An efficient new job release control methodology

This paper presents a new job release methodology: WIPLOAD Control (WIPLCtrl), which is a workload limited release methodology for the overall shop floor. The behaviour of WIPLCtrl is analyzed using the Markov process model of a transfer line system to observe the potential advantage of WIPLCtrl relative to the conventional measure of system workload using the WIP level. A simulation experiment is then conducted considering a simplified semiconductor wafer fabrication system. The performance of WIPLCtrl is compared with that of an open-loop release methodology and three other closed-loop workload limited release methodologies. A case study is also carried out by simulating a real-life wafer fabrication, considering the complexities and the distinguishing features of semiconductor manufacturing. The experimental results indicate that WIPLCtrl is an efficient job release methodology, capable of reducing both the mean and the standard deviation of the cycle time for a given throughput level. The performance of WIPLCtrl appears to be reliable with the increase of system congestion level and the increase of system variability caused by stochastic factors such as machine unreliability and processing time variability.