Friam

Fwd: [Repast-interest] The "Art" of Modeling

Classic

List

Threaded

3 messages Options

Friam mailing list

Fwd: [Repast-interest] The "Art" of Modeling

Thru sheer coincidence I was crafting the request below while our
discussion on the "art" of simulation spun up. I sent it to three
groups: the RePast and NetLogo modelers, and my SFI BusNet ValueNetwork
team.

Boy, I got a GREAT set of responses and they are still rolling in!
I'll stitch them all together into a web page or two shortly. Needless
to say, our FRAM discussion stitches in to it as well!

Owen

Begin forwarded message:

> From: Owen Densmore <[hidden email]>
> Date: Fri Feb 14, 2003 11:16:34 PM America/Denver
> To: [hidden email],
> [hidden email]
> Subject: [Repast-interest] The "Art" of Modeling
>
> I'm looking into a consulting job modeling a particular market
> (Automotive, for example). The idea is to create a model that is
> useful for "decision support", asking "what if" sorts of questions.
>
> The model would include three types/sets of agents: Market Segments,
> Primary Competitors, and Distribution Channels. Market Segments are
> sets of customer agents motivated by Price, Brand, Loyalty and other
> fairly well quantized parameters. Brand for example is simply market
> share percentage, while Loyalty would be a probability for change.
> There would be financial elements such as company capitol, assets,
> debt etc.
>
> The questions asked of the model would be based on strategy: following
> strategy X, would would our return on investment be.
>
> My concern is that I'm comfortable with the "plumbing" of modeling ..
> RePast/NetLogo, Java, and so on, but I have very little detailed
> knowledge of the "art" of modeling deciding on the right set of model
> parameters, agents, and interactions and how much detail is needed for
> them.
>
> So my question is: are there good books or papers discussing the "art"
> side of modeling? Details on the modeling of markets and so on?
>
> Owen Densmore 451 Camino Don Miguel Santa Fe, NM 87505
> Work: 505-983-6305 Cell: 505-570-0168 Home: 505-988-3787
> [hidden email] http://backspaces.net/
>
>
>
> -------------------------------------------------------
> This SF.NET email is sponsored by: FREE SSL Guide from Thawte
> are you planning your Web Server Security? Click here to get a FREE
> Thawte SSL guide and find the answers to all your SSL security issues.
> http://ads.sourceforge.net/cgi-bin/redirect.pl?thaw0026en
> _______________________________________________
> Repast-interest mailing list
> [hidden email]
> https://lists.sourceforge.net/lists/listinfo/repast-interest
>

Owen Densmore 451 Camino Don Miguel Santa Fe, NM 87505
Work: 505-983-6305 Cell: 505-570-0168 Home: 505-988-3787
[hidden email] http://backspaces.net/

Friam mailing list

Fwd: [Repast-interest] The "Art" of Modeling

I'm not familiar with modeling, but brand as I understand it is not
synonymous with market share. I would think you would want some other
parameter(s) to model brand..

Joe Spinden

-----Original Message-----
From: [hidden email] [mailto:[hidden email]] On Behalf Of
Owen Densmore
Sent: Sunday, February 16, 2003 12:04 PM
To: [hidden email]
Subject: [FRIAM] Fwd: [Repast-interest] The "Art" of Modeling

Thru sheer coincidence I was crafting the request below while our
discussion on the "art" of simulation spun up. I sent it to three
groups: the RePast and NetLogo modelers, and my SFI BusNet ValueNetwork
team.

Boy, I got a GREAT set of responses and they are still rolling in!
I'll stitch them all together into a web page or two shortly. Needless
to say, our FRAM discussion stitches in to it as well!

Owen

Begin forwarded message:

Owen Densmore 451 Camino Don Miguel Santa Fe, NM 87505
Work: 505-983-6305 Cell: 505-570-0168 Home: 505-988-3787
[hidden email] http://backspaces.net/

=========================================================
FRIAM Complexity Coffee listserv
Meets Fridays 9AM @ Museum Hill Cafe
Archives, unsubscribe, etc.:
http://www.redfish.com/mailman/listinfo/friam_redfish.com

Friam mailing list

Fwd: [Repast-interest] The "Art" of Modeling

In reply to this post by Friam mailing list

One element of the craft is that models are situated
within a contract. We know from our experience that
the scopes of research contracts are difficult to mange,
the people you talk to tend to change from phase to phase,
and that your access to data and the quality of the data
you do get may change as the contract progresses. Even
better, as you begin to model at the beginning of the
contract, you probably don't know how all this change
will manifest.

This is not at all suprising, given that the research
effort itself is (perhaps subtly) part of the complex
system under study, or, at the very least, is part
of an emergent system of how research is done within
or outsourced by the client. If the client is asking
about strategies or decision support systems upon which
the clients' company will be bet, then there will be
extra burdens of *proof* on the simulator.

It seems, then, that simulation craftsperople require
tools that enable their models to evolve as their
contractual engagements progress. Further, if the
client is hiring you to tell them something they
don't know, the tools should have a good auditing
facility to keep track of the simulation results
vis-a-vis the model versions.

To summarize, we need methods and tools that:
- facilitate ease of fundamental model changes
- facilitate ease of model augmentation
- enable auditing of model changes
- enable auditing of matches between model
variations and simulations and results.

Since this all sounds like tremendous fun, we also need
the methods and tools to be easy to use and unobtrusive.

Recently (well, for awhile now), I've been playing with
the notion of annotated models. Annotations take the form
of statements about resources. Annotations themselves are
resources that may be annotated. Resources (actual ones
in the client environment or annotations or sets thereof) are
defined as "authorities", so the shape of any given model is
the set of authorities that are "believed" by the simulation
environment (or groups of researchers) at the time. All resources
are represented as URI's, so an XML representation is somewhat
"natural". To evade the problems with modifying DTDs and XML
Schema, I've been using RDF Schema (http://www.w3c.org/rdf)
and looking at some of the associated ontology work at the W3C.
The Jena toolkit from HPLabs is a nice Java implementation of
an RDF modeler (http://www.hpl.hp.com/semweb).

This gives some purchase on the above criteria, though the jury's
still out on how this will work out in practice.

Carl

-----Original Message-----
From: [hidden email] [mailto:[hidden email]]On Behalf
Of Owen Densmore
Sent: Sunday, February 16, 2003 11:04 AM
To: [hidden email]
Subject: [FRIAM] Fwd: [Repast-interest] The "Art" of Modeling

Thru sheer coincidence I was crafting the request below while our
discussion on the "art" of simulation spun up. I sent it to three
groups: the RePast and NetLogo modelers, and my SFI BusNet ValueNetwork
team.

Boy, I got a GREAT set of responses and they are still rolling in!
I'll stitch them all together into a web page or two shortly. Needless
to say, our FRAM discussion stitches in to it as well!

Owen

Begin forwarded message: