Linear Regression Tree for Multistage Manufacturing Process Control

Reconfigured Piecewise Linear Regression Tree for Multistage Manufacturing Process Control Ran Jin and Jianjun Shi H. Milton Stewart School of Industrial and Systems Engineering, Georgia Institute of Technology, 765 Ferst Drive NW, Atlanta, GA 30332, USA, E-mail: [email protected] Abstract In a multistage manufacturing process, massive observational data are obtained from the measurements of the product quality features, process variables, and material properties. Those data have temporal and spatial relationships and may have nonlinear data structures. It is a challenge task to model the variation and its propagation from those observational data and further use the model for feedforward control purposes. This paper proposes a methodology of feedforward control based on piecewise linear models in the aforementioned circumstance. An engineering-driven reconfiguration method for piecewise linear regression trees is proposed. The model complexity is further reduced by merging the leaf nodes with the constraint of the control accuracy requirement. A case study in a multistage wafer manufacturing process is conducted to illustrate the procedure and effectiveness of the proposed method.