Previous Page

|

Next Page

Chapter Topics

Response surface designs examine both the linear and the quadratic behavior of the responses over the design region. These designs, which are typically sufficient to fit a quadratic model, enable you to do the following:

• Distribute points in such a way as to provide information about the response throughout the experimental region
• Block runs
• Compute an independent estimate of error
• Use as few runs as feasible
• Estimate model parameters and carry out an analysis of variance
• Fit models that can be used to find optimal factor settings
• Test for model adequacy, including lack of fit

Refer to Myers and Montgomery (1995) for a discussion of response surface designs.

Covered in This Chapter