If a parameter is continuous, we can choose the option “Manufacturable Values” to define a post-processing filter on this parameter. We can take the example of a parameter “thickness” that varies from 5 to 25mm with manufacturable values set to 5, 10, 12, 20 and 25. If “thickness” is enabled for optimization, the optimizer can only choose between the manufacturable values defined previously, even if the parameter itself is continuous. For example, the optimizer cannot take the value 15mm as it is filtered out in “manufacturable values”. In Design of Experiments, these manufacturable values are indeed not respected. DOE generates the surface response using all the values, and then applies the “use manufacturable values” as a post-processing filter. If a parameter is discrete, then its values are neither continuous nor ordered. There is no interpolation possible between these values and thus 1 DOE for every discrete value is generated. Example: if P1 is a discrete parameter and P2 and P3 are continuous, then we will have a DOE over P2 and P3 for every discrete value of P1. In summary, a parameter can be either continuous or discrete. If it is continuous, there is the option to define the filter manufacturable values on it.