Decision Making and Tradeoff Studies

Terry Bahill
Systems and Industrial Engineering
PO Box 210020
University of Arizona
Tucson, AZ 85721-0020, USA
terry@sie.arizona.edu
http://www.sie.arizona.edu/sysengr/slides/tradeoffStudies.ppt
© 2004 Bahill

Most engineering decisions involve choosing between alternatives. Sometimes such decisions have to be made emotionally. However, quantitative data are often available to help justify the decision. When such quantitative data are available they are usually multi-dimensional, so using these data becomes problematic. Often the quantitative data are organized as figures of merit, and some method is devised to combine these figures of merit. Usually the figures of merit are merely summed together with appropriate scoring functions and weights, a technique called the Multi-Attribute Utility Technique, which is used in Quality Function Deployment (QFD). Sometimes the figures of merit are multiplied together, as in a cost to benefit ratio. Most of the time the choice between Sum and Product is not important. For example, suppose people are to be selected using two figures of merit: years of school completed and annual salary. Maximizing either the sum or the product of these figures of merit would be reasonable and it would not make much difference which was chosen. In this case the choice of a combining function would not be important because the two figures of merit are highly correlated. However we usually want to choose figures of merit that are not correlated. This causes problems for identifying the preferred alternative. We applied many techniques to the same decision problem and found that the preferred alternative is a function of the tool chosen.

References [55 and 79]. This lecture is suitable for engineers. This talk requires an overhead projector (or PowerPoint and a computer projection system). This talk takes one to eight hours depending on the format and content.