Arnoud Visser wrote:
> In general I agree, except that those parameters maybe grouped into
> classes, that are known. I found the restrictions about prior knowledge
> on buildings on the map always a bit too strict. In reality the
> firebrigade has made a risk-assesment, and knows for instance where
> dangerous industry is located. How dangerous at this situation they do
> not know, but they will be a big difference with burning
> office-buildings. The same holds for ambulances (busy places as
> stations) and police forces (important routes).
I fully agree, Arnoud. Usually, agents will have vague (abstracted,
generalized, statistical, whatever) knowledge that roughly corresponds
to the more detailed internal data of the simulator. I would have no
problem, if agents knew the relations between their restricted model and
the "real" unknown data. Just like fieryness is a helpful abstraction
from more complex issues that we understand in theory (that is, when we
have a look at the simulator source code) but nevertheless cannot really
use.
michael
-- Michael Brenner Phone: (+49) (761) 203-8226 Institute of Computer Science Fax: (+49) (761) 203-8222 Albert-Ludwigs-University mailto:brenner@informatik.uni-freiburg.de Am Flughafen 17, Geb. 52 http://www.informatik.uni-freiburg.de/~brenner D-79110 Freiburg, Germany
This archive was generated by hypermail 2.1.3 : Wed 21 Apr 2004 - 15:43:28 GMT