In This Issue

 **************** 

A Comparison of Designs for Second-Degree Models

Dr. Selden B. Crary

Update on my May 1999 DOE Perspectives article

In my earlier article (DOE Perspectives, May 1999) I responded to Ray Myers’ paper, which had appeared in the October 1998 issue of the Journal of Quality Technology, by pointing out that optimal-design software now allows one to find designs that are often far superior to standard designs, such as central-composite designs.  I also pointed out some of the features of the optimal-design-generation program I-OPT that is available for Unix workstations from a University of Michigan web site.

One careful reader of my article thought that I should not be so single-minded in the pursuit of optimal designs for prediction because there are other objectives of experimentation.  This reader also thought I was attempting to knock down a straw man by picking central-composite designs and then showing that an I-optimal design could do better for prediction.  The target of my assault lacked substance for this reader because the I-optimal designs I used were generated assuming that the model was precisely a second-degree multivariate polynomial (a “quadratic model”) over an exactly specified region (a multivariate cube).  The reader concluded his remarks by stating that classical designs were “constructed specifically to deal with inadequately specified models,” and he surmised that an I-optimal design would fare poorly under different criteria.

I consider the reader’s comments an expected sort of response, part of a lively debate on DOE issues that will continue as computer-generated application-specific optimal designs replace traditional practices.

I answer my reader by stating that I do not suffer from an idée fixe on optimal designs for prediction, notwithstanding the facts that my group at the University of Michigan developed I-OPT and our primary focus over the last decade has been on optimal designs for prediction.  Rather, my view is that technology is now making it possible for the user to define the objectives of experimentation and to find a suitable design via optimization.  The objective may be parameter estimation, prediction, a hybrid of the two, or something else.  The method may be sequential and/or adaptive.  The missing elements are a means of obtaining the user’s objectives and software for setting up the objective function.  It is inevitable that these elements will be created in the coming decade.

Concerning my reader’s concern that I-optimal designs are based on the assumption that one has specified a perfectly known model function, I can point out that it is perfectly within technological feasibility to include in the user’s input specification to an optimal-design engine that the design be robust against model misspecification.  Of course, what is meant by “model misspecification” must be defined, but this is an advantage because it provides the user with the opportunity to provide input on his/her needs in this regard.

I-OPT includes one means of specifying robustness against model misspecification.  For any problem, I-OPT allows up to ten different model functions to be entered by the user, along with the user’s best estimate of the probability that each is the correct model.  For example, a user can enter that he/she has 90% confidence that the correct model is a full second-degree model and a 10% chance that the model is a full third-degree model.  The I-OPT search engine then uses a simple 90/10 admixture of the objective functions for the two models considered separately in its search for the optimal design.  Gone is the requirement that the user rely on authorities telling him/her that D-optimal or I-optimal, or whatever, are the best method.  The power is given to the user.  Former authorities may be discomfited.

Regarding the reader’s statement that classical designs were “constructed specifically to deal with inadequately specified models,” I can reply that there is no universally good design.  If there were, there would be no debate.  So what is one to do to resolve these different points of view?  One answer lies in careful study of the properties of various designs.  First, there must be software available for interested parties to find designs that meet various criteria, and then various designs must be compared on an objective basis.  The problem is that this demands a lot of work.  In this article, I point out some of the results we are obtaining at the University of Michigan.  The complete story will be published in the peer-reviewed literature.

Comparisons of designs

This past summer, I had the pleasant experience of having two outstanding undergraduate students working with me on summer fellowships supported by my department (Electrical Engineering and Computer Science at the University of Michigan).  One was an entering first-year student named Kimberly Kuether.  Kimberly took on the challenge of finding closed-form algebraic expressions for classical designs of experiments for second-degree models, and she did a wonderful job of it.  She learned about inverting partitioned matrices and then applied a symbolic-manipulation program to find expressions for the integrated variance (IV), D-optimality criterion, and A-optimality criterion for the following types of designs: central composite designs based on both factorial and fractional factorials, Box-Behnken designs, three-level designs in three factors, and fractional factorials of three-level designs.  The purpose here is to let people know that such comparisons are being made and whom to contact about them.  For this article I will focus on the results from her study of central-composite designs.

