Dear Michael
That's great - I wasn't aware of these two sites which are both really
helpful and I'll follow them through. The issue on GRASS's
scriptability is something of a double-edged sword in that it makes it
much simpler to develop models in various languages and call them inside
the script, but then there often seems to be a performance hit in terms
of CPU or file IO. The permutation testing that goes with the
sensitivity analysis of many ABMs probably accentuates this issue.
Perhaps linkage through the new GRASS Extension Manager might speed up
processing a little, but I suspect it's more a case of revisiting some
of my coding, to make it a little less clumsy! Judging by the number of
pdfs on agent-based models on your site, I've a lot of reading to catch
up on.
What's the relationship between the openabm list and the grass-abm one
hosted at osgeo which I've just stumbled across? The activity on the
latter does, unfortunately, seem fairly low.
Best wishes
Roy
On Thu, 2008-02-14 at 11:20 -0500, grass-user-request@lists.osgeo.org
wrote:
------------------------------
Message: 5
Date: Thu, 14 Feb 2008 08:16:51 -0700
From: Michael Barton <michael.barton@asu.edu>
Subject: Re: [GRASS-user] Agent-based modelling and GRASS
To: grass-user@lists.osgeo.org
Message-ID: <1DE57E75-DADB-4807-AC00-1C7C004E9B54@asu.edu>
Content-Type: text/plain; charset=US-ASCII; delsp=yes; format=flowedOn Feb 14, 2008, at 2:53 AM, grass-user-request@lists.osgeo.org wrote:
> Date: Thu, 14 Feb 2008 09:38:54 +0000
> From: Roy Sanderson <r.a.sanderson@newcastle.ac.uk>
> Subject: [GRASS-user] Agent-based modelling and GRASS
> To: grass-user@lists.osgeo.org
> Message-ID: <1202981934.28165.239.camel@clarinet>
> Content-Type: text/plain
>
> Hello
>
> Does anyone know the current state of play with agent-based modelling
> and GRASS? About four years ago there seemed to be considerable
> activity on this front, with for example, Jim Westervelt's work on
> integrating Swarm models into GRASS for cowbird predictions (described
> in R. Gimblett's text on GIS and ABM). A couple of years ago I
> developed a simple discrete-event model of bumblebee foraging,
> using C++
> and GRASS shell scripts, although I will admit that as communication
> between the model and the GIS was primarily via text files that the
> disk
> I/O activity caused huge processing time problems.
>
> Then there has been the approach of using R within GRASS, and given
> that
> R has its own library for individual-based ecological models (via the
> simecol library) this may provide a more productive route for
> developing
> agent-based models operating within a GIS environment. However, I
> have
> the impression that the technology is going through some flux, and
> wonder what users consider to be the best approach?
>
> Best wishes
> Roy
>
> --
> Roy Sanderson
> Institute for Research on Environment and Sustainability (IRES)
> Devonshire Building
> Newcastle University
> Newcastle upon Tyne NE1 7RU
> r.a.sanderson@newcastle.ac.uk
> 0191 246 4835
>Roy,
One of my research projects is coupling a Java-based ABM (DEVSJAVA)
with GRASS quite successfully. Rather than trying to write an ABM in
GRASS, we are using GRASS' scriptability to call it as needed by the
ABM. The ABM manages the agents and their behavior; GRASS manages the
landscape and surface process dynamics. Here is a link to our website
<http://medland.asu.edu>. I'm happy to talk more with you on or off
list. Also, you might want to take a look at the new OpenABM
Consortium site <http://www.openabm.org>. If there is sufficient
interest, I'd be happy to start a forum on ABM and GIS/GRASS there.Michael
____________________
C. Michael Barton, Professor of Anthropology
Director of Graduate Studies
School of Human Evolution & Social Change
Center for Social Dynamics & Complexity
Arizona State UniversityPhone: 480-965-6262
Fax: 480-965-7671
www: <www.public.asu.edu/~cmbarton>------------------------------