A new desirability function-based method for correlated multiple response optimization

Design of process parameters to meet the required specification of quality characteristics has become a crucial issue for many industries. The underlying relationships between process inputs and outputs are one of the main tasks in quality engineering. Most of researches assume independency and consider same relative importance of quality characteristics with constant variances over experimental space. This study represents a novel robust approach based on desirability function and global criterion methods that not only obtains the parameter design but also considers different variance, correlation, and relative importance level of outputs. The suggested method enforces all quality measurements to fall within specification limits. To illustrate computational aspects of the proposed method, two realistic examples have been conducted. The obtained results demonstrate the superiority of the proposed approach with respect to the existing approaches.