Reasoning with assumptions, defeasibly, in model formulation

Abstract

The authors examine the relevance of reasoning with assumptions in two processes that are desired to be supported in model management systems, namely model formulation and model version management. They submit, and illustrate with an example, that the ability to represent and reason with assumptions in modeling languages could lead to significant improvement in the functionality of model management systems. The authors also argue that the process of reasoning with assumptions is non-monotonic and propose that defeasible reasoning is a useful candidate for modeling this process.<>