Quality Engineering deals with the robustness of function of a product over the all customer's usage conditions.

The role of management is to make product designess to design robust products so there would be few market claim after entering actual production and shippig them to market. This article discuss the methods that how to make design people to do robust product design. The methods stated here is to test the robustness of the product by taking a Signal factor/s and a Noise factor/s to measure S/N ratio without showing test conditions of noise for the acceptance of the design.

As the robust design in quality engineering means the product works at various customer's usage conditions. Even we know accurate complete mathematical model, we have to do robust design process for functionality improvement at customer usage conditions. Mathematical modets don't usually include various effects of customers usage conditions. We have to generate variation of each design parameter around each nominal levels selected. Those variations should include not only manufacturing variations but also variations due to environment conditions and deteriorations at customer's conditions. For each setting levels of control factors, two compounded noise levels are set as follows.

N₁ : the level which gives minus side output.

N₂ : the level which gives plus side output.

As there is usually non-linear mathematical terms for signals, Signal-to-Noise ratio on nominal-is-best is used for robustness using computer simulation design.