swarm-modeling
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Swarm Modelling] Project on GRID-mediated access to agent-based models


From: Nick Gotts
Subject: [Swarm Modelling] Project on GRID-mediated access to agent-based models
Date: Tue, 18 Feb 2003 17:54:13 +0000

Apologies if you receive multiple copies of the message below. Since FEARLUS is 
written using the Swarm libraries, I thought members of this list might well be 
interested.

Regards,
Nick Gotts

Dear Colleagues,

As  you are  no doubt  aware,  one of  the key  problems for  computer
modelling in  academic research is  that of repeatability.   For novel
approaches  such as agent-based  modelling in  particular, researchers
need  to be able  to examine,  experiment with,  and use  each others'
models,  (and to  take the  time to  do  so) if  we are  to go  beyond
producing and describing interesting models, to their effective use in
social science  and policy contexts. We consider  that emerging "Grid"
related            technologies            (http://www.gridforum.org/,
http://www.semanticgrid.org/)  have  the potential  to  support a  new
level of collaborative/competitive interactions between research teams
in agent-based modelling  and related areas, and write  to invite your
participation in a proposed  pilot project, exploring the potential of
the  Grid to  contribute  to e-Social  Science.  This is  a matter  of
urgency, as we have an application deadline of 26 February to meet.

In collaboration with colleagues at the University of Aberdeen, we are
writing  a proposal  to a  UK funding  body, for  a 12-18  month pilot
project that would involve making the FEARLUS model of land-use change
(http://www.macaulay.ac.uk/fearlus/) available  via Grid technologies,
specifically  the  emerging Open  Grid  Services Architecture  (OGSA),
which aims to integrate Grid and Web service concepts and technologies
(http://www.globus.org/ogsa/).  This would allow  us to make our model
available as a  service for other researchers to  use, with the effect
that  they   could  run   large-scale  experiments  with   FEARLUS  on
high-performance computing  facilities without having  to download and
compile  it.  The  pilot  project is  being  designed with  a view  to
forming an  EU FP6 Integrated Project  to do the same  with many other
models.

The  proposed work  will include:  (1)  describing FEARLUS  as a  Grid
service, so that  remote users can discover and  invoke it, and obtain
interpretable results; (2)  maintaining histories of interactions with
users,  allowing  experiments  to   be  replayed  and  results  to  be
aggregated; (3) managing the Grid  service, to control user access and
allocation of resources.   It would assist us greatly  in this initial
project if we  could identify a few colleagues willing  to be named in
our pilot project proposal, as  agreeing to help to assess the planned
Grid deployment of FEARLUS, in one of two possible roles:

(a) In a  research role, either  comparing results from your own model
with those from FEARLUS, checking results in FEARLUS publications, or
exploring FEARLUS's capabilities and limitations as a stage in the
process of designing a new model.

(b)  In a  teaching role,  using   FEARLUS as  part of  an appropriate
lecture course (FEARLUS has already been used in teaching at
postgraduate level).

Agreeing to be named in  the proposal would not  constitute or imply a
legal  or  contractual obligation to   take   part in  the  assessment
process, but of course we would want to name anyone who did not expect
to do so. The requirements of doing so would be a negotiable amount of
time for the   work involved in (a)   or  (b) above,  and   to give us
feedback from that  work (perhaps by  answering  a questionnaire); and
possibly a willingness to install a thin client to access FEARLUS over
the Grid.

Those  willing  to  be named will   (if the   application succeeds) be
consulted about  the facilities to be  made available, to maximise the
benefit to  their own teaching and/or  research. We  believe they will
benefit from  our  experience  of  planning and undertaking   the Grid
deployment of an  agent-based   model,  and  will be in   an  improved
position   to  join in  potential   Grid-related  and e-Social Science
projects - ours and others'. Specifically, we would hope they would be
interested in joining us in the follow-up proposal to the EU. Although
it is unlikely   we would   be able to   pay   you for  your time   in
participating in  this exercise, any other  expenses incurred would be
reimbursed.

If you  are interested  in getting involved,  please could you  let us
know as  soon as possible (but  no later than 24  February) which role
you would be prepared to take  (or indeed if there is another role you
can think  of that  you would prefer).  Naturally we will  not include
your  name  without  explicit  permission,  and are  ready  to  answer
questions to the best of our ability.

Yours,
        Nick Gotts and Gary Polhill

Nick Gotts
Macaulay Land Use Research Institute
Aberdeen AB15 8QH
Scotland, UK

Tel: +44 (0)1224 498200 ext. 2229
Fax: +44 (0)1224 311556
e-mail: address@hidden
http://www.macaulay.ac.uk/fearlus
  ---------------------------------------------------------------------
   Swarm-Modelling has migrated to address@hidden, please use this
   address in future postings to this list.


reply via email to

[Prev in Thread] Current Thread [Next in Thread]