A discussion on performance value versus performance order

Ordinal optimization techniques help us locating "good" designs efficiently. Previous studies of ordinal optimization define the "good" designs as those coming at the top of the "order" of performance. In this study, we argue that with the help of practical prior knowledge and observations, we can establish quantitative relationship between the good design found and the performance "values" of it. Therefore, the analysis of ordinal optimization can be carried out to identify designs that are good by some criterion of their performance "values.".