Nearly in all papers of Grey relational analysis, Taguchi method is used to complement it than other types of experimental designs such as central composite design. Why?
There is no scientific explanation to it. Nevertheless, you can easily find out in Researchgate platform papers exposing criticisms to called Taguchi method and difficulties or errors in "classical" Design of Experiments or RSM approach use.
Article Estimating the effect of process parameters on MRR, TWR and ...
Article Optimization of Multiple Performance Characteristics of Fric...
In my understanding, GRA is used to solve multi-attribute problems (ie optimization based on two conditions: ie optimize operation condition to maximum removal and minimal cost for example) where its not possible with traditional Taguchi methods.
It is also possible to integrate GRA with other DOE. However other methods are also possible for other DOE for multie-attribute problems. Optimization could be done via other forms of combinations. (ie in RSM you overlap the optimized contour plots to see the overlapped condition).