Pseudorandom Numbers

Using a Linear Congruential Generator

RandomVariates.gen_prn — Function

gen_prn()

Generate a pseudorandom number.

Notes

Uses a linear congruential generator (LCG) with POSIX parameters:

$X_n = 25214903917 X_{n-1} + 11 \quad mod \quad 2^{48}$

Examples

julia> gen_prn()
 156750217634815

julia> gen_prn()
 63914890472862

source

Using a Tausworthe Generator

RandomVariates.tausworthe_rng — Function

tausworthe_rng(shape::Int=1; r::Int=3, q::Int=128)

Generate a shape element array of random variables from a standard Uniform(0,1) distribution using a Tausworthe RNG.

Notes

Implementation:

$B_i = B_{i-r} \quad XOR \quad B_{i-q}$

Examples

julia> U = tausworthe_rng(1)
1-element Vector{Float64}:
 0.5462285033427179

julia> U = tausworthe_rng((2,2))
2×2 Matrix{Float64}:
 0.782613  0.365878
 0.176636  0.0413817

References

Shu Tezuka and Pierre L'Ecuyer. 1991. Efficient and portable combined Tausworthe random number generators. ACM Trans. Model. Comput. Simul. 1, 2 (April 1991), 99–112. DOI:https://doi.org/10.1145/116890.116892

Law, A. Simulation modeling and analysis, 5th Ed. McGraw Hill Education, Tuscon, 2013.

source