Hi,
Nigel's mail generated 18 answers, not 18 points
to be discussed. The easiest way to figure out is to get to
Simsoc archive and start from Nigel's original mail of
August 10th and follow the track by clicking
on the "next in topic" button.
As earlier mentioned, the answers are very condensed
and a summary would be difficult. The main point that appears
from several contributions is that some kind of density
rather than the total number of individuals is relevant.
2009/2/2 Nigel Gilbert <[log in to unmask]>:
> How about it?
>
> Nigel
>
>
> ________________________________________________________________
> Professor Nigel Gilbert, ScD, FREng, AcSS, Professor of Sociology,
> University of Surrey, Guildford GU2 7XH, UK. +44 (0)1483 689173
>
> ------ Forwarded Message
> From: Dan Olner <[log in to unmask]>
> Date: Mon, 2 Feb 2009 09:33:19 -0000
> To: Nigel Gilbert <[log in to unmask]>
> Subject: RE: [SIMSOC] JASSS special section on whether population size
> matters
>
> Greetings
>
> This sounds great. Quick question: could you possibly post a list of the
> eighteen points in summary form?
>
> Thanks
>
> Dan
> University of Leeds
> School of Geography
>
> ________________________________
> From: News and discussion about computer simulation in the social sciences
> on behalf of Nigel Gilbert
> Sent: Fri 30/01/2009 21:47
> To: [log in to unmask]
> Subject: [SIMSOC] JASSS special section on whether population size matters
>
> On August 10, 2008 the following issue was raised by Nigel Gilbert on the
> SimSoc discussion list <http://www.jiscmail.ac.uk/archives/simsoc.html>.
>
> "In what circumstances might it be necessary to implement agent-based models
> of thousands or millions of agents, rather than the more typical tens or
> hundreds? Do you know of any examples where scaling up the model (in terms
> of N, the number of agents) gives qualitatively different results? Do you
> know of any literature that addresses this issue? I'd be very interested to
> hear of any suggestions."
>
> Some 18 suggestions were proposed, although because they were in discussion
> list format, they were perhaps too briefly explained to convince anyone but
> their authors of their relevance.
>
> Nigel's question is reminiscent of the way questions were posed by Learned
> Societies during the 18th and 19th centuries. Scientists would write papers
> to answer these questions, eventually in competition. This tradition is now
> nearly extinct in our society of professional scientists, but it might be
> adapted to an issue we are encountering in Social simulations: Many
> simulations have been made, but our contributions are difficult to integrate
> into coherent knowledge about the world, and maybe even about the
> "silicon-world". We do have summer schools, meetings and even reference
> books, but no manuals and standard courses in university curriculum.
>
> One possible way towards gradually building a consensual knowledge about the
> result of Social Simulations could be to answer fundamental questions such
> as the one raised by Nigel.
>
> I therefore propose the following:
>
> Contributions answering Nigel's question are requested as manuscripts for
> publication in JASSS <http://jasss.soc.surrey.ac.uk/>. The format of these
> contributions can vary, thanks to the flexibility of electronic publication,
> but they should be self-contained and explicitly describe the model(s) they
> refer to.
>
> At minimum, they should describe the different dynamical regimes observed
> when N, the number of agents, is varied, using, for example, regime diagrams
> plotting some variable with respect to N, or to some power of N, and showing
> that there are threshold effects where behaviour changes in a non-linear way
> with increasing N.
>
> Explicit illustrations with specific models should not prevent authors also
> elaborating on the crucial hypotheses responsible for the existence of
> several regimes, and relating these to more general principles. But
> contributions only based on conjectures such as 'when N increases, one
> should observe that ...' would not be eligible.
>
> A deadline of one year hence for such a set of papers would allow potential
> contributors to reflect on the topic and propose contributions to be
> refereed and published in a special section of JASSS. We might also consider
> publishing reactions to these papers, including part of the referees'
> comments and a short introduction by the editor(s).
>
> A formal call for contributions to this proposed special section of JASSS
> will follow, but meanwhile I would be interested in comments and reactions
> to this proposal and, especially, offers to contribute. This proposal has
> been formulated with Nigel's approval.
>
>
> Gerard Weisbuch
>
>
> ________________________________________________________________
> Professor Nigel Gilbert, ScD, FREng, AcSS, Professor of Sociology,
> University of Surrey, Guildford GU2 7XH, UK. +44 (0)1483 689173
>
>
> ------ End of Forwarded Message
>
--
GĂ©rard Weisbuch Laboratoire de Physique Statistique ENS
24 rue Lhomond 75005 Paris France
http://www.lps.ens.fr/~weisbuch
|