After scaling the ranges of the factors so that the region of interest lies between –1 and 1 in each factor, that is, over a [-1,+1]^k hypercube, a central-composite design in k factors consists of 2^k (2 raised to the power k) vertex points, 2k axial points a distance “a” from the origin, and “nc” center points taken at the origin of the region.  Kimberly exploited the high degree of symmetry of these designs to find her closed-form expressions.

One expression that she found gave the expected normalized integrated variance (IV) of the CCD’s as a function of k, a, and nc, based on the assumption that the model is a true second-degree function.  (The expected normalized integrated variance is a measure of how well the design will perform for predicting the response at untried locations, after data is taken at the design points and fit with least-squares fitting.  It is the I-optimality objective.)  Kimberly’s expression is complex, completely filling a computer screen with algebra, but this could then be used to explore various behaviors of the IV.

We could see in the algebra the well-known singularity in the IV of CCD’s for a=as=sqrt(k) (the symbol “as” is introduced for the value of “a” where the singularity occurs), when no center point is taken (nc=0), and we were able to determine that the singularity behaved, for all k, as (a-sqrt(k))^(-2).  This is a very pronounced singularity, which when plotted as IV vs. “a” has a width of order unity.  Put simply, the IV of a CCD with no center point is very high, from approximately 1<a<2 for k=2 and from 2<a<3 for k=6.  This means that it is critical that there be a good datum at the origin, as is common practice.

What is less well known, perhaps virtually unknown, is that even when one takes one or a few center points, there is a remnant of the singularity in the IV, and this can be a problem for users who use CCD’s for prediction.  This means that the predictive capability of a CCD will be relatively poor in the vicinity of a=sqrt(k) for all k.  We have observed that the effect persists over a width of order unity in a, just as for the original singularity.

How about some numbers?

Consider a commonly recommended design, the rotatable CCD (the contours of expected variance of prediction for a rotatable design are hyperspheres), where the value of the distance of the star points from the origin is ar=2^[(k-p)/4].  (For completeness, I have included the integer variable “p” to indicate the fractional nature of the design.  For a full-factorial CCD, p=0; for a half-fraction CCD, p=1; for a quarter-fraction CCD, p=2; etc.)  A small table is given immediately below that gives the location in “a” of the singularity when nc=0, that is, at as=sqrt(k), as well as the value of “a” for the rotatable CCD, “ar.”  

K

 as=sqrt(k)

 ar=2^[(k-p)/4]

2

 1.414 

 1.414

3

 1.732

 1.682

4

2.000

 2.000

5

 2.236

 2.378

6

 2.449

  2.828

It can be seen that “ar” and “as” are identical for both k=2 and 4, meaning that the rotatable CCD will be in the region of the remnant singularity for these values of k.  For other values of k of importance to industrial DOE the rotatable CCD is still in the region of the remnant singularity.  This is a warning sign.

How bad is the IV?

Immediately below is a table of the expected normalized integrated variances for a few CCD’s, with number of factors, k, ranging from 2 to 6 and the number of center points, nc,  ranging from 1 to 5.  (  Readers with an avid interest in numbers may be enjoy the observation that some of the IV values are rational numbers, as indicated in parentheses in the table.  The value of “p” is unity for all entries in this table.

K

 nc

  N

IV at sqrt(k)

 IV at ar=2^[(k-p)/4]

2

 1

   9

   0.6306 (227/360)

0.6306 (227/360)

3

  2

   16

  0.3701

0.3676

4

  3

   27

  0.2667 (4/15)

0.2667 (4/15)

5

  4

   46

  0.2001

0.1971

6

  5

   81

  0.1520

0.1369

6

  1

   77

  0.5194

0.2988

Is there a better way?

You be the judge.  And remember, this is only part of the story.  More study is needed, but the results that my group has found are instructive.  One can use an optimal-design engine, such as I-OPT or gosset to find the design that minimizes the IV over the region of interest, i.e., [-1,+1]^k, allowing that the design points can expand out to the same value of “a” that is used in the CCD.

