Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!seismo!columbia!rutgers!princeton!allegra!ulysses!unc!mcnc!ecsvax!hes
From: hes@ecsvax.UUCP (Henry Schaffer)
Newsgroups: sci.math
Subject: Re: Need formula for Normal Distribution Simulation
Message-ID: <2198@ecsvax.UUCP>
Date: Mon, 3-Nov-86 17:17:48 EST
Article-I.D.: ecsvax.2198
Posted: Mon Nov  3 17:17:48 1986
Date-Received: Tue, 4-Nov-86 10:01:54 EST
References: <146@helm.UUCP> <5270001@hpfcpf.HP.COM> <401@entropy.UUCP>
Organization: NC State Univ.
Lines: 14
Summary: A simpler way to generate normal deviates.


Take the sum of 12 random numbers (the kind that are usually produced,
i.e., a uniform distribution over the 0,1 interval) and subtract the
number 6.0.  This gives a good approximation to a random deviate with
mean = 0 and standard deviation or variance = 1.

Then multiply it by the desired standard deviation and add in the
desired mean (do it in that order!) and you've got it.

This is a surprisingly good method which can be quite fast if the
uniform random generator used gives more than one random number per
call -- otherwise the time is dominated by subroutine linkage business.

--henry schaffer  n c state univ