Robustness for multiple response problems using a loss model

Posits that methods to achieve robustness for single response are well established. For multiple response problems, however, very few methods have been developed successfully. Presents a new method of achieving robustness for multiple response problems using Taguchi’s loss model. Develops loss models for all the responses involved and identifies a design solution that minimizes the total cost loss. This formulation leads to a solution with minimal variance and also has an inherent property of optimizing the manufacturing yield of the design. Compares the proposed method with another formulation. Shows that the proposed method gave better results.