An orthogonal arrays approach to robust parameter designs methodology

Robust parameter design methodology was originally introduced by Taguchi [14 Taguchi, G. 1986. Introduction to Quality Engineering: Designing Quality Into Products and Process, Tokyo: Asian Productivity Organization.  [Google Scholar]] as an engineering methodology for quality improvement of products and processes. A robust design of a system is one in which two different types of factors are varied; control factors and noise factors. Control factors are variables with levels that are adjustable, whereas noise factors are variables with levels that are hard or impossible to control during normal conditions, such as environmental conditions and raw-material properties. Robust parameter design aims at the reduction of process variation by properly selecting the levels of control factors so that the process becomes insensitive to changes in noise factors. Taguchi [14 Taguchi, G. 1986. Introduction to Quality Engineering: Designing Quality Into Products and Process, Tokyo: Asian Productivity Organization.  [Google Scholar] 15 Taguchi, G. 1987. System of Experimental Design, Vol. I and II, New York: UNIPUB.  [Google Scholar]] proposed the use of crossed arrays (inner–outer arrays) for robust parameter design. A crossed array is the cross-product of an orthogonal array (OA) involving control factors (inner array) and an OA involving noise factors (outer array). Objecting to the run size and the flexibility of crossed arrays, several authors combined control and noise factors in a single design matrix, which is called a combined array, instead of crossed arrays. In this framework, we present the use of OAs in Taguchi's methodology as a useful tool for designing robust parameter designs with economical run size.     

Some robust parameter designs from orthogonal arrays

Robust parameter design, originally proposed by Taguchi [System of Experimental Design, Vols. I and II, UNIPUB, New York, 1987], is an offline production technique for reducing variation and improving a product's quality by using product arrays. However, the use of the product arrays results in an exorbitant number of runs. To overcome this drawback, several scientists proposed the use of combined arrays, where the control and noise factors are combined in a single array. In this paper, we use non-isomorphic orthogonal arrays as combined arrays, in order to identify a model that contains all the main effects (control and noise), their control-by-noise interactions and their control-by-control interactions with high efficiency. Some cases where the control-by-control-noise are of interest are also considered.     

Some resolvable orthogonal arrays with two symbols

A method of constructing a resolvable orthogonal array (4λk2,2) which can be partitioned into λ orthogonal arrays (4,k 2,1) is proposed. The number of constraints kfor this type of orthogonal array is at most 3λ. When λ=2 or a multiple of 4, an orthogonal array with the maximum number of constraints of 3λ can be constructed. When λ=4n+2(n≧1) an orthogonal array with 2λ+2 constraints can be constructed. When λ is an odd number, orthogonal arrays can be constructed for λ=3,5,7, and 9 with k=4,8,12, and 13 respectively.     

A semi-parametric approach to robust parameter design

Parameter design or robust parameter design (RPD) is an engineering methodology intended as a cost-effective approach for improving the quality of products and processes. The goal of parameter design is to choose the levels of the control variables that optimize a defined quality characteristic. An essential component of RPD involves the assumption of well estimated models for the process mean and variance. Traditionally, the modeling of the mean and variance has been done parametrically. It is often the case, particularly when modeling the variance, that nonparametric techniques are more appropriate due to the nature of the curvature in the underlying function. Most response surface experiments involve sparse data. In sparse data situations with unusual curvature in the underlying function, nonparametric techniques often result in estimates with problematic variation whereas their parametric counterparts may result in estimates with problematic bias. We propose the use of semi-parametric modeling within the robust design setting, combining parametric and nonparametric functions to improve the quality of both mean and variance model estimation. The proposed method will be illustrated with an example and simulations.     

Space-filling orthogonal arrays of strength two

This paper considers the use of orthogonal arrays of strength two as experimental designs for fitting a surrogate model. Contrary to standard space-filling designs or Latin hypercube designs, the points of an orthogonal array of strength two are well distributed when they are projected on the two-dimensional faces of the unit cube. The aim is to determine if this property allows one to fit an accurate surrogate model when the computer response is governed by second-order interactions of some input variables. The first part of the paper is devoted to the construction of orthogonal arrays with space-filling properties. In the second part, orthogonal arrays are compared with standard designs for fitting a Gaussian process model.     

The lattice of N-run orthogonal arrays

If the number of runs in a (mixed-level) orthogonal array of strength 2 is specified, what numbers of levels and factors are possible? The collection of possible sets of parameters for orthogonal arrays with N runs has a natural lattice structure, induced by the “expansive replacement” construction method. In particular the dual atoms in this lattice are the most important parameter sets, since any other parameter set for an N-run orthogonal array can be constructed from them. To get a sense for the number of dual atoms, and to begin to understand the lattice as a function of N, we investigate the height and the size of the lattice. It is shown that the height is at most ⌊c(N−1)⌋, where c=1.4039…, and that there is an infinite sequence of values of N for which this bound is attained. On the other hand, the number of nodes in the lattice is bounded above by a superpolynomial function of N (and superpolynomial growth does occur for certain sequences of values of N). Using a new construction based on “mixed spreads”, all parameter sets with 64 runs are determined. Four of these 64-run orthogonal arrays appear to be new.     

