# File rng.hpp¶

Class for holding different random number generators.

Author
Oliver W. Laslett

class Rng
#include <rng.hpp>

Abstract class for random number generators.

Subclassed by RngArray, RngMtDownsample, RngMtNorm

Public Functions

virtual double get() = 0

Get a single random number.

Return
a single draw from the random number generator

class RngMtNorm
#include <rng.hpp>

Uses Mersenne Twister to generate normally distributed values.

Random numbers have a mean of zero and a user specified standard deviation.

Inherits from Rng

Public Functions

RngMtNorm(const unsigned long int seed, const double std)

Default constructor for RngMtNorm.

Parameters
• seed: seed for random number generator
• std: the standard deviation of the normally distributed numbers

double get()

Draw a single normally distributed value from the RNG.

Return
single normally distributed number

Private Members

std::mt19937_64 generator

A Mersenne twister generator instance.

std::normal_distribution<double> dist

A normal distribution instance.

class RngMtDownsample
#include <rng.hpp>

Generate normally distributed values with downsampling.

Uses the Mersenne Twister to generate normally distributed random numbers. Down-samples the stream of random numbers by summing consecutive draws along each dimension. Function is usually used for generating coarse Wiener processes

Inherits from Rng

Public Functions

RngMtDownsample(const unsigned long int seed, const double std, const size_t dim, const size_t down_factor)

Default constructor for RngMtDownsample.

double get()

Get a single downsampled value from the random number generator.

Private Functions

void downsample_draw()

Private Members

std::mt19937_64 generator

A Mersenne Twister generator instance.

std::normal_distribution<double> dist

A normal distribution instance.

size_t current_dim

Stores the current state of the output dimenstion.

std::vector<double> store

Stores consecutive random numbers.

const size_t D

The number of dimensions required.

const size_t F

The number of consecutive random numbers to downsample.

class RngArray
#include <rng.hpp>

Provides an Rng interface to a predefined array of numbers.

Inherits from Rng

Public Functions

RngArray(const double *arr, size_t arr_length, size_t stride = 1)

Default constructor for RngArray.

Parameters
• _arr: a predefined array of random numbers
• _arr_length: length of the predefined array
• _stride: the number of consecutive elements to stride for each call to .get()

double get()

Get the next (possibly stridden) value from the array.

The first call will always return the value at the 0th index. Each subsequent call will stride the array (default value is 1) and return that value. A call that extends beyond the end of the array will result in an error.

Exceptions
• std::out_of_range: when a call attempts to get a value beyond the maximum length of the predefined array.

Private Members

unsigned int i =0

Internal state.

const size_t max

Maximum number of draws available.

const double *arr

Pointer to the predefined array of numbers.

const size_t stride

Number of values to stride for each draw.