Design of Experiments

1
Design of Experiments
Douglas M. Stewart, Ph.D. The Anderson Schools of
Management University of New Mexico
2
Robust Design

• Designing for Variance
• Tolerance Design
• Parameter Design (Robust)
• Example Shot Glass vs. Dinner Plate
• Measuring robustness
• Taguchis Signal to Noise Ratio
• Methodology
• Experimentation
• Monte Carlo Simulation

3
Other S/N ratios

• Smaller is better s/n
• Nominal is best s/n
• Larger is better s/n

4
Design of Experiments

• Example Making Espresso
• Full Factorial Design
• Number of Experiments
• Calculation of Effect
• Average at Level - Overall Average
• Partial Factorial Design
• Confounding the Interactions
• Tables

Y

X
Z
5
Robust Design

• Running The Experiment
• Choosing Factors Judgment
• Choosing Levels Judgment
• Controlling for Variation
• Randomizing
• Control of External Factors
• Example Taguchi Airplanes

6
Computer Aided Robust Design

• Requires a mathematically known relationship
  between experimental variables and outcome
• Wheatstone bridge
• Spring stiffness
• Design the experiment as normal
• Generate noise for each experiment
• Monte Carlo simulation
• Noise orthogonal array

7
Example Wheatstone Bridge