Adaptive pull-type production control using Kalman Filters

Abstract

The purpose of this research was to design a closed-loop process control methodology for a pull-type assembly line. It will aid in meeting the required production quota and in minimizing shipping and work-in-process inventory, particularly in environments with high levels of absenteeism and turnover that directly affect work station yield rates. The adaptive pull-type process control system (APPCS) allows pull-type production, with its associated benefits, to be successfully implemented in environments with low and variable yield rates. This is accomplished by dynamically feeding back production information to adjust the number of parts that should be included in a production container. The production control methodology uses a Kalman filter that predicts the amount of pieces the first workstation must place in a production tote in order to meet the target production quota. The key to the algorithm is the weighting function, which statistically dampens the effects of going beyond or below the target production quota. The APPCS was simulated on a pull-type production line with high levels of turnover, absenteeism, and poor yield rates. The results, compared to historical data, were encouraging with future work needed on the development of a forcing function and a covariance matrix for inspection error.