Classification of orthogonal arrays by integer programming

The problem of classifying all isomorphism classes of OA(N,k,s,t)$\mathrm{OA}(N,k,s,t)$'s is shown to be equivalent to finding all isomorphism classes of non-negative integer solutions to a system of linear equations under the symmetry group of the system of equations. A branch-and-cut algorithm developed by Margot [2002. Pruning by isomorphism in branch-and-cut. Math. Programming Ser. A 94, 71–90; 2003a. Exploiting orbits in symmetric ILP. Math. Programming Ser. B 98, 3–21; 2003b. Small covering designs by branch-and-cut. Math. Programming Ser. B 94, 207–220; 2007. Symmetric ILP: coloring and small integers. Discrete Optim., 4, 40–62] for solving integer programming problems with large symmetry groups is used to find all non-isomorphic OA(24,7,2,2)$\mathrm{OA}(24,7,2,2)$'s, OA(24,k,2,3)$\mathrm{OA}(24,k,2,3)$'s for 6?k?11$6?k?11$, OA(32,k,2,3)$\mathrm{OA}(32,k,2,3)$'s for 6?k?11$6?k?11$, OA(40,k,2,3)$\mathrm{OA}(40,k,2,3)$'s for 6?k?10$6?k?10$, OA(48,k,2,3)$\mathrm{OA}(48,k,2,3)$'s for 6?k?8$6?k?8$, OA(56,k,2,3)$\mathrm{OA}(56,k,2,3)$'s, OA(80,k,2,4)$\mathrm{OA}(80,k,2,4)$'s, OA(112,k,2,4)$\mathrm{OA}(112,k,2,4)$'s, for k=6,7$k=6,7$, OA(64,k,2,4)$\mathrm{OA}(64,k,2,4)$'s, OA(96,k,2,4)$\mathrm{OA}(96,k,2,4)$'s for k=7,8$k=7,8$, and OA(144,k,2,4)$\mathrm{OA}(144,k,2,4)$'s for k=8,9$k=8,9$. Further applications to classifying covering arrays with the minimum number of runs and packing arrays with the maximum number of runs are presented.     

A linear programming bound for orthogonal arrays with mixed levels

We show how the Delsarte theory can be used to obtain a linear programming bound for orthogonal arrays with mixed levels. Even for strength 2 this improves on the Rao bound in a large number of cases. The results point to several interesting sets of parameters for which the existence of the arrays is at present undecided.     

Generalized minimum aberration mixed-level orthogonal arrays: A general approach based on sequential integer quadratically constrained quadratic programming

Orthogonal fractional factorial designs and in particular orthogonal arrays (OAs) are frequently used in many fields of application, including medicine, engineering, and agriculture. In this article, we present a methodology and an algorithm to find an OA, of given size and strength, which satisfies the generalized minimum aberration criterion. The methodology is based on the joint use of polynomial counting functions, complex coding of levels, and algorithms for quadratic optimization and puts no restriction on the number of levels of each factor.     

The Hamming distances of saturated asymmetrical orthogonal arrays with strength 2

Abstract

Orthogonal arrays have many connections to other combinatorial designs and are applied in coding theory, the statistical design of experiments, cryptography, various types of software testing and quality control. In this paper, we present some general methods to find the Hamming distances for saturated asymmetrical orthogonal arrays (SAOAs) with strength 2. As applications of our methods, the Hamming distances of SAOA parents of size less than or equal to 100 are obtained. We also provide the Hamming distances of the SAOAs constructed from difference schemes or by the expansive replacement method. The feasibility of Hamming distances is discussed.     

On the construction of balanced and orthogonal arrays

Two series of three symbol balanced arrays of strength two are constructed. Using special classes of BIB designs, two classes of two symbol orthogonal arrays of strength three are constructed.     

Construction of some new families of nested orthogonal arrays

Nested orthogonal arrays have been used in the design of an experimental setup consisting of two experiments, the expensive one of higher accuracy being nested in a larger and relatively less expensive one of lower accuracy. In this paper, we provide new methods for constructing two types of nested orthogonal arrays.     

