I'd like to produce pseudo-random numbers with different distributions for a Monte Carlo simulation.

I've got the poisson distribution working nicely with an algorithm from Knuth. I'm having trouble getting a nice easy and fast algorithm for a power distribution. The gamma distribution should do, but the article in wikipaedia gives an algorithm, but remarks that it's not a good one, without providing a link to a better one.


Is there a good, fast algorithm for a gamma distribution?

  • $\begingroup$ Luc Devroye has often commented on the challenges of producing gamma variates and is purported to maintain an open offer of fine Belgian beer to the discoverer of a "one-line algorithm". If you are interested in random-variate generation and don't know the name Luc Devroye, you need to. See also this small appetizer. $\endgroup$
    – cardinal
    Nov 8, 2014 at 14:02

1 Answer 1


The difficulty mentioned in Wikipedia refers to gamma distributions with small shape parameter; this has been addressed in arXiv:1302.1884:

The gamma distribution with small shape parameter can be difficult to characterize. For this reason, standard algorithms for sampling from such a distribution are often unsatisfactory. In this paper, we first obtain a limiting distribution for a suitably normalized gamma distribution when the shape parameter tends to zero. Then this limiting distribution provides insight to the construction of a new, simple, and highly efficient acceptance--rejection algorithm. Pseudo-code and an R implementation for this new sampling algorithm are provided.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.