There is a long standing issue with my {dqrng} package: weighted sampling. Since implementing fast un-weighted sampling methods quite some time ago, I have now started looking into possibilities for weighted sampling.
The issue contains a reference to a blog post that is by now only available via the wayback machine. This blog post shows a stochastic acceptance method suggested by Lipowski and Lipowska (2012) (also at https://arxiv.org/abs/1109.3627), which appears very promising.
Today dqrng version 0.3.0 made it unto CRAN and is now propagating to the mirrors.
This release contains a breaking change: The initial state of dqrng’s RNG is based on R’s RNG, which used to advance R’s RNG state. The implementation has been changed to preserve R’s RNG state, which is less surprising but can change the outcome of current scripts. (#44 fixing #43)
In addition, the generation of uniform random numbers now takes a short-cut for min == max and throws an error for min > max (#34 fixing #33)
I have blogged about dqrng before, but I forgot to write about the v0.2.0 release, even though that brought a lot of changes including the fast sampling methods to CRAN:
Add R side support for selecting multiple streams for parallel usage. Implement long_jump() for Xo(ro)shiro as alternative to jump() providing fewer streams with longer period. Handle R’s RNG scope properly during initialisation. New functions dqsample and dqsample.int using an unbiased sampling algorithm.