Re: Are there theorems for Agent Based Models

[Ken Cline]

(Sorry for the late response; I got a little behind on my
email.)

On Mon, 12 Jul 1999, Paul Johnson wrote:

> 1. Assume all agents survey the environment as it was after the last
> time step.  If agents are in time 2, for example, they look around at
> all other agents as they were at the end of time 1.  (Example:Heatbugs
> heat display is updated once per time period, all agents refer to that
> snapshot)  
>
> 2. Assume that all agent actions have "instantaneous effects" on each
> other.  If we are in time 2, each agent who observes is able to see all
> agent actions up to that instant.  If in time 2, a long list of agents
> is called upon to act (possibly in random order), each one is aware of
> the actions of the previous agents.  Agents who acted in step 2 before
> are seen and their changes are taken into account.

I'm not claiming to be a Heatbugs expert, but I think the
reason that model appears "synchronized" is because the bugs
have a simple interaction with their environment (eg they
don't observe other bugs directly) and what interaction they
do have (ie adding heat) must propagate before it affects
other agents.

What I mean is that (in Heatbugs) the agents do three basic
things, all their `step' method,  (1) choose next position,
(2) add heat at current location (3) move to next position.  
Because each bug does all three substeps before the next bug
has a chance to take its `step', this means that there does
exist the potential for a bug to take into account what
other bugs have done at the current time step. Its just that
in practice, they don't.

Well in fact, they do account for other bugs if the position
they initially selected is occupied, but this collision
resolution is unavoidable (no pun intended).  That is, even
if the bugs selected their next position in one method and
then moved in a second method, you would still have to use
some type of collision resolution.

Bugs could also gain "fore-knowledge" about others 
indirectly thru the Heatspace because they add heat to the
environment during their `step' method as well. However,
currently the bugs only query the heat at their exact
location and heat added by other bugs does not diffuse until
the end of the cycle (ie after all the bugs have stepped).

My point is just that while the display only updates once
per cycle, the agents are not technically referring to that
snapshot.  Of course this doesn't have much bearing on the
interesting discussion (that followed your message) about
synchronous and asynchronous modelling and theorems
pertaining to them, except to say that it can difficult
create a truely synchronous simulation.  No revelation
there, eh?

Ken.

_________________________________________________________
Ken Cline                             address@hidden
SAIC                                 VOICE (410) 571-0413
Annapolis, MD                          FAX (301) 261-8427









                  ==================================
   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.
                  ==================================