Options
2004
Journal Article
Titel
Quality prognosis in rapid product development
Abstract
A concept for a strategy for quality prediction in distributed product development processes is presented in this paper. The research involves design and implementation of predicition tools for the optimal context-sensitive simulation of product quality containing a forecasting model based on conditional probabilities. The prognosis model utilises the Bayesian probability network that analyses both current and past data and considers existing interdependences between the subsystems. The network refers to the context of a holistic analysis of prototype development procedures and considers varying and imperfectly foreseeable conditions. Parameter learning of the network as a special characteristics of the concept, permits the optimisation of the forecasting process. The prognosis tool contributes to recognition of potential erroneous trends, so that the enterprise can react in time to changes caused by assignment of individual development steps. This approach has been exemplary implemented in a Bayesian Network tool, can however be applied in many problem domains using some software development environments.