Determining job completion time distributions in stochastic production environments

Abstract

As responsiveness becomes more of a competitive issue, knowledge of completion times of jobs and orders on the shop floor is key to the success and even survival of many manufacturing firms. In this paper, we present two approaches - a fast sample path generation technique and a numerical technique - to determine the distribution of completion times of jobs in stochastic production environments. We compare our results with those obtained from conventional discrete-event simulation and also compare speed of execution. We also show that the distribution obtained from the numerical integration technique provides a lower bound to the actual distribution of the completion times.