Re: [Swarm-Support] Heatbug code question

[Steve Railsback]

Paul Johnson wrote:

> I agree with the spirit of your question. It can be done either way.  I 
> think we should consider re-designing Heatbugs to have the agents make 
> that calculation. It would be more in tune with the agent-based philosophy.
>
> It does not happen in this model, but there is a potential benefit to 
> having the HeatSpace make the calculation.  There is a major slow-down 
> if each agent searches each cell finding what's best.  In some models, 
> you can have the space calculate various values and give them to the 
> agents and if you write the model "just so", it avoids repeated 
> calculations.  That's the idea behind this code:

Ah, that's the difference between heatbugs and our fish model: for our 
fish, the answer they get (the choice of which cell is best) depends on 
the state of the fish as well as the state of the cell. So each fish 
must evaluate each potential destination cell anyway, and the slow-down 
Paul mentions is unavoidable.

So in heatbugs there is a tradeoff between what's intuitive and 
'realistic', vs. what executes fast.

Steve

--
Lang Railsback & Associates
250 California Ave.
Arcata CA 95521
707 822 0453