Job shop flowtime prediction and tardiness control using queueing analysis

Abstract

This paper investigates flowtime prediction under conditions where Jackson's decomposition principle can be applied. Four models in which due-date setting rule parameters are based on predicted flowtime are developed and compared. Simulation results show both job characteristic and dynamic shop load information to be useful in predicting flowtimes. Analysis of prediction deviations shows that good predictions lead to errors which are approximately normally distributed. The variance of prediction errors can also be analytically determined. Therefore, quoted delivery dates can be set which are consistent with a desired level of delivery performance.