Context-sensitive selection of methods during product design

In this article a probabilistic approach for detecting suitable methods to support the tasks in different stages of product design will be presented. The problem domain is modeled with the aid of an influence diagram, which is an extension of a Bayesian network. This influence diagram contains the objects and parameters, which are relevant to the decision-making process, namely choosing an appropriate method. First these parameters and objects have to be identified. Secondly the knowledge about dependencies and independencies among the collected variables is used to build up the topology of the influence diagram. In a final step the net has to be quantified, i.e. the probability values have to be specified. Now the decision model is completely defined and after supplying values for the observed or known variables, i.e. the conditions and costraints, utility values for the alternative methods are computed and the method with the highest value will be selected.