Name: Georges Fadel
Email: gfadel@ces.clemson.edu

Comments:
We all agree decisions are taken at multiple levels in the design process. It is not rational to believe that the objective of Profit or any other single utility can be derived at all levels and help drive the design process. At the highest level, such a utility could exist and maybe be correctly formulated, but at sublevels, there is a need to understand the trade-offs and only by exploring the Pareto set can one see the various solutions and make a decision.
I also have a fundamental issue with the single objective approach since it implies only one optimal solution. This means there can be a perfect product for a set of customers and for a company. This is obviously impossible since a better product will surface once this perfect product is out. One could argue a time factor, but then we revert to a multi-criteria approach with multiple "single" objectives of profit at various times. (I call this multiple scenarios).
My other big hangup on the single utility approach is that it presumes the designer can come up with all the individual relevant objectives and bring them together in this utility. This is probably a flawed assuption. Typically, a designer will concentrate on the most relevant and critical objectives and figure out how to do a trade-off between those. During that trade-off exercise, he/she may observe the effect of changing a solution on some objective not considered in the study, and let that objective influence the choice with the others. A Utility based approach for me is just one objective, and cannot reasonably capture all the criteria.

Comments:
The questions posed for debate are not for debate, the answers should be derived. Debating these is like debating whether nature should use F=ma or not. It is or it isn't. It is to be found out, not to be debated. Where does a question like, "All multi objective decision methods are flawed," come from? That is not a meaningful question.

Comments:
In order to address this question, one should also think about the context of decision situation. By context, I'm refereeing to framework and level of which the decision situation is taking place. At the highest level, the objective of every design activity is to make profit. However, the next question that should be asked is "how to make this profit?" which leads the designer to elaborate this high level objective and identify several aspects of it which are relevant to the current decision situation. These lower level objectives collectively, will guide the designer towards achieving his/her ultimate goal (some literature refer to ultimate objectives as strategic objectives [Keeney, 92]). In my view, design process is a cascading process of elaborating one (or more) high-level objectives, using the strategic objectives as guidelines for the elaboration process. The strategic objective acts like an umbrella that all lower level objectives will contribute to achieving some aspect of it. For example, when designer faces a decision to design an axle to transmit power from gearbox to wheels, objectives could be listed as: maximizing transmission torque, and minimizing cross section diameter of the axle. In identifying these objectives, designer implicitly considered the "maximize profit" objective as the strategic objective of the design problem. But the context of design problem requires that designer considers more than one objective to satisfy the requirements.