Impacts of quality and processing time uncertainties in multistage production system

The primary objectives of this research are to develop simulation models for multistage production system under processing time and quality variation, to identify areas of potential bottleneck in production system and to determine the optimum production lot size for each station in a multistage production system under the uncertainties to minimize the WIP level and lead time and thereby the total system cost. A simulation model is developed based on a live case from a Malaysian company. Taguchi approach for orthogonal array is used in designing experiments and these are executed in WITNESS. The models are verified and validated by face validity, the historical data from the company and analytical model. The delivery performances, average lead time and work-in-progress (WIP) in the system, are examined for different experimental scenarios. Interaction effects and confirmation tests are also performed. The optimal batch sizes respectively for polishing unit (PU), quality control (QC) and packing stations are 8, 16 and 3 at minimum WIP and lead time. If the company uses these optimal batch sizes, a total of 24% improvement can be obtained. The simulation models show that Gasketing station is the bottleneck and batch size selection in PU station is the most critical decision in the system. The interaction effects are insignificant. The main contribution of this research is determination of the optimal lot sizes under imperfect quality of product and stochastic processing time. This approach can be generalized to any multistage production system, regardless of the precedence relationships among the various production stages in the system.