Re: space: the initial frontier

[Ginger Booth]

Tim,
    Sorry for the slow response--I took a couple days off.

> I'm hoping to implement several simulation models in SWARM based on
> the notion of a set of interconnected "habitat patches."  What I need
> are some suitable space objects.  Since I don't want to reinvent the
> wheel, I'd like to know what "spaces" other people are working on.

    It sounds like your stuff has a lot in common with my Gecko stuff.
In our models, the landscape produces resources (habitat patches being one
such landscape).  Some agents acquire those resources at a rate determined
by their size, having to divvy up resources if they overlap, which they
generally do.  We also have agents that prey on other agents rather than 
having the capacity to eat sun, as it were.  Typically, a predator and
prey are coincident.  At least until one dies....

> For example, I'd like to have a 2D grid space where each grid-cell
> contained a list (or hash) of agents.  This seems fairly strait
> forward.  I noticed there is a List object in the collections library,
> so I could just subclass grid2D (listGrid2D?) and place a list at each
> location in the grid.  Has anyone else worked on this?

    I'm implementing our own space.  I'm not using a 2d grid space with
lists of agents.  Rather, I'm letting my agents roam free, having arbitrary
sizes (radii).  The infrastructure to do this isn't as simple as subclassing 
grid2D.  I'm still working out the details, but the basic data structure is
a "quadding" (zoning) tree, keeping agents sorted by area, for use in finding
neighbors.  At the moment, I'm working out how to feed the tree without 
gross errors or overhead.  The landscape itself (resource production and
memory by area) is probably a separate structure.  At the top of this nasty
bit of work, I'd like a fairly clean code interface gateway to the rest of
Swarm.

    If you're interested, please let me know.
    
Ciao,
    Ginger