Stationary Point

Stationary Point - The point (if it exists) on a Response Surface which may be located at a maximum, minimum, or saddle. The stationary point may be outside the experimental range. When the stationary point is far from the data region, it should be used only to indicate a general direction, requiring additional data for validation.

The stationary point is in the geometric n-dimensional space of the n-factors of the regression. For that reason, it may not be apparent in the 2-D contour (or in a 3-D surface) plot. It is usually advisable to use the n-dimensional space response surface result with the n(n-1)/2 contour plots to understand where additional data may be required to optimize the process. For example, when 4 main factors are found to be significant, you should investigate the 6 contour plots, plot (x, y, z), as (A, B, response), (A, C, response), (A, D, response), (B, C, response), (B, D, response), (C, D, response). In some cases, the mathematical stationary point may lie in a non-existent region, such as negative percent moisture, indicating the optimal region may be at the closest bound (zero percent moisture). Infeasible stationary points may also be an indication that a more precise experiment is needed over a smaller region.

Learn more about the DOE tools for designed experiments in Six Sigma Demystified (2011, McGraw-Hill) by Paul Keller, in his online Intro. to DOE short course (only $99) or online Advanced Topics in DOE short course (only $139), or his online Black Belt certification training course ($875).