About Possibilities for Real and Alleged Uncertainties within Numerical Optimization Processes

摘要

Within the usual product design process uncertainties are quite common and can be handled by well standardized procedures like Six Sigma or Robust Design methods. During the last years, numerical optimization is increasingly applied for the improvement of product qualities. However, there exist some kinds of real and also alleged uncertainties in this context that stem from different sources and may harm the successful utilization of the latter. It is necessary to understand theses coherencies in order to work properly within the optimization design process. This survey points out possible causes of uncertainties within the optimization process, their sources and impacts with respect to simulation, as well as alleged uncertainties, which are often dealt with in an inappropriate way. The focus lies on the design process itself, represented by the simulation and optimization workflow, as well as human factors, represented by uncertain goals respectively knowledge deficiency. The impact of these factors will be shown and possible solutions presented.