned seed)

Public Member Functions inherited from SerializableSingleton< RandomNumberGenerator >
SerializableSingleton < RandomNumberGenerator > *	GetSerializationWrapper () const

Static Public Member Functions
static RandomNumberGenerator *	Instance ()

static void	Destroy ()

Protected Member Functions
	RandomNumberGenerator ()

Private Member Functions
template<class Archive >
void	save (Archive &archive, const unsigned int version) const

template<class Archive >
void	load (Archive &archive, const unsigned int version)

Private Attributes
boost::mt19937	mMersenneTwisterGenerator

boost::variate_generator < boost::mt19937 &, boost::uniform_real<> >	mGenerateUnitReal

boost::variate_generator < boost::mt19937 &, boost::random::normal_distribution_v165<> >	mGenerateStandardNormal

Static Private Attributes
static RandomNumberGenerator *	mpInstance = nullptr

Friends
class	boost::serialization::access

Detailed Description

A special singleton class allowing one to generate different types of random number in a consistent way across platforms and different versions of compilers, standard libraries, and our dependencies. As a result Chaste developers should always use this class to get random numbers rather than accessing C/C++/boost methods directly.

This class is a singleton and an instance should be retrieved with: RandomNumberGenerator* p_gen = RandomNumberGenerator::Instance();

Note for developers that might want to change this class!:

We use boost random number library for random number generation under the hood http://www.boost.org/doc/libs/1_65_0/doc/html/boost_random.html

Generally boost is consistent between our supported versions (listed on https://chaste.cs.ox.ac.uk/trac/wiki/InstallGuides/DependencyVersions) but on a couple of occasions boost has broken backwards compatibility between random numbers for a given seed.

To get around this, we have copied a minimal set of boost's more recent random include headers into the folder global/src/random

If the user is on a recent version of boost, the boost libraries/includes are used by default, if the user is on an older version of boost some of the Chaste copies of newer boost distributions in global/src/random are used instead to give consistent random numbers across boost versions.

Definition at line 81 of file RandomNumberGenerator.hpp.

Constructor & Destructor Documentation

RandomNumberGenerator::RandomNumberGenerator ( )

protected

Protected constructor. Use Instance() to access the random number generator.

Definition at line 47 of file RandomNumberGenerator.cpp.

References mpInstance.

Referenced by Instance().

Member Function Documentation

void RandomNumberGenerator::Destroy ( )

static

Destroy the current instance of the random number generator. The next call to Instance will create a new instance and re-seed. This method *must* be called before program exit, to avoid a memory leak.

Definition at line 68 of file RandomNumberGenerator.cpp.

References mpInstance.

Referenced by AbstractCellBasedTestSuite::tearDown().

double RandomNumberGenerator::ExponentialRandomDeviate ( double scale )

Returns: Generate a random number from an exponential distribution with specified scale parameter.

Parameters

scale The scale parameter of the exponential distribution from which the random number is drawn, often named lambda

Definition at line 138 of file RandomNumberGenerator.cpp.

References mMersenneTwisterGenerator.

Referenced by CellCycleTimesGenerator::GenerateCellCycleTimeSequence(), and ExponentialG1GenerationalCellCycleModel::SetG1Duration().

double RandomNumberGenerator::GammaRandomDeviate	(	double	shape,
		double	scale
	)

Returns: Generate a random number from a gamma distribution with specified shape and scale parameters.

Parameters

shape	the shape parameter of the gamma distribution from which the random number is drawn
scale	the scale parameter of the gamma distribution from which the random number is drawn

Definition at line 124 of file RandomNumberGenerator.cpp.

References mMersenneTwisterGenerator.

Referenced by GammaG1CellCycleModel::SetG1Duration().

RandomNumberGenerator * RandomNumberGenerator::Instance ( )

static

Returns: a pointer to the random number generator object. The object is created the first time this method is called.

Definition at line 59 of file RandomNumberGenerator.cpp.

References mpInstance, and RandomNumberGenerator().

template<class Archive >

void RandomNumberGenerator::load	(	Archive &	archive,
		const unsigned int	version
	)

inlineprivate

Load the RandomNumberGenerator and its member variables.

Note: do not serialize this singleton directly. Instead, serialize the object returned by GetSerializationWrapper().

Parameters

archive	the archive
version	the current version of this class

Definition at line 140 of file RandomNumberGenerator.hpp.

References mGenerateStandardNormal, and mMersenneTwisterGenerator.

double RandomNumberGenerator::NormalRandomDeviate	(	double	mean,
		double	stdDev
	)

Returns: Generate a random number from a normal distribution with given mean and standard deviation.

Parameters

mean	the mean of the normal distribution from which the random number is drawn
stdDev	the standard deviation of the normal distribution from which the random number is drawn

Definition at line 119 of file RandomNumberGenerator.cpp.

References StandardNormalRandomDeviate().

Referenced by StochasticOxygenBasedCellCycleModel::GenerateStochasticG2Duration(), StochasticWntCellCycleModel::GenerateStochasticG2Duration(), and SimpleWntCellCycleModel::SetG1Duration().

unsigned RandomNumberGenerator::randMod ( unsigned base )

Returns: Generate a random number modulo base (i.e. an integer within the range [0, base) == [0,1,..,base-1] ).

