Modeling ordered decision making in product development

Abstract

This paper models the process of ordered product design decision making and shows that the order in which product development decisions are made has a significant effect on the product designed. The problem is formulated by introducing the notion of Quality Loss to quantify the loss of design freedom incurred by the decision makers in the later stages of a cross-functional decision process. In this context, the optimal order is the decision order with the lowest quality loss, whose characterization is one of the contributions of this paper. Design variables are partitioned into exclusive groups based on their connectivities, and the design sensitivities and connectivities are combined to develop two properties called order invariance and task invariance. Sufficient conditions, relating exclusive group structure to these invariance properties, are presented to help enhance our understanding of the optimal order. When all decision variables are order invariant, the optimal order identification problem is shown to map to the well known shortest path problem. The invariance properties derived in the paper can also be used to simplify subsequent design iterations, as shown with a simple industrial product design example.