Newsgroups: comp.ai.neural-nets
Path: cantaloupe.srv.cs.cmu.edu!rochester!cornellcs!newsstand.cit.cornell.edu!news.kei.com!simtel!news00.sunet.se!sunic!psinntp!psinntp!psinntp!psinntp!megatest!news
From: Dave Jones <djones>
Subject: (no subject)
Content-Type: text/plain; charset=us-ascii
Message-ID: <DIBp6v.9qG@Megatest.COM>
Sender: news@Megatest.COM (News Admin)
Nntp-Posting-Host: pluto
Content-Transfer-Encoding: 7bit
Organization: Megatest Corporation
Mime-Version: 1.0
Date: Mon, 20 Nov 1995 03:47:18 GMT
X-Mailer: Mozilla 1.1N (X11; I; SunOS 5.4 sun4m)
X-Url: news:comp.ai.neural-nets
Lines: 15

I've got an idea for a variation on simulated annealing that I want to
try out.

I need an algorithm that takes an arbitrary standard deviation as input,
and returns random numbers with a distribution that is normal (a bell
curve), has a mean of zero, and has the given standard deviation.

How do I do that? Sounds easy. Should be able to use a uniform random
number generator somehow, but I haven't figured out a way to do it without
making far too many calls to the uniform random number generator. One should
be enough.


          Dave