Some new constructions of strength 3 mixed orthogonal arrays

Orthogonal arrays of strength 3 permit estimation of all the main effects of the experimental factors free from confounding or contamination with 2-factor interactions. We introduce methods of using arithmetic formulations and Latin squares to construct mixed orthogonal arrays of strength 3. Although the methods could be well extended to computing larger arrays, we confine computing to at most 100 run orthogonal arrays for practical uses. We find new arrays with run sizes 80 and 96, each has many distinct factor levels.     

Robust estimation for non-homogeneous data and the selection of the optimal tuning parameter: the density power divergence approach

The density power divergence (DPD) measure, defined in terms of a single parameter α, has proved to be a popular tool in the area of robust estimation [1 A. Basu, I.R. Harris, N.L. Hjort and M.C. Jones, Robust and efficient estimation by minimizing a density power divergence, Biometrika 85 (1998), pp. 549–559. doi: 10.1093/biomet/85.3.549[Crossref], [Web of Science ®] , [Google Scholar]]. Recently, Ghosh and Basu [5 A. Ghosh and A. Basu, Robust estimation for independent non-homogeneous observations using density power divergence with applications to linear regression, Electron. J. Stat. 7 (2013), pp. 2420–2456. doi: 10.1214/13-EJS847[Crossref], [Web of Science ®] , [Google Scholar]] rigorously established the asymptotic properties of the MDPDEs in case of independent non-homogeneous observations. In this paper, we present an extensive numerical study to describe the performance of the method in the case of linear regression, the most common setup under the case of non-homogeneous data. In addition, we extend the existing methods for the selection of the optimal robustness tuning parameter from the case of independent and identically distributed (i.i.d.) data to the case of non-homogeneous observations. Proper selection of the tuning parameter is critical to the appropriateness of the resulting analysis. The selection of the optimal robustness tuning parameter is explored in the context of the linear regression problem with an extensive numerical study involving real and simulated data.     

Mean squared error of prediction approach to the analysis of a combined array

SUMMARY The combined array provides a powerful, more statistically rigorous alternative to Taguchi's crossed-array approach to robust parameter design. The combined array assumes a single linear model in the control and the noise factors. One may then find conditions for the control factors which will minimize an appropriate loss function that involves the noise factors. The most appropriate loss function is often simply the resulting process variance, recognizing that the noise factors are actually random effects in the process. Because the major focus of such an experiment is to optimize the estimated process variance, it is vital to understand the resulting prediction properties. This paper develops the mean squared error for the estimated process variance for the combined array approach, under the assumption that the model is correctly specified. Specific combined arrays are compared for robustness. A practical example outlines how this approach may be used to select appropriate combined arrays within a particular experimental situation.     

A note on embedding of orthogonal arrays of strength two

The problem of embedding an orthogonal array of strength 2 into a complete orthogonal array is discussed. It is shown that for any n≠4 any orthogonal array of strength 2 and deficiency 2 can always be embedded into a corresponding complete orthogonal array.     

On the construction of mixed orthogonal arrays of strength two

The generalized Kronecker sum was used by Wang and Wu (J. Amer. Statist. Assoc. 86 (1991) 450) and Dey and Midha (Statist. Probab. Lett. 28 (1996) 211; Proc. AP Akad. Sci. 5 (2001) 39) to construct mixed orthogonal arrays. We modify their methods to obtain several families of mixed orthogonal arrays. Some new arrays with run size less than 100 are found.     

A note on the minimum size of an orthogonal array

It is an elementary fact that the size of an orthogonal array of strength t on k factors must be a multiple of a certain number, say L_t, that depends on the orders of the factors. Thus L_t is a lower bound on the size of arrays of strength t on those factors, and is no larger than L_k, the size of the complete factorial design. We investigate the relationship between the numbers L_t, and two questions in particular: For what t is L_t < L_k? And when L_t = L_k, is the complete factorial design the only array of that size and strength t? Arrays are assumed to be mixed-level.

We refer to an array of size less than L_k as a proper fraction. Guided by our main result, we construct a variety of mixed-level proper fractions of strength k ? 1 that also satisfy a certain group-theoretic condition.     

Efficient arrangements of two-level orthogonal arrays in two and four blocks

When orthogonal arrays are used in practical applications, it is often difficult to perform all the designed runs of the experiment under homogeneous conditions. The arrangement of factorial runs into blocks is usually an action taken to overcome such obstacles. However, an arbitrary configuration might lead to spurious analysis results. In this work, the nice properties of two-level orthogonal arrays are taken into consideration and an effective method for arranging experimental runs into two and four blocks of the same size is proposed. This method is based on the so-called J-characteristics of the corresponding array. General theoretical results are given for studying up to four experimental factors in two blocks, as well as for studying up to three experimental factors in four blocks. Finally, we provide best blocking arrangements when the number of the factors of interest is larger, by exploiting the known lists of non-isomorphic orthogonal arrays with two levels and various run sizes.