Publications des agents du Cirad


Integrated modeling : from models to theories and back

Müller J.P.. 2010. In : Proceedings of the iEMSs Firth Biennial Meeting : International Congress on Environmental Modelling and Software (iEMSs 2010), Ottawa, Canada, 5-8 July 2010. s.l. : s.n., 2 p.. Conference of the International Environmental Modelling and Software Society. 5, 2010-07-05/2010-07-08, Ottawa (Canada).

"A is a model of B for X if manipulating A allows to answer the questions of X on B"(Minsky 65). It follows that the formal structure A, the system B, and A being a model B depends on X and his questions. Integrated modeling is thought as the process of putting together several models (say A1 to An) or to build a new model A'which subsumes the models which would have been designed separately. We argue the main limitation lies in the lack of explicitation of how Ai or A' relates to Bi or B' for the Xs performing integrated modeling.To make the relationship between A and B formally explicit, we need a formal proxy of B. The usual proxy is the set of observations we have about B. We suggest that another proxy of B is the theory in the philosophical sense of a conceptual explanation of how the system B behaves associated with its ontology. We argue that the theories in the philosophical sense must be made explicit because integrated modeling is about articulating theories and not only about integrating models. From our experience in multidisciplinary contexts, we are using ontologies, and combining and articulating these becomes the heart of the integrated modeling process. The formalism (differential equations, multi--?agents systems, etc.) is chosen after and associated with the ontologies. Both the articulation among theories, and the process of abstraction between the theories and the chosen formal structures is in the center of the integrated modeling process. (résumé d'auteur)

Documents associés

Communication de congrès

Agents Cirad, auteurs de cette publication :