From: <HIDDEN>
Path: news.mathworks.com!newsfeed-00.mathworks.com!webcrossing
Newsgroups: comp.soft-sys.matlab
Subject: Re: Pseudo random generator
Message-ID: <ef557ed.3@webcrossing.raydaftYaTP>
Date: Tue, 1 May 2007 15:55:41 -0400
References: <ef557ed.-1@webcrossing.raydaftYaTP>
Lines: 36
NNTP-Posting-Host: 216.208.252.5
MIME-Version: 1.0
Content-Type: text/plain; charset="ISO-8859-1"
Content-Transfer-Encoding: 8bit
Xref: news.mathworks.com comp.soft-sys.matlab:406288



All random numbers generated are by definition pseudo random. Whether
you need to use rand or randn depends on the application at hand.

  Walter Roberson wrote:
>
>
> In article <ef557ed.1@webcrossing.raydaftYaTP>,
> Alissa gotenberg <fete7a_85@hotmail.com> wrote:
>>THNX FOR YOUR REPLY , I KNEW BEFORE BOUT THE rand() & randn() ,
> but
>>the project specification said it must be pseudo random
generated
>>code ,
>>so i wanted to know if there is a difference between the
noramlly
>>distributed random or uniform distributed randoms & the pseudo
> random
>>.
>
> Both rand() and randn() return pseudo-random sequences.
>
>>
>>cause i guess it is a different random generation algorithm ,
>
> Both of them use the same internal random number generators, but
> randn() post-processes (sometimes requiring multiple values) to
> generate normally distributed numbers.
>
> In a recent thread (a week ago) 'random variable', Peter Perkins of
> The
> MathWorks posted a reference to the algorithms used in both cases:
>
> <http://www.mathworks.com/moler/random.pdf>
> --
> Prototypes are supertypes of their clones. -- maplesoft
>