Six Sigma for Business Excellence: Approach, Tools and Applications

15

Design of Experiments Part 2: Test Less, Learn More—Fractional Factorial Designs

Introduction

Full factorial experiments are very powerful when it comes to evaluating the effects of various factors and interactions. However, as the number of factors and levels increases, the required number of runs too increases rapidly, in geometric proportion. A large number of runs with more factors make the experiment even more expensive and difficult to administer, as is evident from Table 15.1 which shows the number of runs in full factorial experiments for two-level and three-level designs.

Let us consider a three-factor two-level 2³ experiment. In a 2^k experiment, we can evaluate k main effects. The number of r-factor interactions which can be evaluated ...