Re: [Swarm-Modelling] beginner

[glen e. p. ropella]

hilit wrote:
> I wish to explore stray cat dynamics in my city by using an individual
> based spatially explicit model.
> 
> Problem is, I don't even know how to start!!
> 
> I don't have any background in programming.
> 
> I was told by a colleague who has done such a model to begin with a flow
> chart of all the cat behaviours I know, but I feel it is too early for
> me because I know so much about the cats and so little about "making a
> model".
> 
> Do you have any suggestions for a "green" like me?

The first thing to do when considering whether you can build a model is
to find out what type of data is available.  That's not to say a list of
cat behaviors isn't a good thing to do; but, models should be largely
data driven.  So, if you have published (or your own) data on stray
cats, that's where you should start.  And if you're lucky enough to find
a published _model_ of cat behavior in cities, then the first step is to
re-create that model.  This contributes to the "community of cat
modelers" (if there is such a thing) and allows you to learn a bit about
the modeling tools during the re-creation of that model.

My next step would be to find/build a geographical model of a city.  It
doesn't have to be _your_ city; but, some city even if it's
over-simplified and/or imaginary.  This is particularly important in
modeling something like stray cat behavior because their behavior will
be largely a function of the terrain in which they live.

With some type of geographical model, you can begin to list cat
behaviors in a very concrete way, referring explicitly to geographical
structures.  So, after you have the geography, you should list the cat
behaviors.

The next step might be to define any other actors, if you intend to
represent them.  Humans and prey are important food sources.  Dogs might
be predators, etc.

Alternatively, you could implement some magical functions to randomly or
periodically place food and danger sources into various spots in the
geography in lieu of implementing actors that generate the food or danger.

The final step toward a first iteration would be to define the schedules
and internal logic of the various actors.  These will be hypothetical
and you might need to play around with them quite a bit to get a
reasonable model.  Hence, when you define them, you should make them
easy to turn on/off and replace.

Since geography will be such a large part of this model, I recommend
picking a tool that has some GIS capability already integrated.  Take
the tutorials (if they exist) for that tool and build 1 or 2 very simple
models using that tool before you re-implement the published model.

-- 
glen e. p. ropella, 971-219-3846, http://tempusdictum.com
Every normal man must be tempted at times to spit upon his hands, hoist
the black flag and begin slitting throats. -- H.L. Mencken