Centre for Policy Modelling
Manchester Metropolitan University
----------------------------------
Contact:
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
Report on R&D in information agents area at the Center:
*******************************************************
The Centre for Policy modelling was established some five years ago to
develop a modelling software environment, methodology and techniques for
the analysis of complex social enviroments. As the name of the
institute indicates, a key element in our mission is to develop
capbilities for modelling that will inform policy analysis and
formation.
We provide two major publicly available resources:
1. SDML: a modelling language which corresponds to a fragment of
strongly grounded epistemic logic. SDML is sponsored by
ObjectShare (UK) Ltd. (formerly ParcPlace (UK) Ltd., the commercial
arm of Xerox PARC) who provide us with their current version of
Smalltalk. With that sponsorship, we make SDML available free of
charge for purposes of academic research. There are currently some
30 universities and research institutes in the SDML user community
including universities in Belgium, Spain, France, Switzerland and
the USA.
2. Bibliography on Complexity, developed and maintained by Bruce
Edmonds, Senior Research Fellow in the Centre. In the six months
to the end of May, 1998, there were some 70,000 requests to the
bibliography site at
http://www.cpm.mmu.ac.uk/~bruce/combib/combib.html.
The Centre for Policy Modelling is actively developing an alternative to
economic approaches to multi-agent systems in general. Examples of
recent papers are:
Scott Moss, "Social Simulation Models and Reality: Three
Approaches", MABS98, Paris, July 1998.
Scott Moss and Bruce Edmonds, "Modelling Economic Learning as
Modelling", Cybernetics and Systems, 1998, pp. 214-147.
Edmonds, B. . Modelling Socially Intelligent Agents . Applied
Artificial Intelligence, 12(7), 1998, pp. 677-699.
Scott Moss, Helen Gaylard, Steve Wallis, S and Edmonds, B. (1998).
SDML: A Multi-Agent Language for Organizational Modelling,
Computational and Mathematical Organization Theory, 4, 43-69.
Projects currently under development include
* to integrate the CPM's social modelling approach with physical
modelling to create a social-physical model for the analysis of
policies for dealing with climate change (with collaborators
including the Swiss Federal Institute of Science and Technology,
Carnegie Mellon Univesity, CNR in Rome, the Environmental
Change Unit in Oxford and the University of Surrey (see
http://www.cpm.mmu.ac.uk/cpmrep47.html);
* to integrate the CPM's organizational modelling with low level
robotics control for an entry into the next RoboCup competition
(with the Department of Cybernetics, University of Reading).
* to prove theorems concerning individual relationships emerging
from simulation models in order to determine, for example, which
properties are independent of random elements in each experiment,
which elements of model specification imply particular emergent
relationships and to formulate conjectures concerning relationships
which are general with respect to classes of formal logics.
* to develop canonical task environments supporting the nesting of a
variety of simulation models in order to determine which properties
of the models are general for that class of models, which differences
among task environments require what differences in representations of,
for example, agent cognition.
* to develop methods for the validation and verification of simulation
models involving agent interaction in complex social environments.