The Relationship Between Decision Making and Optimization

• Anonymous - A decision can certainly benefit from application of optimization but we should remember that no engineering design can be said to be optimized.
• Anonymous - Both have theire place but we need to define these.
• Anonymous - Optimization is a tool that can be used for decision making. Decision theory and utility theory should be used to structure the problem and optimzation to solve it.
• Anonymous - Design is supposed to be a systematic procedure.
• Bufardi - If the decision involves mor than one objective which is often the case then there is no optimal solution. At most we can seek a good compomise solution.
• Howard Duhon - Optimization is rarely the appropriate goal of decision making. A typical project involves many thousands of decisions most of which are made via engineering judgement.
• Ng Kok Weng - Decision made can merely be done based on requirement/s met and optimisation is a different basis of consideration.
• Louis Steinberg - not a; it is often impossible to find global optima or even to quantify all aspects of a problem. Not c: formal optimization is a method for determing where optimal utility is on b, I would prefer "many" in place of "most"

• Anonymous - Both a and b are important. If we are thinking about decision making, uncertainty is very important. It is important to identify and model all kinds of uncertainties involved in making decisions. I think that any single theory is not enough for modeling all kinds of uncertainties. Then, what is needed is formal methods for combining the uncertainties.
• Bufardi - If the uncertainty is due to incomplete information then it is completely wrong to use probability theory.The different methods of handling uncertainty are complementary and not competing since they apply to diffrent types of uncertainty.
• Howard Duhon - The most dangerous uncertainties in engineering design are the ones we don't indentify during the design process. What we need most are structured methods for imagining what might go wrong.

• Anonymous - I think the role of LEARNING should receive more emphasis in the DBD community. The quality of a decision is strongly affected by the quality of information which is available to the design maker. A more effective learner therefore will often make better decisions.
• Anonymous - Build a little, test a little - measure the results!
• Hai Shi - Find individual limited tasks, define the decision-making process. I'm not sure if it is possible to define a uniform or universal decision-making process mathematically at this time. This is like to build a general problem solver vs. expert systems.
• Anonymous - Before this question, we need to make sure or find out who is the decision maker in design. The designers, the managers? If the designer is the decision maker, does his preference reflect the company's interest? Should we relate every decision in design (selection of geometry layout, material, tolerance, etc.) to the final criterion, assumingly the profit of the product? What's the relationship between part/component and product/assembly?...to name a few. I feel the current discussion of DBD emphasizes too much on 'decisions' without exploring the practice of engineering design, such as the types of decsions made every day and how to apply our theories to this. Only after exploring all these issues can we start defining a rational, normative decision making process in engineering design.
• Howard Duhon - Attempts to define a rational, normative decision-making process are misguided. Engineering teams are social entities and engineering decisions are largely socially driven. The decision process cannot be and should not be completely rational. An effective engineering process must address mainly the effective identification of objectives (such as Keeney's Value Focused Thinking approach) and must account for the peoples need for control (as provided by the action science research of Argyris).