Flow-time estimation by synergistically modeling real and simulation data

Abstract

The ability to quote a competitive and reliable lead time for a new order is a key competitive advantage for manufacturers and plays a significant role in customer acquisition and satisfaction. Quoting a precise and reliable lead time requires a good prediction for the flow time of a new order. This research focuses on quantifying the dependence of the flow time upon observed job shop status variables, the size of a new order, and the arrival rate of future orders. An iterative fitting procedure based on stochastic kriging with qualitative factors, is developed to synergistically model simulation and real manufacturing data, for the prediction of a new order's flow time.