@saehrimnir/druidjs / Randomizer

Class: Randomizer

Defined in: util/randomizer.js:45

Seeded pseudo-random number generator.

Implements sfc32 (Small Fast Counting, Doty-Humphrey), a 128-bit-state counter-based generator that passes TestU01 BigCrush. It is preferred here over the more familiar Mersenne Twister, which needs a 2.5 KB state, is markedly slower to seed, and whose GF(2)-linear output fails BigCrush's matrix-rank and linear-complexity tests.

Every operation is 32-bit integer arithmetic (^ << >>> + and Math.imul), all of which ECMAScript specifies exactly. The stream is therefore identical on every engine for a given seed — unlike floating point transcendentals, which are only "implementation-approximated". See DR for what that does and does not guarantee about the output of an algorithm.

See

where sfc32 originates

Example

const R = new Randomizer(1212);
R.random;      // float in [0, 1)
R.random_int;  // uint32

Constructors

Constructor

new Randomizer(_seed?: number): Randomizer;

Defined in: util/randomizer.js:63

Parameters

Parameter	Type	Description
_seed?	number	The seed for the random number generator. If _seed == null then the actual time gets used as seed. Default is new Date().getTime()

Returns

Randomizer

Properties

Property	Type	Default value	Defined in
_a	number	0	util/randomizer.js:47
_b	number	0	util/randomizer.js:49
_c	number	0	util/randomizer.js:51
_d	number	0	util/randomizer.js:53
_seed	number	undefined	util/randomizer.js:55
_val	number | null	null	util/randomizer.js:57

Accessors

random

Get Signature

get random(): number;

Defined in: util/randomizer.js:95

Returns a float between 0 and 1.

Returns

number

• A random number between [0, 1]

---

random_int

Get Signature

get random_int(): number;

Defined in: util/randomizer.js:104

Returns an integer between 0 and MAX_INTEGER.

Returns

number

• A random integer.

---

seed

Get Signature

get seed(): number;

Defined in: util/randomizer.js:86

Returns the seed of the random number generator.

Returns

number

• The seed.

Set Signature

set seed(_seed: number): void;

Defined in: util/randomizer.js:69

Parameters

Parameter	Type
_seed	number

Returns

void

Methods

choice()

choice<T>(A: T[], n: number): T[];

Defined in: util/randomizer.js:157

Returns n samples drawn from A without replacement.

Uses a partial Fisher-Yates shuffle over a scratch index array, which runs in O(n): A itself is never touched, and exactly n random values are consumed. Removing each picked index with Array.prototype.splice instead would shift the tail on every draw and make the call O(n²) — measurably worse from a few hundred elements upward, and 27× slower at n = 4000.

Type Parameters

Type Parameter	Description
T	Returns samples from an input Matrix or Array.

Parameters

Parameter	Type	Description
A	T[]	The input Matrix or Array.
n	number	The number of samples.

Returns

T[]

A random selection form A of n samples.

---

gauss_random()

gauss_random(): number;

Defined in: util/randomizer.js:127

Returns a normally distributed number with mean 0 and standard deviation 1.

Uses the Marsaglia polar method, which yields two values per iteration; the spare is cached and returned by the following call.

Returns

number

A standard normal variate.

---

choice()

static choice<T>(
   A: T[], 
   n: number, 
   seed?: number): T[];

Defined in: util/randomizer.js:189

Type Parameters

Type Parameter	Description
T	Returns samples from an input Matrix or Array.

Parameters

Parameter	Type	Default value	Description
A	T[]	undefined	The input Matrix or Array.
n	number	undefined	The number of samples.
seed	number	1212	The seed for the random number generator.

Returns

T[]

• A random selection form A of n samples.