제안 방법

• section. The classification is performed in two aspects: statistical properties and the DM’s preference. Through the classification, limitations of the existing work and thus insights on the new direction in multiresponse optimization could be identified.
• In the optimization stage, two questions must be addressed: “what-to-optimizee (optimization goal) and “how-to-optimize-it'' (solution technique). This paper focuses on the “what-to-optimize' aspect, assuming that the data have been collected and the response models have been fitted reasonably well. A multiresponse optimization problem is formally defined as:
• The review and the classification aim to provide useful information about multiresponse optimization and to show the future directions, respectively. This paper has described the existing work under the category of the graphical, priority-based, desirability function, loss function, process capability, and probability-based approach. Then, it has been classified in two aspects: statistical properties and the DM’s preference.