Optimal combined usage of formulas, fuzzy-logic and neural networks for calculation in product development

This paper describes the optimal representation and application of calculation-related knowledge in product development. Classification into verified, uncertain and empirical knowledge supports the choice of suitable representation methods: formulas, fuzzy logic and neural networks. Potential use of neural networks will also be described in further detail. Joint application of different knowledge types is achieved by means of a hybrid simulation model combining the advantages of different representation methods to provide the best possible supply of knowledge. The modularization of knowledge components, which significantly facilitates further use, and their representation according to the quality of information are significant potential applications of the hybrid simulation model.