Case-Based Adaptation for Product Formulation
Résumé
The fact that Case-based reasoning (CBR) adaptation in design domains is knowledge-intensive is one of the major factors that has limited the industrial application of CBR systems. Nevertheless, inductive techniques can ease the adaptation knowledge acquisition bottleneck by enabling useful knowledge to be elicited from the case-base (CB). The use of neural networks that use the knowledge available in the CB to (i) generate a desired mapping from differences between a query and retrieved cases, (ii) to minimise those differences and hence (iii) adapt retrieved cases so that an optimal solution to a query is found is studied in this paper. This adaptation method is suitable for CBR systems that use numerical-valued attributes for describing a case.
Origine : Fichiers produits par l'(les) auteur(s)
Loading...