IT-Based Knowledge Management Systems to Support the Design of Product Development Processes - Information Technology and Product Development - page 49

Shoemaker, Leung Tsui, & Wu, 1991). Thus it is important to store the knowledge
obtained in these experiments for reuse later. The execution of such experiments
requires a combination of tacit and explicit knowledge. An explicit knowledge of
the methodology and tools is required but the tacit knowledge experience gained
over time from conducting these experiments is invaluable. For example, based
on knowledge gained over time by conducting experiments, an engineer may
obtain significant tacit knowledge and intuitively know the relationships between
the environmental factors (or process-related factors) and the product development
outcomes. Thus it is important to capture this knowledge. The following section
examines the use of DOE and RDM in greater depth.
3.3 DOE and Robust Design in Product Development
Engineers execute experiments to enhance their knowledge of a particular process
or product. One of the most common techniques of doing so is by varying one factor
or one variable at a time while holding the other factors constant. This method of
experimentation is commonly known as one-factor-at-a-time (OFAT) and can be
regarded as a form of “trial and error” requiring a mixture of luck, experience, and
intuition for its success (Clements, 1995).
The OFAT approach to experimentation for determining main setting parameters
is still very popular in today's organizations for several reasons (Antony & Tzu-
Yao, 2003). It is commonly thought that the most accurate way of measuring the
effect of a design change is to keep all the other factors fixed while one factor is
being assessed. Furthermore, it is believed that OFAT techniques are easily con-
ducted and do not need any advanced statistical knowledge in their application. The
OFAT process also provides a “quick-fix” solution that managers are often content
with. This is due to the fact that the significance of DOE is not stressed enough to
engineers within academic institutions. Antony and Tzu-Yao (2003) highlights that
many companies are not ready for the implementation of advanced quality improve-
ment techniques such as DOE. However, the use of an easy-to-use information
technology-based system to support the DOE process may address such issues.
Statistical experimentation such as the DOE and the analysis of variance
(ANOVA) technique dates back to the 1930s (Fisher, 1935), “DOE” is one of
the most powerful quality improvement techniques for reducing process variation,
enhancing process effectiveness, and process capability (Antony, 2006). Fishers'
approach to experimentation was a direct alternative to the OFAT approach. Since
then, his approach has evolved into a number of techniques for improving process
performance/capability and reducing process variation (Montgomery, 2001b).
DOE offers a number of advantages to experimentation over the OFAT approach
(Antony & Tzu-Yao, 2003). First, DOE requires much less resources (i.e., a number
of experiments, time, material cost) than OFAT experiments for the same amount
of information (or insights) obtained. For example, with three factors at two levels,
a full factorial design requires only eight runs compared to 16 runs for an OFAT