 |
Author |
Topic: Agent-based modelling (Read 1174 times) |
|
jkeirstead
|
Agent-based modelling
« on: Jan 10th, 2004, 12:44pm » |
|
Specifically - does anyone know if Processing can accommodate the simultaneous action of multiple agents? The setup/loop syntax is a bit different from the hierarchial SWARM method (http://www.swarm.org).
For example, in the cellular automation examples all the bugs move at different times i.e. one moves, then another, and so on through the collection. Whereas in SWARM, all the bugs would move at the same time (well "time" in the model anyway).
So yeah, just wondering if anyone has tried agent-based modelling with Processing and how they've gone?
|
|
|
|
|
TomC
|
Re: Agent-based modelling
« Reply #1 on: Jan 10th, 2004, 1:05pm » |
|
If you want all your bugs to move at the same time, you need to keep a copy of the old state, and do all the calculations based on that, putting the results in a new state. That way the first bug and the last bug act on the same information.
You might want to take a look at MultiProcessing on the Model-based Design page at Yale - http://www.architecture.yale.edu/872a/
I don't know if it's going to be released to the public, but the source code is up... maybe Simon, or somebody else, could fill us in on if it's possible to use it with the current processing release?
|
| « Last Edit: Jan 10th, 2004, 1:05pm by TomC » |
|
|
|
|
|
