抄録
In the last edition, I explained that we used the SN Ratio to show system functionabilities and dispersions of output characteristics in "Quality Engineering". ln this edition, I would like to explain how we should evaluate the functionabilities and the dispersions in various conditions. When people evaluate the characteristic dispersions, they make many samples repeatedly, and wait for random errors to occur. However, engineers who know Robust Design forcefully change the levels of noise(the causes of dispersion) and input signal to see how the output actually vary.
