Hi,
>What does it mean to say that a language "has been shown to bear
>sufficient relations to"
>a theory?
I'm sorry if that isn't a precise description of what he did. He
actually considered two implemented BDI architectures (PRS and dMARS)
and designed a language, called AgentSpeak(L) which abstracts those
implementation (the reverse of the usual process, if you like). He
then gave an operational semantics which makes it clearly that the
language abstracts those practical architecture and showed a
one-to-one relation to a proof-theoretic semantics (based on
transition systems). The rules of this proof-theory can then be used
to show that the agents defined in that language actually satisfy the
properties of BDI agents which have been proved using the "expressive
BDI logics". But I no expert at all in BDI agents: you'd better see
this reference (I hope the bibtex format is useful):
@INPROCEEDINGS{RaoASL,
AUTHOR = "Rao, Anand S.",
TITLE = "{AgentSpeak(L)}: {BDI} Agents Speak Out in a Logical
Computable Language",
BOOKTITLE = "Proceedings of the Seventh Workshop on Modelling
Autonomous Agents in a Multi-Agent World
(MAAMAW'96), 22--25 January, Eindhoven,
The Netherlands",
PUBLISHER = "Springer-Verlag",
YEAR = "1996",
EDITOR = "Van de Velde, Walter and Perram, John",
NUMBER = "1038",
SERIES = "Lecture Notes in Artificial Intelligence",
PAGES = "42--55",
ADDRESS = "London",
NOTE = "URL:
\mbox{\texttt{http://www.aaii.oz.au/bios/rao.html}}" }
>Was Castelfranchi addressing the scale problem?
I actually don't know the answer. I only hope I haven't quoted him in
vain. The best you can hope now, as far as an answer is concerned, is
that he will reply to your e-mail himself.
(In any case, if the only scalable solution we have at present is not
expressive enough, we don't have a solution at all -- better to
devise a good one and then concentrate on the practical issues?)
Rafael
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|