Stochastic re-entrant line modeling for an environment stress testing in a semiconductor assembly industry

摘要

Environmental stress testing (EST) is a reliability test for the semiconductor products that is concerned with how products perform its intended functions under some environmental conditions. Planning, implementation and delivery schedules are affected by the outcome of these tests and it is crucial to determine the throughput of the EST process. In this paper we present an efficient modified analytical model based on approximate mean value analysis (MVA) with probabilistic re-entrant line to predict the total mean waiting time and subsequently the mean throughput rate for the EST process. Using the analytical and simulation method, we analyse a five-stage queuing system with re-entrant to the second stage under various stochastic routing. The MVA algorithm has been modified to deal with this situation. Results show that the modified algorithm can deal with situations involving small and large number of lots respectively.