Design of Experiments Lean Six Sigma

Y = f(x)

DOE was originally developed in 1930’s by Sir Ronald Fisher to improve agricultural methods Fisher used DOE to maximise the yield of agricultural crop (Y) by changing the key process inputs; fertilizers & seed type (x’s) he DOE approach allowed Fisher to understand the main effects of the inputs, and the interactions between the inputs which impact the process output

The objective is to logically organise changes to 2 or more input variables (x’s) and evaluate if any variable, or any combination of the variables, significantly affect the output (Y)

What is Design of Experiments?

A DOE is a set of tests on the process output with at least 2 process inputs, each set at 2 or more levels

The key principle behind the DOE technique is to create a perfectly balanced design which includes an equal combination of process settings

Consider the example below with 3 key process inputs, each set at the high end (1) and low end (0) of their respective specification (or process variation) limits