EVALUASI RANCANGAN EKSPERIMEN PADA MODEL PERMUKAAN MULTIRESPON YANG MELIBATKAN FAKTOR NOISE
Abstract
This article aim to study process design having the character of steady (robust design), in the effort is optimal overall of or response. Applying method merging between inner arrays and outer arrays in one matrixes designs so-called combined arrays can close over insuffiency from products arrays. Taguchi based on the design made by response model as design of experiments integration Taguchi with method of surface of response ( RSM), then determine valuation mean response variance and response based on the response model appraiser. Some bottoms design of experimentses evaluated by using criteria Variance dispersion graphs (VDGs) with amount of appropriate experiment. Application earnings yield indicate that modified CCD have variance slope compared is smaller with bottom design MR. But, from Prediction scale earnings yield Variansi Residual maximum and the average of design augmented MR compared by is smaller modified CCD. So that design augmented MR consider as effort cost effective and time, because amount of point of small relative experiment. However in general both the design is proper consider, if paying attention to earnings yield from minimum variance..
Keyword:Robust-Variance-Dispersion-Graphs ( VDGs), multirespon, combined-array,
This license allows authors to copy, redistribute, remix, transform, and build upon the Work, in any format or medium, for any purpose including commercial purpose, on a perpetual basis provided they credit the Work and the authors. Authors
must explain any changes that were made from the original and may not suggest the authors endorse the use. The resultant work must be made available under the same terms, and must include a link to the CC BY 4.0 International License.