Neural-network-driven fuzzy reasoning for product development processes

Abstract

The quantitative and qualitative dependency measures of serial and parallel product development processes are analyzed. And the neural-network-driven fuzzy reasoning mechanism of dependency relationships is developed in the case that there is no sufficient quantitative information or the information is fuzzy and imprecise. In the reasoning mechanism, a three-layer feedforward neural network is used to replace fuzzy evaluation in the fuzzy system. A hybrid learning algorithm that combined unsupervised learning and supervised gradient-descent learning procedures is used to build the fuzzy rules and train membership functions. © Springer-Verlag Berlin Heidelberg 2005.