Using Parameterized Pareto Sets to Model Design Concepts

Decisions made during conceptual design can have a major impact on the success of a design project, and designers must take care to select a concept that leads to a desirable design solution. However, the inherently imprecise nature of design concepts complicates decision making. A single concept relates to a large set of specific design implementations, each of which has a different level of desirability based on the tradeoffs designers are willing to make. Thus, designers must consider tradeoffs across the many possible implementations of a design concept in order to decide between concepts rigorously. To accomplish this efficiently, designers require an abstract understanding of the characteristics of a design concept. In this paper, we describe an approach to modeling design concepts that is based on an extension of the notion of a Pareto set, called a parameterized Pareto set. Using this construct, designers can generate a model based on information about prior implementations of a design concept in a way that includes tradeoff information while being independent of implementation details and reusable for different design problems. We demonstrate the approach on the conceptual design of a gearbox. The example involves two different design scenarios that serve to demonstrate the reusability of the model and effectiveness of the overall approach.