Geometric and Feature modelling

---

History-independent boundary evaluation for feature modeling

Abstract

Most current feature modeling systems strongly rely on a history-based interpretation of the feature model, in order to maintain its evaluated boundary representation. This dependency on the model history is undesirable, as it forces the user of the modeling system to reason in terms of a strict chronological feature creation order. Moreover, re-evaluation of the boundary representation, as performed in such systems, has a computational cost proportional to the size of the model history. Such drawbacks suggest that current feature modeling systems are still too tied to conventional geometric modeling techniques.
In this paper, it is argued that to overcome the drawbacks mentioned above, a declarative feature model is required, whose structure is dynamically adapted as modeling operations create or modify relations among its features. Operations performed on this feature model can then be efficiently propagated to an evaluated non-manifold geometric representation, without invoking model history considerations. The paper describes how such a geometric model the socalled Cellular Model can be maintained throughout model evolution. For each modeling operation, this is achieved in two phases. First, the Cellular Model is incrementally re-evaluated. Second, the Cellular Model is interpreted, according to the feature information stored in its cellular entities and the current dependencies among the features. The advantages of the use of this history-independent boundary evaluation, implemented within the semantic feature modeling approach, are illustrated with some modeling examples.