New Beta Tester -- Intro

[Martin J. Maiers]

Greetings SWARM folks!

My name is Martin Maiers; I'm a grad student in Computer (and Cognitive)
Science at the University of Minnesota.  

I work in the Advanced Biosciences Computing Center (ABCC) within the 
College of Biology where (among other things) I tend to the care and 
feeding of the College's UNIX boxes.

I am a new beta tester for SWARM as of the latest release.

I heard about SWARM at ALIFEIV and have been looking
forward to trying it out ever since.

I haven't been able to build SWARM yet due to problems with 
building gcc 2.7.2 under Solaris2.4, but I should have it running
any day now...

--

Here at the ABCC we do more Biochemistry than anything else:
  Databank Searching/Sequencing
  Protein Analysis/Folding/Structure Prediction
  Molecular Modeling
  Sequence Alignment

but my primary interest is in Ecology, particularly modeling social
networks.

--

The first project that I'd like to use SWARM for is a model
of social behavior (grooming networks) of Vervet Monkeys.

I am working with an Anthropologist (my father) to develop a model 
that can be used to explain how dominance hierarchies are represented and
maintained.  Most of the computer modeling in this are has been
done with static agents (usually with a pre-ordained rank).

I'd like to use SWARM to situate the agents spatially and endow
them dynamic memory of their interactions with each other and their
environment.  

Dominance hierarchies can be subtle and (may) depend on having a 
memory of past interactions with individuals which in turn requires
a local model of the "minds" of the other individuals in order to 
take into account their mental states (beliefs, desires, mood).

In terms of a simulation, this means each individual would maintain
a simple history of what happened the last time they interacted
with another individual.  Their decision to groom or not to groom,
or to fight or to be groomed depends on how they "feel" about that
person (summing up several weights in a network of fuzzy (aka Neural)
rules).

--

The inspiration for this research comes from a number of sources:

  Dan Dennett - for his model of Consciousness and Intentionality.
  Rodney Brooks - for his work on Situatedness and Subsumption.
  Lotfi Zadeh - for his computational model of fuzzy logic.
  Cheney & Seyfarth - for Vervet field work and behavior models.
  Jane Goodall - for a rich source of observational Chimpanzee data.

In fact, just last month the "Jane Goodall Center for Primate Research" 
was opened here at the University of Minnesota.  35 years of Chimpanzee
field data is being computerized, indexed and published electronically.

In the future I'd like to model Chimpanzees as well but for now
we are just working with Vervets.

--

The second project I'd like to try with SWARM is a model of 
Cell Wall formation.  I've been working with a couple colleagues 
of mine on data structures for modeling the interaction of 
Phosphlipid molecules but really have lacked a "simulation engine" 
to run it on.

Actually, what we really need is a way around the 3-D N-body problem...

I've looked at some statistical physics (Spin Glass) stuff but haven't
found a way to apply it directly to our problem.

--

Anyhow, thats a brief intro to this Beta tester's site and what 
I'm hoping to use SWARM for.

I'm sure I'll have a question or six about the actual program if 
and when I ever get gcc2.7.2 built....



Cheers,
Martin

-- 
Martin Maiers, System Administrator             address@hidden
Advanced Biosciences Computing Center           (612) 624-7471
247 Gortner Labs, 1479 Gortner Ave              (612) 625-9284 (HelpLine)
St. Paul, MN 55108-1022                         (612) 625-5780 (FAX)