[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: guassian distribution generator
From: |
Rick Riolo |
Subject: |
Re: guassian distribution generator |
Date: |
Mon, 26 Aug 1996 07:26:19 -0400 (EDT) |
Manor,
thanks for pointing that out. (I guess I did miss that
in your previous message...)
I would suggest/request that you add an additional method:
-(float) getGuassianWithMean: (float) mean andSD: (float) sd;
(or something like that syntax)
so that one can easily/cleanly make many requests with
different means/stddevs (eg when each agent has its own
distribution).
thanks!
- r
Rick Riolo address@hidden
Program for Study of Complex Systems (PSCS)
1061 Randall Lab University of Michigan
Ann Arbor MI 48109-1120
http://pscs.physics.lsa.umich.edu/rlr-home.html
On Fri, 23 Aug 1996, Manor Askenazi wrote:
> Date: Fri, 23 Aug 96 12:26:22 MDT
> From: Manor Askenazi <address@hidden>
> To: address@hidden
> Cc: address@hidden
> Subject: Re: guassian distribution generator
>
> > Has anyone ported a guassian distribution generator (given some mean and
> > sd) to Swarm,
>
> Hi Rick!
>
> Did you miss this (from a previous message I sent)?
>
> ------------------------------------------------------------------------
>
> Also with respect to 'serious' scientific simulation (especially in the
> humanities :-) we have the Gaussian object:
>
> /************************** Gaussian *****************************/
>
> @interface Gaussian: SwarmObject {
> id uniform ;
> float mean, std ;
> int stored ;
> float stored_result ;
> }
>
> +create: aZone withMean: (float) aMean
> Std: (float) aStd
> andSeed: (unsigned) aSeed ;
>
> +create: aZone withSeed: (unsigned) aSeed ;
>
> -initSeed: (unsigned) aSeed ;
>
> -setMean: (float) theMean ;
> -setStd: (float) theStd ;
>
> -setUniform: aUniform ;
> -getUniform ;
>
> -(float) rFloat ;
>
> @end
>
> /****************************************************************/
>
>
> Note that unlike the other RNGs, Guassian is a Swarmobject and also has
> a very visible and straightforward way to change the seed (Nelson's original
> RNG's were meant to be distributable separately from Swarm so they are not
> created in the usual way, also, the way in which users change the seed is
> not intuitive enough - this may be 'corrected' by the next release).
>
> ------------------------------------------------------------------------
>
> I was expecting to release this along (with the other new stuff which I
> have reported on in that email).... Also, Sven Thommesen is working on a
> few additional Distributions as we speak!
>
> Regards,
>
> Manor.
>