High there,
I agree with Scott but I would say that discrepancy between e.g. agent
theories (BDI) and impementation is also due to a problem of formal
description and definition of agent properties...R. Brooks and others proved
the concept of nonsymbolic platform for agents on a "useful scale" producing
some interesting kind of behaviours...maybe this is also true for socsim
systems...maybe the desired behaviour is the result of interaction of small
and simple components (agents) under some conditions bulding agencies of
behaviours...it is possible to construct formalisms of agents' (not only)
social properties but "the life is elswhere"...I'd like to know about your
suggestions, Scott (and from others too :-) what is the way of solving the
theory-praxis gap in building agent systems and how the work on methodology
of socsim systems should start...
hope to receive response
Alesh Kubiik
Department of Applied Informatics
University of Economics
Bratislava
Slovakia
-----Original Message-----
From: Scott Moss <[log in to unmask]>
To: simsoc <[log in to unmask]>
Date: 16. znora 1999 14:44
Subject: formalisms and applications
>In both the multi-agent systems literature in general and the social
>simulation literature in particular, we find a large number of papers
>representing agents by some formalism -- a logic or a mathematical
>representation. The usual procedure is to use the formalism to capture
>some intuition about human behaviour and then to derive some results
>either analytically or computationally. In effect, the authors of these
>papers start with their chosen agent representation and set it in a
>problem space that enables them to apply those representations to
>achieve some set of results.
>
>Sometimes explicitly, but more usually implicitly, the reported analyses
>of these representations, problem spaces and results are intended to be
>steps towards a more powerful and general implementation of agents that
>will function effectively in useful problem spaces such as in
>information filtering on large databases or the Web, in robotics, in
>planning or in modelling social and economic institutions.
>
>There is another distinct strand in the multi-agent systems literature
>in which a usefully-scaled problem is identified and the authors develop
>ad hoc techniques to reason about those problems either to devise useful
>software or to simulate real social issues.
>
>Do these two strands in the literature ever meet? Are there cases where
>(say) BDI or deontic or other logics or finite cellular automata have
>been used to represent agents that perform some function(s) on a useful
>scale?
>
>I ask these question because I see as a key reason for the scientific
>failure of economics the acceptance by economists of virtually every
>stripe the need to restrict the problem spaces they specify to the
>limited domains of application of their particular techniques of
>analysis and associated agent representations. Applied policy analyses
>typically use the semantics of the theory but there is no direct link
>between the applications and the theory.
>
>In the absense of an intertwining ot these two strands, I see this
>scientific failure as a danger for the multiagent systems community in
>general and the social simulation community in particular.
>
>regards to all
>scott
>
>--
>Scott Moss
>Director
>Centre for Policy Modelling
>Manchester Metropolitan University
>Aytoun Building
>Manchester M1 3GH
>UNITED KINGDOM
>
>telephone: +44 (0)161 247 3886
>fax: +44 (0)161 247 6802
>
>http://www.cpm.mmu.ac.uk/~scott
>
>
>
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|