A Comparison of Three-Level Orthogonal Arrays in the Presence of Different Correlation Structures in Observations

When the experimenter suspects that there might be a quadratic relation between the response variable and the explanatory parameters, a design with at least three points must be employed to establish and explore this relation (second-order design). Orthogonal arrays (OAs) with three levels are often used as second-order response surface designs. Generally, we assume that the data are independent observations; however, there are many situations where this assumption may not be sustainable. In this paper, we want to compare three-level OAs with 18, 27, and 36 runs under the presence of three specific forms of correlation in observations. The aim is to derive the best designs that can be efficiently used for response surface modeling.

[1]  John Eccleston,et al.  AN ALGORITHM FOR THE DESIGN OF 2? FACTORIAL EXPERIMENTS ON CONTINUOUS PROCESSES , 1995 .

[2]  Michael Jackson,et al.  Optimal Design of Experiments , 1994 .

[3]  Steven G. Gilmour,et al.  Projective three-level main effects designs robust to model uncertainty , 2000 .

[4]  John Eccleston,et al.  Some results on two-level factorial designs with dependent observations , 1998 .

[5]  Kenny Q. Ye,et al.  Geometric isomorphism and minimum aberration for factorial designs with quantitative factors , 2004, math/0503678.

[6]  Damaraju Raghavarao,et al.  Optimal s , 2006, Comput. Stat. Data Anal..

[7]  D. Raghavarao,et al.  Balanced 2 n Factorial Designs When Observations are Spatially Correlated , 2009, Journal of biopharmaceutical statistics.

[8]  Testing Moving Average Against Autoregressive Disturbances in the Linear-Regression Model , 1991 .

[9]  P. Goos,et al.  A variable-neighbourhood search algorithm for finding optimal run orders in the presence of serial correlation and time trends , 2006 .

[10]  Lih-Yuan Deng,et al.  Orthogonal Arrays: Theory and Applications , 1999, Technometrics.

[11]  C. Koukouvinos,et al.  18-run nonisomorphic three level orthogonal arrays , 2007 .

[12]  Steven G. Gilmour,et al.  Some new three-level orthogonal main effects plans robust to model uncertainty , 2004 .

[13]  David M. Steinberg,et al.  Trend robust two-level factorial designs , 1991 .

[14]  R. J. Martin,et al.  An algorithm for the design of factorial experiments when the data are correlated , 1999, Stat. Comput..

[15]  R. Mead,et al.  Statistical isomorphism of three-level fractional factorial designs , 2006 .

[16]  Gregory M. Constantine,et al.  Robust designs for serially correlated observations , 1989 .

[17]  A. Dean,et al.  Projection properties of certain three level orthogonal arrays , 2005 .

[18]  Yun Shen,et al.  Comparison of designs in presence of a possible correlation in observations , 2006 .

[19]  Julie Zhou,et al.  A Robust Criterion for Experimental Designs for Serially Correlated Observations , 2001, Technometrics.

[20]  G. Jenkins,et al.  The Estimation of Slope When the Errors are Autocorrelated , 1962 .