For k=2, the N=9, I-optimal design is the following (or any of three equivalently good designs found by rotating this design by 90, 180, or 270 degrees about the origin):

[4x(-0.098, +0.098), (-1.000, -0.828), (-1.000, +1.000), (+1.000, -0.336), (+0.828, +1.000), (+0.336, -1.000)].  The integrated variance of this design is 0.2945, which is more than a factor of two lower that the IV for the rotatable CCD (0.6306).

In more factors, the benefit of using an I-optimal design as opposed to a rotatable, full-factorial CCD grows rapidly.  By k=6, the factor of improvement is greater than a factor of ten.  This means that you have the choice of using an I-optimal design with the same number of  points as the corresponding rotatable CCD and you can expect to get a better fit, or you may choose to use an I-optimal design with many fewer points, since the minimum number of points for a full second-degree I-optimal design in six factors is only 28.  So you could use, say 32, instead of the 77 to 81 points for the CCD, and still expect to have a better prediction, since the IV’s of the optimal designs scale roughly linearly with N.

Let me add that if the objective of experimentation is not prediction over the entire region of the factor space, then another objective may be more appropriate, such as D-optimality.  Kimberly has found that the D-optimal designs also perform much better than CCD’s and other classical designs, when the objective is minimizing the error in the parameter estimates.  Thus, what we are observing seems generic to optimal designs.  With a properly defined objective function optimal designs generally perform much better than classical designs.

The bottom line is that optimal designs can improve your experimentation greatly and/or reduce the amount of experimentation needed to succeed in your objective.

Since I’m on vacation in Denmark right now, I don’t have all the numbers from all of my group’s work available, but the comprehensive results will be included in a forthcoming report.  Anyone wishing a few numbers from my group may contact me.

**************** 

Short Course Notes

Note that our limited resources mean that these are the ONLY public classes that we are offering through the end of the year.  Space is limited so sign up now.  Note that completion of an approved basic class is a pre-requisite for attending the advanced class.

These classes are sponsored and run by Math Options (http://www.mathoptions.com).  Check the Math Options web site for full details.

Basic Statistics for Industry

Las Vegas, NV
Nov. 1

Performing Objective Experiments

(AKA A Modern DOE Workshop)
Las Vegas, NV
Nov. 2-4

**************** 

Long Distance Information

The free teleclass “How to Tell the Truth with Statistics” with Bill Kappele will be offered by Math Options on October 26, 1999, at 10 AM Pacific time (1:00 Eastern time).

For more information and to register please call

Bill Kappele at (888) 764-3958 or visit

http://www.MathOptions.com.

--

Bill Kappele also passes along this tip:  Note that reasonable is high praise coming from Bill.

Thomas Pyzdek’s article in the August issue of “Quality Digest” is a an excellent article.  Mr. Pyzdek discusses the dilemma faced by all researchers, namely, “How can I utilize data that wasn’t collected using DOE?”  His conclusions are extremely reasonable.

You can find this article, “Applying Virtual DOE in the Real World,” at http://www.qualitydigest.com/aug99/html/sixsig.html

**************** 

More Free Designs

Several designs have been added to the Math Options I-Optimal Design Library.  These are:

1.   Process Factor designs for spherical regions.

2.   Designs for mixtures and discrete factors.

I am also in the process of adding information about the designs, specifically the log condition number, the maximum VIF, the model, and whether the region is cubical or spherical.  You will find this information for some of the designs now.

Access to the library will now be much easier!  Those of you who download the library periodically will no longer need to provide contact information every time!  At the top right corner of the page for the Library, http://www.mathoptions.com/i-optima1.htm, you will find an entrance specifically for you in red.  This should make your life easier!

You can also download the entire library with one click now.  You will need a password to unzip the file.  The password is mailed to first-time registrants when they register.

 **************** 

Jobs

An R&D unit based just outside of Paris is looking for someone to head up their Molecular Modeling team.  For complete details, contact Marie De Maistre at mailto:transearch@transearch.com.

**************** 

Conference Notes

XIV Compstat
Utrecht, The Netherlands
21-25 August 2000
http://neon.vb.cbs.nl/rsm/compstat

 

*   *   *   *   *   *   *   *

Alan Brice Corwin

7349 16th Avenue NE

Seattle, WA 98115

206-525-9093

mailto:abcorwin@processbuilder.com

http://www.processbuilder.com/