 |
 |
|
 |
 |
|
 |
 |
 |
The many design types included in DOE++ are sufficient for most applications. However, sometimes limitations on resources and factor level combinations will require you to build a custom design that is based on one of the included design types.
Starting in DOE++ 9, you can use a tool to build a custom design that is optimized for a selected regression model (i.e., the design will minimize the uncertainty of the regression coefficients) or optimized to cover as large of a design space as possible (i.e., the distance between design points is maximized). For example, if you only have 11 units available to test using a two level factorial design with five factors, you could build a two level full factorial design (which would require 32 runs), and then use this tool to build a modified version of the design that uses 11 runs and takes into account only the effects you want to investigate, as well as any restrictions on the factor level combinations that can be used in the experiment.
The following topics explain how to create custom optimal designs, including:
© 1992-2016. ReliaSoft Corporation. ALL RIGHTS RESERVED.
 |
E-mail Link |