Vol. 11, Special Issue 5 (2022)

Author(s):
Ankita Verma, Rajeev Kumar, Vinayaka, Kishor Gavhane and Bulbul Ahmed

Abstract:
The goal of many studies is to choose a design that allows a good estimate of the relationship between explanatory factors and the response of interest. The design of the RESPONSE surfaces allows adjustment of the response surfaces and provides a means of checking their suitability. There are many measure to study the predictive performance of responsive surface designs for comparison of designs, but these are mainly the single-valued efficiency type criteria. The graphical method is a useful tool for assessing the overall predictability of an experimental design over the entire area of interest. Variance dispersion graph (VDG) are useful for comparing competing designs over a fixed design space. However, they do not provide all useful information about the predictability of a design. The fraction of design space (FDS) technique provides more detailed information by quantifying the portion of the design space where the scale prediction variance (SPV) is less than or equal to any predefined value. The FDS technique complements the use of VDGs to better understand the predictability of a design. Here, these two graphical tools have been described. Several standard response surface designs with varying number of elements are discussed with these two methods.