Abstract

Qualitative reasoning is an alternative problem-solving technique useful for the conceptual design of structures. Qualitative reasoning represents the relationships between parameters in a model, and a search computation assigns values represented by intervals and relevant points in the behavior. The traditional difference between analysis and design or input and output parameters in a procedural computation is not existent in qualitative reasoning, since all the parameters in a model are equally represented. Qualitative reasoning derives values for parameters even with incomplete and imprecise knowledge about the model. This work presents a qualitative structural analysis framework, suitable for the evaluation of conceptual designs as well as for tutoring systems. The framework has been implemented in a computer program called Agrippa using the computer language Prolog. Based on a representation of fundamental principles of equilibrium, compatibility, and force-deformation and an incomplete knowledge of geometry and topology, Agrippa derives the signs and relative magnitude of forces and displacements for three-dimensional models of structures.