An increasing number of manufacturers offers customizable product spectra instead of common invariant products. This can be explained e.g. by growing market competition and a steady trend towards individualization in society. Product spectra allow the customization of products to meet customers’ requirements by adapting specific attributes, the so-called degrees of freedom. Different methods have already been developed to set up a product spectrum, but so far, no efficient proceeding is available to determine a customized product from a spectrum. One basic assumption of common methods for product specification is the availability of products in discrete specifications, e.g. building blocks or type series. These methods focus on the selection of one product variant amongst predetermined possibilities. However, customizable products have to be adapted to customers’ requirements in contrast to selecting between given alternatives. For this reason, established procedures do not meet actual demands. In this contribution, we present a neural network approach for specification of customized products. These networks are highly adaptable and represent powerful tools for modeling complex interdependencies, which are often hard to quantify.     «

An increasing number of manufacturers offers customizable product spectra instead of common invariant products. This can be explained e.g. by growing market competition and a steady trend towards individualization in society. Product spectra allow the customization of products to meet customers’ requirements by adapting specific attributes, the so-called degrees of freedom. Different methods have already been developed to set up a product spectrum, but so far, no efficient proceeding is availabl...     »