, the Rule Engine for the JavaTM
|Hey Scott ... tell us about yourself :-) -JM
I work in the area of social simulation and am the Director of the Centre for Policy Modelling in the Business School in Manchester Metropolitan University (UK). My interest in Jess stems from the view developed with my colleagues over the past decade and more that modelling individual behaviour and social interaction should be done declaratively. The reason for this is that building models around social evidence means (a) that the evidence is usually expressed qualitatively and (b) that engaging with stakeholders in (say) a policy development process is most easily achieved if the models incorporate the language the stakeholders use and return results in that language as well. At the same time, numbers are naturally used to represent other phenomena such as the physical and biological aspects of the environment and numbers are best manipulated procedurally.
Jess is really good for the way it allows and supports the integration of declarative and procedural representations. Being able to embed Jess models in Java adds enormous flexibility. The drawback of Jess for these purposes is that the only easy and natural way to implement agents with individual cognition that interact socially is to give each its own Rete engine. Using a single engine for all is cumbersome because every fact has to state explicitly whether it is private to the individual (and which individual) or publicly obtainable. We wondered about defining a module of each agent. This was discouraged by Dr Friedman-Hill. We need to ask him why.
Front Page |
Recent Changes |
Powered by Friki |
Last Edited: 09 June 2006