ABSTRACT
A common task performed during preliminary design is the selection, based on customer preference, between two alternative design concepts. This selection process is used as a means to improve the design, but is complicated by several factors: (1) most designs seek to satisfy multiple quality attributes, (2) the attributes may be subjective in nature and not immediately quantified, (3) the attribute objectives may be in conflict, and (4) there is typically uncertainty in the level of performance achieved by each design concept. The methodology described in this paper links design variables-tangible features of the design concept which can be manipulated by the design-er-to multi-attribute value, a number indicating customer preference. The design comparison is made by mapping the competing designs into customer value space, and applying a measure of overall customer preference which considers both cost and a quality attribute. The uncertainty of this mapping process is addressed, and an estimate of the most preferred design is provided.