[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Swarm futures (was Re: ABM/Toolkit futures [was: Swarm futures (was
From: |
gepr |
Subject: |
Re: Swarm futures (was Re: ABM/Toolkit futures [was: Swarm futures (was Re: GNUstep and MacOS X Port Effort)]) |
Date: |
Sat, 5 Oct 2002 14:34:58 -0700 |
(Moving this to Swarm-Modelling, which is where it really belongs.
Sorry for abusing the support list like that.)
Gulyas Laszlo writes:
> On Fri, 4 Oct 2002 address@hidden wrote:
> > BTW, I really do mean "Swarm futures" when I talk about this stuff.
>
> OK, I think this is really a terminology issue here...
> Notice that I used ABM (!) in my subject and that's what I meant.
> To me Swarm _is_ the implementation. True, it set the example
> and terminology for most agent-based modeling/simulation tools,
> but it's still just a particular realization of a method.
> That is, as I see it...
Hmmm. To me Swarm is the generic name for the type of modeling, which
is why "Swarm Intelligence" is appropriately named. There are
different kinds of "Agent-Based Modeling", some of which are very
Swarm-like and some of which are completely different.
When designing BDI agents, for example, many people use ABM toolkits.
None of those ABM toolkits are at all Swarm-like. Swarm Modeling is a
subset of ABM. And it's a pretty specific subset.
Marcus (I think) introduced the term "fine-grained agents" to indicate
the difference between Swarm Modeling and other ABM. True, the SDG's
implementation of Swarm, MAS, AScape, and RePast all have the
capability of doing non-Swarm ABMs. But, it's useful, especially when
having a conversation about methods, to distinguish between the two
types of modeling. (In fact, what gives the Swarm and its various
implementations a bit of power, in my opinion, is it's open-ness to
*integration* of all types of modeling, including other types of
ABM.... a kind of multi-modal, multi-scale, and multi-frequency
modeling.)
When talking about what the SDG should do next, it is important to
talk explicitly about Swarm futures and try to discover where Swarm
(as a modeling style) fits in with things like AgentSheets, Aglets,
and AUML, to name a few, as well as the Swarm-like tools.
Plus being sloppy in our language doesn't help refine the agenda. So,
I prefer the more precise term "Swarm" over the more ambiguous term
"ABM".
I also don't think it's relevant to restrict the conversation to
"toolkits", for similar reasons.
OK... enough nit-picking about terminology. [grin] But, in
self-defense, terminology can often make or break an effort. So, I'd
rather be clear and too wordy than vague and terse.
--
glen e. p. ropella =><= Hail Eris!
H: 831.335.4950 http://www.ropella.net/~gepr
M: 831.247.7901 http://www.tempusdictum.com
==================================
Swarm-Modelling is for discussion of Simulation and Modelling techniques
esp. using Swarm. For list administration needs (esp. [un]subscribing),
please send a message to <address@hidden> with "help" in the
body of the message.
==================================
- Re: Swarm futures (was Re: ABM/Toolkit futures [was: Swarm futures (was Re: GNUstep and MacOS X Port Effort)]),
gepr <=