Page 455 - Design for Six Sigma a Roadmap for Product Development

414 Chapter Twelve

Controllable factors
……
x 1 x 2 x n
• • •

Input Output y
Process
Response

• • •
……
z 1 z 2 z p
Uncontrollable factors
Figure 12.1 A process model.

where ε is experimental error, or experimental variation. The existence
of ε means that there may not be an exact functional relationship
between y and (x 1 ,x 2 ,...,x n ). This is because

1. Uncontrollable factors (z 1 ,z 2 ,…,z p ) will influence the response y but
are not accounted for in Eq. (12.1).
2. There are experimental and measurement errors on both y and
(x 1 ,x 2 ,…,x n ) in the experiment.
A DOE project will take many steps, described below.

Step 1: Project definition
This is the first but certainly not a trivial step. We need to identify the
objective of the project and find the scope of the problem. For example,
in a product design, we need to identify what we want to accomplish.
Do we want to reduce defect? Do we want to improve the current prod-
uct’s performance? What is the performance? What is the project
scope? Do we work on a subsystem or a component?

Step 2: Selection of response
variable (output)
After project definition, we need to select the response variable y. In
selecting this variable, the experimenter should determine if it could
provide useful information about the process under study. Usually, the