YAI swarm: yet another intro

[address@hidden (Bohdan Durnota)]

Hi,

I am a newcomer to the Swarm user arena. My name is Bohdan
Durnota, and I am a lecturer in the department of software
development at Monash University in Melbourne, Australia.
Apart from interests in (object-oriented and formal) software
engineering, I have a considerable interest in the modelling
of complex systems, ranging from geological structures, to
biodiversity and financial asset markets.

I have had a quick look at som eof the intros and I am
impressed with the variety of interesting projects people from a
range of disciplines are working on.

>From the point of view of Swarm, my immediate interests lie in 4 
particular projects:

1. Intelligent traffic systems/CA traffic flow models.
My honors student last year worked on using a classifier system
within an agent to control traffic flow at intersections.
We have some ideas we want to experiment on -- on communities of 
such traffic light agents located at sites, and interacting and
evolving their control behaviors. In addition, we have been
running some two-dimensional CA models of traffic flow in which
there is adaptive route finding by individual drivers. We have
just begun implementing this model into Swarm.

2. Biodiversity modelling.
I have been doing some modelling work with a biologist colleague
of mine on biodiversity in tropical rainforests, especially with
reference to Australian ones. We are interested in what happens
over relatively long time periods (say Tertiary->Holocene), and
to co-dependencies amongst guilds and functional groups. It is a
spatially explicit model at the community level. At present,
I have worked on developing a formal object-oriented specification
of the assembly and other rules we have in mind, and we have done
some simulation work using both stochastic Petri nets and the
Linda coordination language with C. However, there are a number of 
*simpler* questions and formulations which one would like an
ECHO-like model to explore with. For example, the role of landscape
heterogeneities (like soil types) and light gaps as generators of
biodiversity. I would hope we will be able to use Swarm to explore
questions like these.

3. Strategic Interactions in asset markets.
Over the last half-year, an economic colleague, I and a Masters
student have been working on investigating the effect of strategic
interactions amongst traders in financial asset markets. We have
a Windows 95/NT software system built to help us run a range of
simulations. However, we have not concentrated on any sophisticated
learning nor evolution of these strategic interactions. Hopefully
using Swarm and an Echo/Gecko framework, we can start exploring
the evolution of these strategic interactions.

4. GIS agents
Randy Gimblett from Univ Arizona's School for Renewable Natural
Resources will be on sabbatical for a year
here in Melbourne, and we and some other colleagues will be working
on a project which will combine Swarm with a GIS system (probably 
ARC/INFO) and software developed by the Australian AI Institute
to encode intentional agents. The project will aim at exploring
conflicts amongst users of a reserved forest. I am sure we'll have
plenty of technical issues here to confront!

5. ?ALife models of the software development process?
I have only had tentative conversations with one of our students
on how we can formulate CAS-type models of how we actually develop
software under changing and uncertain requirements, and modify
our development processes along the way, ....

Also, I have a pedagogical interest in Swarm. I have been teaching
object-oriented modelling and simulation, and object-oriented
formal specification, and I have found these topics can combine
nicely so as to give computing students a flavor of the ideas
behind, and excitement in, work on CAS. Eventually, I would like to
be using Swarm or something similar in my teaching.

Seeing that most/all of us in this group are interested in CAS, let
me mention that there is a forthcoming conference on Complex Systems
to be held in Australia in July (abstracts due MArch 1), which you
might be interested in. Chris Langton will be our principal speaker.
If you are interested, then please browse the web page:

        http://life.csu.edu.au/complex/cs96no2.html

Anyway, I hope all of you have begun the New Year on a good
note.

Cheers, Bohdan Durnota

_________________________________________________________________

Bohdan Durnota                          PO Box 197
Department of Software Development      Caulfield East 3145
Monash University                       Australia

email:  address@hidden
http:   //www.sd.monash.edu.au/~bdurnota/bdurnota.html/
phone:  +61-3-99032041
fax:    +61-3-99031077
_________________________________________________________________