DOE - Design of Experiments? What does that mean?

A

amanbhai

what is design of experiment? what's the use of it. where is it applicable in practical word.
How can i describe it DOE. Can anyone help me in that?
:thanks:
 

apestate

Quite Involved in Discussions
DOE was DOA

amanbhai

There is much more information on this subject in threads than anyone could hope to nail down in one post. Please try reading some of the information through this link:

DOE - Design of Expriments

Brief story:

We work from blueprint drawings, and manufacture machined parts. We have a giant filing cabinet system, and one day I started wondering how many specifications were in that set of cabinets. 28 +/- 0.8 as a specification for example.

I happened to have a list of all these files, 1 through 9160. Then I picked 20 at random, using a random number generator. Then I took those 20 randomly picked files and counted the number of specifications on each print, and got an average. On average, I got 21.8 specifications per blueprint, with a rough average of 199,688 specifications total.

During the design of this experiment, I looked for critical problems, such as how random were my picks? Am I accidentally sampling prints with more or less specifications? Etc. DOE.
 

Statistical Steven

Statistician
Leader
Super Moderator
amanbhai said:
what is design of experiment? what's the use of it. where is it applicable in practical word.
How can i describe it DOE. Can anyone help me in that?
:thanks:
Design of experiments from a statistical perspective (most commonly used) is a systematic process to understand the variability in different factor levels.

DOE has many different uses, but it usually used to characterize the process and ultimately make importants to optimize the process based on the results.

It is applicable in ALL industries
 
B

Bill Pflanz

Steve,

Do you have any examples for using DOE methods in a non-manufacturing area?

Bill Pflanz
 

Statistical Steven

Statistician
Leader
Super Moderator
Mr

Bill Pflanz said:
Steve,

Do you have any examples for using DOE methods in a non-manufacturing area?

Bill Pflanz
As a matter of fact I do. I gave a talk once on using DOE in shipping.

The factors that were controlled was shipping company, # of packages shipped, distance to the destination and costs. Distance and costs are actually covariates in the model. We measured on-time performance and number of days to be received.

The purpose of the study was to find the best carrier based on the number of units being shipped. Some carriers are better with small shipments (FedEx) versus big shipments (Yellow).

I hope that helps.
 
Top Bottom