Capturing and implementing human algebraic reasoning for qualitative simulation

This paper addresses some shortcomings of current qualitative simulation being compared to human reasoning about dynamics. Some are related to control-theoretical approaches that remain at a low level of abstraction and others to the need of richer knowledge representation and inference on quantity-based information. The components of human algebraic reasoning beyond that of current qualitative models are also discussed. A general quantity base management system that implements those ideas for more practical qualitative modeling is presented with its design rationales