HOW TO ACHIEVE A ROBUST PROCESS USING RESPONSE-SURFACE METHODOLOGY

Taguchi's experimental system is used to develop robust processes, that is, processes which are insensitive to variations in uncontrollable variables, such as environmental variables. Taguchi's recommended designs can be considered to be response surface designs. These designs are compared to composite designs, which are classical response surface designs, and a new class of composite designs having mixed resolution is suggested as a competitor to Taguchi designs. An example is given of types of problems where it is better not to achieve a robust process, and customizing for specific customers can lead to greater customer satisfaction. An introduction to response surface methodology and to screening designs is used to set the stage.