Online automatic process control using observable noise factors for discrete-part manufacturing.

1 Introduction

In manufacturing processe there are many proces variables that interact in a complicated nonlinear fashion. In general, these variables are classified into controllable factors x (variables which can be easily manipulated online) and noise, or uncontrollable, factors n (variables that vary randomly and are difficult to manipulate online). If y is used to denote the a whole response, the relationship between x n and y can be generally press outed as:

y = f (x n) (1)

For a discrete-part manufacturing proces the relationship in Equation (1) can ofttimes be expressed in the format of a regression type (Box and Draper, ...