Design of Experiments Guide  • Mixture Designs • Examples of Mixture Design Types • Optimal Mixture Design
Previous
 • 
Next
Optimal Mixture Design
Optimal mixture designs use the Custom Design platform with the mixture variables entered into the response and factors panels.
To create an example optimal mixture design with three factors:
1.
Select DOE > Classical > Mixture Design.
2.
(Optional) To match the output of this example, click the Mixture Design red triangle and select Set Random Seed, and then enter 1409.
3.
Use three factors for this example. No changes need to be made to the Factors section.
4.
Click Continue.
5.
Click Optimal on the Choose Mixture Design Type panel.
6.
In the Define Factor Constraints panel, click Specify Linear Constraints, and then click Add.
7.
Enter a 1 in the X1 and X2 boxes and enter 0.8 in the constraint box.
This constrains X1+X2 to less than 80% of the mixture.
8.
In the Model panel, click Interactions > 2nd.
This adds interaction effects to the model.
9.
(Optional) To match the output of this example, click the Mixture Design red triangle and select Number of Starts, and then enter 2.
10.
In the Design Generation panel, enter 2 for the Number of Center Points.
11.
Click Make Design.
Figure 12.6 12-Run Optimal Design Runs
12.
Click Make Table.
Note: In practice, you would round the mixtures amounts to the appropriate level of precision.
Visualize the Design
1.
From the design table, select Graph >Ternary Plot.
2.
Select X1, X2, and X3 and click X, Plotting, and then click OK.
Figure 12.7 Ternary Plot for Optimal Design
The ternary plot shows the 7 unique mixtures that make up the design. For details about ternary plots see Ternary Plot Overview.

Previous
 • 
Next
Help created on 7/12/2018