[Modeling] agents vs actors
Mon, 22 Sep 2003 21:11:27 +0200
Hi Jim, See below after FT>
----- Original Message -----
From: "James Odell" <firstname.lastname@example.org>
Sent: Monday, September 22, 2003 6:18 PM
Subject: Re: [Modeling] agents vs actors
> Hi Frits,
> If I understand your definition correctly you are saying:
> An agent is a technological implementation of an actor.
FT> No Jim, you did not understand me correct. I am only referring to one
particular project where we try to use agent technology to do something in
our field of research that seems somewhat of track. In our definition a
virtual actor is an technological implementation of a human actor (like an
architect, structural engineer, building advisor, and such), and a virtual
actor consists of a rather large number of
agents each performing one of the roles of the (virtual) actor.
> And by "technology" do you mean using some human-fabricated mechanism?
FT>No, by technology I mean agent technology, i.e. the concepts, models and
tools used to create multi agent systems.
> So, that if I were able to create a human, say Frankenstein's monster,
> that human would be an agent? And if the human were not created with a
> technology (e.g., god or evolution), then it would not be an agent?
FT>So if you were able to create a monster and call it Frankenstein youre
name probably would not be Odell ;-/
> Please don't think that I am not trying be a troublemaker here. I am just
> trying to understand the fine line.
> Also, why do you think that an agent can only have one role? Most living
> creature have multiple roles. Why should agents?
FT>What we call virtual actors you call agents, compound agents, complex
agents, or whatever. To me 'agent' is a word mostly connected to the ICT
world (travel agents do not play a role in our field). So we prefer Virtual
Architect, Virtual Building Advisor, and such because that clearly displays
the performances required.
Moreover, there is not really a reason why a travel agent would not be
called an actor or, a virtual actor if the roles of the actor are performed
by some multi agent system. Or is there?
> On 9/22/03 11:44 AM, f.tolman scribed:
> > In our research effort we distinghuish between an actor and a virtual
> > Our actor notion follows the UML use case idea. If both fysical and
> > actor are interchangeable without corrupting the system at least some
> > definitions of 'intelligence' have been satisfied. Agent technology is
> > to implement this concept. Many agents may be required to create a good
> > working virtual actor. Mostly one role is assigned to each agent.
> Modeling mailing list