Parameters

base	the order of the field of positive integers from which the random number is drawn. This should be no greater than INT_MAX

Definition at line 77 of file RandomNumberGenerator.cpp.

References mMersenneTwisterGenerator, and NEVER_REACHED.

Referenced by CryptShovingCaBasedDivisionRule::CalculateDaughterNodeIndex(), MutableVertexMesh< ELEMENT_DIM, SPACE_DIM >::PerformProtorosetteResolution(), MutableVertexMesh< ELEMENT_DIM, SPACE_DIM >::PerformRosetteRankDecrease(), Shuffle(), PottsBasedCellPopulation< DIM >::UpdateCellLocations(), and CaBasedCellPopulation< DIM >::UpdateCellLocations().

double RandomNumberGenerator::ranf ( )

Returns: Generate a uniform random number in (0,1].

Definition at line 109 of file RandomNumberGenerator.cpp.

References mGenerateUnitReal.

Referenced by RandomDirectionVertexBasedDivisionRule< SPACE_DIM >::CalculateCellDivisionVector(), CryptCentreBasedDivisionRule< ELEMENT_DIM, SPACE_DIM >::CalculateCellDivisionVector(), RandomDirectionCentreBasedDivisionRule< ELEMENT_DIM, SPACE_DIM >::CalculateCellDivisionVector(), CryptVertexBasedDivisionRule< SPACE_DIM >::CalculateCellDivisionVector(), ExclusionCaBasedDivisionRule< SPACE_DIM >::CalculateDaughterNodeIndex(), ShovingCaBasedDivisionRule< SPACE_DIM >::CalculateDaughterNodeIndex(), RandomCellKiller< DIM >::CheckAndLabelSingleCellForApoptosis(), CryptCellsGenerator< CELL_CYCLE_MODEL >::Generate(), CellsGenerator< CELL_CYCLE_MODEL, DIM >::GenerateBasicRandom(), HoneycombMeshGenerator::GetCircularMesh(), CryptProjectionStatistics::GetCryptSection(), CryptStatistics::GetCryptSection(), VertexMesh< ELEMENT_DIM, SPACE_DIM >::GetShortAxisOfElement(), CryptSimulationBoundaryCondition< DIM >::ImposeBoundaryCondition(), PlaneBoundaryCondition< ELEMENT_DIM, SPACE_DIM >::ImposeBoundaryCondition(), BernoulliTrialCellCycleModel::ReadyToDivide(), UniformCellCycleModel::SetCellCycleDuration(), UniformG1GenerationalCellCycleModel::SetG1Duration(), PottsBasedCellPopulation< DIM >::UpdateCellLocations(), CaBasedCellPopulation< DIM >::UpdateCellLocations(), SimpleOxygenBasedCellCycleModel::UpdateHypoxicDuration(), and CaBasedCellPopulation< DIM >::WriteVtkResultsToFile().

void RandomNumberGenerator::Reseed ( unsigned seed )

Reseed the random number generator.

Parameters

seed	the new seed

Definition at line 153 of file RandomNumberGenerator.cpp.

References mGenerateStandardNormal, mGenerateUnitReal, and mMersenneTwisterGenerator.

Referenced by CellCycleTimesGenerator::GenerateCellCycleTimeSequence(), and AbstractCellBasedTestSuite::setUp().

template<class Archive >

void RandomNumberGenerator::save	(	Archive &	archive,
		const unsigned int	version
	)		const

inlineprivate

Save the RandomNumberGenerator and its member variables.

Note: do not serialize this singleton directly. Instead, serialize the object returned by GetSerializationWrapper().

Parameters

archive	the archive
version	the current version of this class

Definition at line 112 of file RandomNumberGenerator.hpp.

References mGenerateStandardNormal, and mMersenneTwisterGenerator.

template<class T >

void RandomNumberGenerator::Shuffle ( std::vector< boost::shared_ptr< T > > & rValues )

inline

Produce a permutation of the integers and non-empty std::vector, using the Knuth-algorithm (also called the Fisher-Yates algorithm), a linear time unbiased method. The values are shuffled in place.

Parameters

rValues the initial values and the output permutation of shuffled values. Must be non-empty

Definition at line 214 of file RandomNumberGenerator.hpp.

References Instance(), and randMod().

Referenced by TetrahedralMesh< ELEMENT_DIM, SPACE_DIM >::PermuteNodes(), PottsBasedCellPopulation< DIM >::UpdateCellLocations(), and CaBasedCellPopulation< DIM >::UpdateCellLocations().

void RandomNumberGenerator::Shuffle	(	unsigned	num,
		std::vector< unsigned > &	rValues
	)

Produce a permutation of the integers 0,1,..,num-1, using the Knuth-algorithm (also called the Fisher-Yates algorithm), a linear time unbiased method. The shuffled values are returned in rValues, which doesn't need to be correctly-sized when passed in.

Parameters