home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!wupost!zaphod.mps.ohio-state.edu!sample.eng.ohio-state.edu!purdue!yuma!lamar!mglacy
- From: mglacy@lamar.ColoState.EDU
- Newsgroups: sci.math.stat
- Subject: Re: random number generator
- Summary: What about the L'Ecuyer algorithm?
- Keywords: pseudo-random, uniform, normal
- Message-ID: <Jul27.161826.26402@yuma.ACNS.ColoState.EDU>
- Date: 27 Jul 92 16:18:26 GMT
- Sender: news@yuma.ACNS.ColoState.EDU (News Account)
- Organization: Colorado State University, Fort Collins, CO 80523
- Lines: 32
-
-
- I've been using a random number generator that I translated to Pascal
- from some FORTRAN code available on STATLIB. Here's the
- prologue comment to the subroutine as obtained from STATLIB.
- ---------------------------------------
- Generate uniformly distributed random numbers using the 32-bit
- generator from figure 3 of:
- L'Ecuyer, P. Efficient and portable combined random number
- generators, C.A.C.M., vol. 31, 742-749 & 774-?, June 1988.
-
- The cycle length is claimed to be 2.30584E+18
- ---------------------------------------
-
- It uses two 32 bit seeds--Is this likely to make it dependable
- over long sequences, as claimed?
-
- Presuming this is something you recognize, I'd appreciate any
- comments on its quality from you RNG specialists out there.
- I've been using it as a source of uniform variates to supply a Gaussian
- deviate generator (double precision) using the polar method (as described
- in Numerical Recipies). It's all part of some Monte Carlo investigations
- of sampling distributions, with repetitions in the range of say
- 10,000 to 100,000 samples of N = 15.
-
-
- Thanks.
-
-
- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
- Mike Lacy Dept. of Sociology Colo. State Univ. Fort Collins CO 80523
- Opinions and truth-claims offered above are mine, not my employer's.
- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
-