Multi response simulation and optimization of gas tungsten arc welding

Patil, Mukundraj V.

doi:10.1016/j.apm.2016.10.033

Cited by 14 publications

(9 citation statements)

References 5 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…While the 6-factor design with 17 runs of Stone et al (2014) provides a smaller average absolute correlation than the 6-factor sDSD, it has a larger maximum absolute correlation (0.66 versus 0.5). The 7-factor design with 21 runs and k = 3 of Patil (2017) has larger maximum and average absolute correlations between its pairs of 2FI contrast vectors than the 7-factor design with 17 runs and k = 1.…”

Section: Aliasing Of 2fismentioning

confidence: 92%

“…Table 5 shows that we can drop any single column, any pair of columns and any triplet of columns from a (m + k)-factor sDSD without affecting the 2FIs' contrast vectors' correlations (in other words, without affecting the aliasing among the 2FIs), except when starting from the 16-factor sDSD. As a result, Dougherty et al (2015), Fidaleo et al (2016), Patil (2017, and Stone et al (2014) coincidentally used the best possible pDSD for their experiment. Any other choice of columns to drop would have resulted in an equivalent pDSD for their experiments.…”

Section: Best Sets Of K Columns To Dropmentioning

confidence: 99%

“…Figure 4 shows that the design options used by Patil (2017) and Stone et al (2014) were not optimal in terms of the maximum absolute correlation between pairs of 2FI contrast vectors. While the 6-factor design with 17 runs of Stone et al (2014) provides a smaller average absolute correlation than the 6-factor sDSD, it has a larger maximum absolute correlation (0.66 versus 0.5).…”

Section: Aliasing Of 2fismentioning

confidence: 99%

“…For instance, when comparing different cost-efficient screening designs, Stone et al (2014) preferred a 6-factor design with 17 runs constructed by dropping two columns from the 8-factor sDSD to a 6-factor 13-run sDSD, due to the relatively high aliasing between pairs of 2FIs and between a QE and a 2FI in the 13-run design. Patil (2017) studied the impact of seven factors on a welding process where some 2FIs were expected to be active, and observed that the 7-factor design formed by dropping one column from the 8-factor sDSD exhibited a substantial amount of aliasing among the interactions. To reduce the aliasing, he dropped three columns from the 10-factor sDSD in Table 1, and thus used a 21-run design instead of a 17-run design.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Projections of Definitive Screening Designs by Dropping Columns: Selection and Evaluation

2019

View full text Add to dashboard Cite

Definitive screening designs are increasingly used for studying the impact of many quantitative factors on one or more responses in relatively few experimental runs. In practical applications, researchers often require a design for m quantitative factors, construct a definitive screening design for more than m factors and drop the superfluous columns. This is done when the number of runs in the standard m-factor definitive screening design is considered too limited or when no standard definitive screening design exists for m factors. In these cases, it is common practice to arbitrarily drop the last column of the larger definitive screening design. In this paper, we show that certain statistical properties of the resulting experimental design depend on which columns are dropped and that other properties are insensitive to the exact columns dropped. We perform a complete search for the best sets of 1-8 columns to drop from standard definitive screening designs with up to 24 factors. We observed the largest differences in statistical properties when dropping four columns from 8-and 10-factor definitive screening designs. In other cases, the differences are moderate or small, or even nonexistent. Our search for optimal columns to drop necessitated a detailed study of the properties of definitive screening designs. This allows us to present some new analytical and numerical results concerning definitive screening designs.

show abstract

Section: Aliasing Of 2fismentioning

confidence: 92%

Section: Best Sets Of K Columns To Dropmentioning

confidence: 99%

Section: Aliasing Of 2fismentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Projections of Definitive Screening Designs by Dropping Columns: Selection and Evaluation

2019

View full text Add to dashboard Cite

show abstract

“…Since their conception, a considerable effort has been invested into the further development of the definitive screening designs; references include Jones and Nachtsheim (2013), Georgiou, Stylianou and Aggarwal (2014), Nguyen and Pham (2016), Jones and Nachtsheim (2017) and Nachtsheim, Shen and Lin (2017). Recent applications have been described by Dougherty et al (2015), Fidaleo et al (2016) and Patil (2017).…”

mentioning

confidence: 99%

A classification criterion for definitive screening designs

Schoen

Eendebak²,

Goos³

2019

Ann. Statist.

View full text Add to dashboard Cite

A conference design is a rectangular matrix with orthogonal columns, one zero in each column, at most one zero in each row, and −1s and +1s elsewhere. A definitive screening design can be constructed by folding over a conference design and adding a row vector of zeroes. We prove that, for a given even number of rows, there is just one isomorphism class for conference designs with two or three columns. Next, we derive all isomorphism classes for conference designs with four columns. Based on our results, we propose a classification criterion for definitive screening designs based on projections into four factors and illustrate its potential by studying designs with 24 factors. * Supported by the Research Foundation -Flanders FWO MSC 2010 subject classifications: Primary 62K20; secondary 05B20, 94C30

show abstract