Decision Support Problems (DSPs) are used to model design decisions involving multiple trade-offs. In practice, such design decisions are also coupled, that is, these decisions must be modelled by identifying and addressing the influence they exert on one another. Hence, we need to classify coupled decision problems and to introduce methods for managing uncertainty for such problems. Classification of coupled decision problems allows for the development and execution of decision templates to effect design and to archive design-related knowledge on a computer. Incorporating robustness metrics allows for the exploration of robust design solutions for coupled decision problems by managing uncertainty.

In this paper, we present a classification scheme for coupled decisions using DSPs, called the Decision Scenario Matrix and we illustrate its utility by solving a coupled problem using DSPs. The design of a beam to be used as a fender is used to illustrate the efficacy of the formulation of coupled problems. In the first example, we determine a robust design, that is, determine the dimensions of the fender and simultaneously design the material recognizing that the computational models are incomplete and inaccurate. In the second example, we determine robust design solutions when design decisions are coupled, that is, determine the dimensions of the fender and select the material concurrently. Our focus, in this paper, is on illustrating the efficacy of the method rather than on the results.

This content is only available via PDF.
You do not currently have access to this content.