A dynamic forecasting model for job shop flowtime prediction and tardiness control

This paper presents a forecasting approach to flowtime prediction in a job shop. The flowtime precision relationship developed considers both job characteristic and shop loading information. Forecast errors are shown to be approximately normally distributed. A lateness feedback approach is also developed to dynamically estimate the variance of forecast error. The estimated distribution of forecast error is used to set delivery safety allowances which are based on a desired level of delivery performance. Results show that the lead times required to maintain a desired level of delivery performance are lowest when due-date dependent dispatch is used.