Re: Thesis project... long winded.

[John Eikenberry [MSAI]]

On Thu, 27 Feb 1997, Holly Arrow wrote:

> Since the agents will be interacting with one another, they will need a
> system for assessing the risk of interaction -- i.e., whether they should
> trust another agent or not.  A basic friend/enemy detector.  Come to think
> of it, they may need an even more basic detector that will tell them
> whether they have encountered another agent or a "inanimate" object like a
> database.  

I plan on implementing something of a heirarchal identification scheme. 
Each agent will have a unique identity. This will include info on what 
type of object/agent it is, what node the agent is from, and a unique mark.
There may also be additional info based on the metaphor (I'm thinking of 
a pack identifier for the wolf metaphor and a 'caste' marking for the 
bees (ie. if its a worker, solder, etc)).

> What are the agents trying to achieve?  Are they foraging for
> information?  What kind of information?  Will the interactions with other
> agents involve an exchange of information?

I hopefully answered these in the responce to Nelson's mail.

BTW, Nelson, Holly, thanks for the responces. :)

John Eikenberry
address@hidden
______________________________________________________________
"A society that will trade a little liberty for a little order
 will deserve neither and lose both."
                                         --B. Franklin