Re: Instantiating multiple ModelSwarms

[Paul E Johnson]

If you use Unix, there is a program called drone which can run your
program over and over under many conditions and it keeps excellent
records on what has been done.  The only part you need to do is enable
the command line parameter processing and write a drone control file.

On windows, we have trouble getting Drone to work, because the required
library Expect won't compile.  So I wrote a perl program that can
replicate a program and sweep through parameters, keeping good records
on what's been done.  That's available

http://lark.cc.ukans.edu/~pauljohn/Swarm/MySwarmCode/replicator.pl

We have tested that on windows and unix, works fine.

I don't mean to say that Steve's approach won't work, but I spent a long
time trying to make things like that work and never got it.  As I
recall, the hangup is that you want a control panel for the multi-run
experiment, and in addition you need a control panel for the individual
run of the model, and making those two work together was a major
problem. You will see, if you dig into it.... 

Fred Wan wrote:
> 
> Thanks for your reactions. The reason the ExperSwarm model is important is
> that I need to parameterize _a lot_. Not only do I need to be able to test
> different environments for the Multi-Agent System (e.g., obstacles, food
> density), but I also need to be able change the design of the agent in a
> systematic way and to (meticulously) record the results. As far as I'm
> concerned there needs only to be only one swarm (maybe together with one
> guiswarm) in which the model runs. The problems I encountered arise from
> confusion about how the control flows are running from, to and between
> swarms, activities, schedules, etc. Besides the obvious technical
> intricacies, the whole conceptual model underlying Swarm is quite hard to
> understand (though it is growing).
> 
> All input is appreciated, so I'd like to receive the code for the Experiment
> Swarm.
> 
> Thanks,
> Fred Wan.
> 
> address@hidden
> 
>                   ==================================
>    Swarm-Support is for discussion of the technical details of the day
>    to day usage of Swarm.  For list administration needs (esp.
>    [un]subscribing), please send a message to <address@hidden>
>    with "help" in the body of the message.

-- 
Paul E. Johnson                       email: address@hidden
Dept. of Political Science            http://lark.cc.ukans.edu/~pauljohn
University of Kansas                  Office: (785) 864-9086
Lawrence, Kansas 66045                FAX: (785) 864-5700

                  ==================================
   Swarm-Support is for discussion of the technical details of the day
   to day usage of Swarm.  For list administration needs (esp.
   [un]subscribing), please send a message to <address@hidden>
   with "help" in the body of the message.